HPCWIKI - User contributions [en]

Node.js tips and tricks

2026-10-30T05:32:02Z

Admin:

== Overview ==

Node.js tips and tricks에 대한 기술 문서입니다.

== Key Concepts ==

{| class="wikitable"
! Concept
! Description
! Related
|-
| Node.js tips and tricks
| HPC/서버 환경에서 중요한 기술 개념
| [[Linux]], [[Server]]
|}

== Detailed Explanation ==

<syntaxhighlight lang="bash">
# Install npm-check-updates globally
$ npm i -g npm-check-updates
# Inside of nodejs project folder where package.json is located
$ ncu -u
Upgrading /opt/imp/home/prj/package.json
[====================] 39/39 100%
...
async ^2.6.3 → ^3.2.6
cheerio ^1.0.0-rc.2 → ^1.0.0
cookie-parser ^1.4.4 → ^1.4.7
....
# Run npm install to install new versions.
$ npm install
</syntaxhighlight>
<syntaxhighlight lang="bash">
curl https://raw.githubusercontent.com/creationix/nvm/master/install.sh | bash
source ~/.bashrc
# List up node version
nvm list-remote

== Best Practices ==

* 최신 버전 사용 권장
* 공식 문서 참고
* 테스트 환경에서 먼저 검증

== References ==

* [https://wiki.hpcmate.com Node.js tips and tricks]

== Related Pages ==

* [[Linux]]
* [[Server]]
* [[Hardware]]
* [[Network]]

[[Category:Server]]
== Knowledge Graph ==

Related

→ [[Linux]]
→ [[Server]]
→ [[Hardware]]
→ [[Network]]

[[Category:Reference]]

Warranty

2026-10-27T10:20:20Z

Admin: /* 서비스요금 산정 기준 */

== Overview ==

HPCMATE 제품보증 및 지원은 다음과 같습니다.

== 제품 보증 ==
제품 보증기간은 제품 보증규정에 명시되어 있는 기준일로부터 시작됩니다.
{| class="wikitable"
|-
! 구분 !! 정의 및 품목 !! 보증기간
|-
| SYSTEM|| 시스템 구동에 필요한 최소한의 H/W로, Chassis, Main Board,CPU, RAM, HDD가 조립된 제품 || 3년 (무상 1년, 실비/유상 2년)
|-
| 단일 구성품 || 조립되어 있는 상태가 아닌 단일품목 Chassis, Main Board, Add-on-card, Power supply, Switch 등 || 1년(무상 3개월, 실비/유상 9개월)
|}
<nowiki>**</nowiki>유/무상 서비스에 명시되어있는 예외 기준외 사용자 부주의 등에 대한 불량 발생은 보증기간내에 발생하더라도 별도의 실비/유상 비용이 청구 됩니다.

=== 교체 및 A/S입고 공동사항 ===

* RMA 제품 교체 : 하자 부품은 당사에 먼저 입고하면, 업무일 5일 이내 하자 제품 내용 확인 (제조사가 지정한 특정 상황은 예외, 예- SSD의 TWB 한계 초과) , 하여 보증기간 내 제품은 1:1 교환하여 출고 드립니다. 입고 방법은 택배를 이용해 주시기 바라며 입고시 발생되는 물류비는 당사 50%, 고객사 50% 가 원칙 입니다. ( 당사, 고객사 선불로 이용) 단, 그밖에 퀵 또는 중량화물을 이용하는 경우 고객사에서 100% 부담 해주셔야 합니다
* 방문 진단 및 교체 : 하자 부품을 당사 직원이 방문해서 교체해 드리는 방법은 지역에 따라 경비 (출장비/진단비)가 청구 됩니다.
* 온라인 기술지원 : 온라인 기술지원은 당사 Wiki 및 기술지원창구 (http://support.hpcmate.com) 을 통해 지원받을 수 있습니다.
* 오프라인 기술지원 : 고객 사이트에 직접 방문하여 기술지원하는 방식으로 사전 일정예약 및 출장비, 진단비, 조치비등 서비스 요금이 부가 됩니다.

=== 서비스요금 산정 기준 ===
서비스 요금은 출장비, 진단비, 처리비 및 기타로 구분될 수 있으며 2026년 기준 지역별 기본 출장비는 다음과 같습니다.
{| class="wikitable"
|+
!지역
!오프라인 기본 출장비
(부가세별도)
!온라인 원격 서비스
(부가세별도)
!노트
|-
|서울강남지역
|150,000 만원
| rowspan="8" |50,000/시간당
| rowspan="8" |진단/처리비 별도
|-
|서울강북지역
|170,000 만원
|-
|경기지역
|220,000 만원
|-
|강원지역
|270,000 만원
|-
|충정지역
|300,000 만원
|-
|전라북도/경상북도
|350,000 만원
|-
|전라남도/경상남도
|450,000 만원
|-
|제주도 및 기타지녁
|협의
|}
[[Category:Server]]
[[Category:Reference]]

Warranty

2026-10-27T10:19:53Z

Admin:

== Overview ==

HPCMATE 제품보증 및 지원은 다음과 같습니다.

== 제품 보증 ==
제품 보증기간은 제품 보증규정에 명시되어 있는 기준일로부터 시작됩니다.
{| class="wikitable"
|-
! 구분 !! 정의 및 품목 !! 보증기간
|-
| SYSTEM|| 시스템 구동에 필요한 최소한의 H/W로, Chassis, Main Board,CPU, RAM, HDD가 조립된 제품 || 3년 (무상 1년, 실비/유상 2년)
|-
| 단일 구성품 || 조립되어 있는 상태가 아닌 단일품목 Chassis, Main Board, Add-on-card, Power supply, Switch 등 || 1년(무상 3개월, 실비/유상 9개월)
|}
<nowiki>**</nowiki>유/무상 서비스에 명시되어있는 예외 기준외 사용자 부주의 등에 대한 불량 발생은 보증기간내에 발생하더라도 별도의 실비/유상 비용이 청구 됩니다.

=== 교체 및 A/S입고 공동사항 ===

* RMA 제품 교체 : 하자 부품은 당사에 먼저 입고하면, 업무일 5일 이내 하자 제품 내용 확인 (제조사가 지정한 특정 상황은 예외, 예- SSD의 TWB 한계 초과) , 하여 보증기간 내 제품은 1:1 교환하여 출고 드립니다. 입고 방법은 택배를 이용해 주시기 바라며 입고시 발생되는 물류비는 당사 50%, 고객사 50% 가 원칙 입니다. ( 당사, 고객사 선불로 이용) 단, 그밖에 퀵 또는 중량화물을 이용하는 경우 고객사에서 100% 부담 해주셔야 합니다
* 방문 진단 및 교체 : 하자 부품을 당사 직원이 방문해서 교체해 드리는 방법은 지역에 따라 경비 (출장비/진단비)가 청구 됩니다.
* 온라인 기술지원 : 온라인 기술지원은 당사 Wiki 및 기술지원창구 (http://support.hpcmate.com) 을 통해 지원받을 수 있습니다.
* 오프라인 기술지원 : 고객 사이트에 직접 방문하여 기술지원하는 방식으로 사전 일정예약 및 출장비, 진단비, 조치비등 서비스 요금이 부가 됩니다.

=== 서비스요금 산정 기준 ===
서비스 요금은 출장비, 진단비, 처리비 및 기타로 구분될 수 있으며 2026년 기준 지역별 기본 출장비는 다음과 같습니다.
{| class="wikitable"
|+
!지역
!오프라인 기본 출장비
(부가세별도)
!온라인 원격 서비스
(부가세별도)
!노트
|-
|서울강남지역
|150,000 만원
| rowspan="8" |50,000/시간당
| rowspan="8" |진단/처리비 별도
|-
|서울강북지역
|170,000 만원
|-
|경기지역
|220,000 만원
|-
|강원지역
|270,000 만원
|-
|충정지역
|300,000 만원
|-
|전라북도/경상북도
|350,000 만원
|-
|전라남도/경상남도
|400,000 만원
|-
|제주도 및 기타지녁
|협의
|}
[[Category:Server]]
[[Category:Reference]]

Warranty

2026-10-27T09:55:18Z

Admin:

== Overview ==

HPCMATE 제품보증 및 지원은 다음과 같습니다.

== 제품 보증 ==
제품 보증기간은 제품 보증규정에 명시되어 있는 기준일로부터 시작됩니다.
{| class="wikitable"
|-
! 구분 !! 정의 및 품목 !! 보증기간
|-
| SYSTEM|| 시스템 구동에 필요한 최소한의 H/W로, Chassis, Main Board,CPU, RAM, HDD가 조립된 제품 || 3년 (무상 1년, 실비/유상 2년)
|-
| 단일 구성품 || 조립되어 있는 상태가 아닌 단일품목 Chassis, Main Board, Add-on-card, Power supply, Switch 등 || 1년(무상 3개월, 실비/유상 9개월)
|}
<nowiki>**</nowiki>유/무상 서비스에 명시되어있는 예외 기준외 사용자 부주의 등에 대한 불량 발생은 보증기간내에 발생하더라도 별도의 실비/유상 비용이 청구 됩니다.

=== 교체 및 A/S입고 공동사항 ===

* RMA 제품 교체 : 하자 부품은 당사에 먼저 입고하면, 업무일 5일 이내 하자 제품 내용 확인 (제조사가 지정한 특정 상황, 예- SSD의 TWB 한계) 하여 보증기간 내 제품은 1:1 교환하여 출고 드립니다. 입고 방법은 택배를 이용해 주시기 바라며 입고 시 발생되는 물류비는 당사 50%, 고객사 50% 가 원칙 입니다. ( 당사, 고객사 선불로 이용) 단, 그밖에 퀵 또는 화물로 이용하는 경우 고객사에서 부담 해주셔야 합니다
* 방문 진단 및 교체 : 하자 부품을 당사 직원이 방문해서 교체해 드리는 방법은 지역에 따라 벼도의 경비 (출장비/진단비)가 청구 됩니다.
* 온라인 기술지원 : 온라인 기술지원은 당사 Wiki 및 기술지원창구 (http://support.hpcmate.com) 을 통해 지원받을 수 있습니다.
* 오프라인 기술지원 : 고객 사이트에 직접 방문하여 기술지원하는 방식으로 사전 일정예약 및 출장비, 진단비, 조치비등 서비스 요금이 부가 됩니다.

=== 서비스요금 산정 기준 ===
서비스 요금은 출장비, 진단비, 처리비 및 기타로 구분될 수 있으며 2026년 기준 지역별 기본 출장비는 다음과 같습니다.
{| class="wikitable"
|+
!지역
!기본 출장비 (부가세별도)
!노트
|-
|서울강남지역
|150,000 만원
| rowspan="8" |진단/처리비 별도
|-
|서울강북지역
|170,000 만원
|-
|경기지역
|220,000 만원
|-
|강원지역
|270,000 만원
|-
|충정지역
|300,000 만원
|-
|전라북도/경상북도
|350,000 만원
|-
|전라남도/경상남도
|400,000 만원
|-
|제주도 및 기타지녁
|협의
|}
[[Category:Server]]
[[Category:Reference]]

Warranty

2026-10-27T09:54:57Z

Admin:

== Overview ==

HPCMATE 제품보증 및 지원은 다음과 같습니다.

== Detailed Explanation ==

= 제품 보증 =
제품 보증기간은 제품 보증규정에 명시되어 있는 기준일로부터 시작됩니다.
{| class="wikitable"
|-
! 구분 !! 정의 및 품목 !! 보증기간
|-
| SYSTEM|| 시스템 구동에 필요한 최소한의 H/W로, Chassis, Main Board,CPU, RAM, HDD가 조립된 제품 || 3년 (무상 1년, 실비/유상 2년)
|-
| 단일 구성품 || 조립되어 있는 상태가 아닌 단일품목 Chassis, Main Board, Add-on-card, Power supply, Switch 등 || 1년(무상 3개월, 실비/유상 9개월)
|}
<nowiki>**</nowiki>유/무상 서비스에 명시되어있는 예외 기준외 사용자 부주의 등에 대한 불량 발생은 보증기간내에 발생하더라도 별도의 실비/유상 비용이 청구 됩니다.

=== 교체 및 A/S입고 공동사항 ===

* RMA 제품 교체 : 하자 부품은 당사에 먼저 입고하면, 업무일 5일 이내 하자 제품 내용 확인 (제조사가 지정한 특정 상황, 예- SSD의 TWB 한계) 하여 보증기간 내 제품은 1:1 교환하여 출고 드립니다. 입고 방법은 택배를 이용해 주시기 바라며 입고 시 발생되는 물류비는 당사 50%, 고객사 50% 가 원칙 입니다. ( 당사, 고객사 선불로 이용) 단, 그밖에 퀵 또는 화물로 이용하는 경우 고객사에서 부담 해주셔야 합니다
* 방문 진단 및 교체 : 하자 부품을 당사 직원이 방문해서 교체해 드리는 방법은 지역에 따라 벼도의 경비 (출장비/진단비)가 청구 됩니다.
* 온라인 기술지원 : 온라인 기술지원은 당사 Wiki 및 기술지원창구 (http://support.hpcmate.com) 을 통해 지원받을 수 있습니다.
* 오프라인 기술지원 : 고객 사이트에 직접 방문하여 기술지원하는 방식으로 사전 일정예약 및 출장비, 진단비, 조치비등 서비스 요금이 부가 됩니다.

=== 서비스요금 산정 기준 ===
서비스 요금은 출장비, 진단비, 처리비 및 기타로 구분될 수 있으며 2026년 기준 지역별 기본 출장비는 다음과 같습니다.
{| class="wikitable"
|+
!지역
!기본 출장비 (부가세별도)
!노트
|-
|서울강남지역
|150,000 만원
| rowspan="8" |진단/처리비 별도
|-
|서울강북지역
|170,000 만원
|-
|경기지역
|220,000 만원
|-
|강원지역
|270,000 만원
|-
|충정지역
|300,000 만원
|-
|전라북도/경상북도
|350,000 만원
|-
|전라남도/경상남도
|400,000 만원
|-
|제주도 및 기타지녁
|협의
|}
[[Category:Server]]
[[Category:Reference]]

Server Diagnostics

2026-10-14T05:51:29Z

Admin:

= Server Diagnostics =

== Overview ==

The HPCMATE Portal runs a '''read-only system diagnostic''' over SSH to produce a structured health report for any registered server. It never modifies the target — it connects, runs a battery of non-destructive checks, and scores the result.

=== Summary ===

* What: a deep, read-only health check that connects over SSH and inspects CPU, memory, disks, GPU, [[network]], services, the [[kernel]] log, and the /var/log files for hardware faults — and can be exported as a self-contained HTML report.
* Why: it turns raw system state into a single verdict (HEALTHY / DEGRADED / CRITICAL) plus per-item findings and recommendations, so a technician can see '''what is wrong and why it matters''' at a glance.
* When: before shipping a server, during pre-ship validation, or on demand from the Portal → Diagnostic view.

== Purpose ==

* Goal: give operators and customers a defensible, explainable health assessment.
* Scope: hardware health (GPU, disks, memory, thermals, network) and software/service state on a single host, over SSH.
* Non-goals: it does not install, repair, reboot, or change configuration. Multi-node cluster health is out of scope.

== How It Works ==

The Portal spawns a Python diagnostic over the serversetup venv (paramiko). It opens an SSH session to the target and runs a fixed set of shell probes. Every probe that crosses a threshold becomes a '''finding''' with a severity. Findings are aggregated into one overall verdict and a numeric health index. The whole run is read-only and typically takes a few seconds to a minute.

{| class="wikitable"
! Stage
! What happens
! Output
|-
| Connect
| TCP reachability probe, then SSH authentication
| reachable / unreachable
|-
| Collect
| run the battery of read-only checks
| raw measurements
|-
| Evaluate
| compare each measurement to thresholds
| findings (severity + detail + recommendation)
|-
| Score
| roll findings up into one verdict + health index
| HEALTHY / DEGRADED / CRITICAL
|}

The report is shown in the Portal's Diagnostic view — verdict, KPI cards, findings, a readable /var/log '''Log Hardware Analysis''' panel (scanned sources, per-bucket counts, matched raw lines, or a green all-clear note), and collapsible raw data. A '''⬇ Export HTML''' button downloads it as a self-contained .html file: the stylesheet is inlined and interactive controls are stripped, so it opens offline and can be shared or archived.

== Verdict Model ==

Every finding carries one of four severities:

{| class="wikitable"
! Severity
! Meaning
! Weight
|-
| critical
| Hardware failure or data-integrity risk; act now
| 25
|-
| warning
| Degrading or out-of-policy; schedule maintenance
| 8
|-
| [[optimization]]
| Works fine, but a known improvement is available
| 1
|-
| info
| Informational; no action required
| 0
|}

The overall verdict is the highest severity present: any '''critical''' → CRITICAL; else any '''warning''' → DEGRADED; else HEALTHY. The health index (0–100) shown in the report is computed as 100 minus 25 per critical, 8 per warning, and 1 per optimization, clamped to 0–100. A 100 means nothing was flagged. Optimization findings come in two kinds: tuning notes, which appear in the findings table, and '''Upgrade Opportunities''' (category Upgrade), which the Portal renders in a separate banner with an HPCMATE service offer and a [[contact]]-sales call-to-action — see Upgrade Opportunities below.

== Check Reference ==

This is the explanation of '''why''' the diagnostic judges each item the way it does. Each check below names its trigger condition, why it matters, the severity it can raise, and the common false positives. The Portal report links each finding here.

=== CPU Load ===

* Trigger: the 15-minute average load is compared against the core count.
* Why: sustained load above the core count means work is queued for longer than it can be serviced; above 80% of capacity it starts to hurt interactivity.
* Severity: critical if the 15-min load exceeds the core count; warning if it is above 80%; otherwise info.
* False positives: a burst from a parallel job (MPI/NCCL) can briefly push the 1-minute load high even when the 15-minute load is nominal — the check deliberately uses the 15-minute value to avoid reacting to short bursts.

=== Memory ===

* Trigger: available RAM (from '''free -m''') is compared to total.
* Why: low free memory means the kernel is close to OOM-killing workloads; high swap usage means RAM pressure is spilling to disk.
* Severity: critical if available is under 1 GB or under 10% of total; warning if under 20%; otherwise info. Swap over 50% used is a warning.
* False positives: on a machine running GPU inference, the page cache and buffer cache look like "used" memory but are reclaimable — the check uses the '''available''' column, not used, precisely so cached memory is not misread as pressure.

=== Swap ===

* Trigger: swap used over 50% of swap total.
* Why: sustained swapping degrades performance and is usually a sign the host is undersized for the [[workload]].
* Severity: warning.
* False positives: a host with a very large [[swap partition]] may use a little swap for cold pages without any real pressure — watch the trend, not a single sample.

=== Disk Space ===

* Trigger: usage percentage of the filesystems mounted at /, /home, /var, /opt, /data.
* Why: a near-full filesystem can stop services, fail job writes, or wedge logs.
* Severity: critical at 95% or more; warning at 85% or more.
* False positives: a tmpfs (tmp/ram) that is mostly in use is often intentional (scratch space) and is not flagged; only the listed real filesystems are scored.

=== Disk SMART ===

* Trigger: '''[[smartctl]] -H''' per physical disk, plus the Reallocated/Pending/Offline-Uncorrectable attribute counters.
* Why: SMART is the early-warning system for disk failure; uncorrectable errors mean data integrity is already at risk.
* Severity: critical if health is FAILED or there are any Offline_Uncorrectable sectors; warning for any Reallocated or Pending sectors; info on PASSED.
* False positives: '''smartctl missing or unable to read the disk reports UNKNOWN, not a failure''' — on hosts where smartmontools is not installed or the disk does not expose SMART, the check is simply not informative. A non-zero pending count can also be a transient that the disk corrects; monitor the trend before replacing.

=== GPU Temperature ===

* Trigger: per-GPU temperature from '''nvidia-smi'''.
* Why: sustained high temperature causes throttling and, over time, physical damage; it is a direct signal of inadequate cooling.
* Severity: critical above 90°C; warning above 80°C.
* False positives: a short heavy burst can push a card briefly into the warning band and settle back down; a single sample is a snapshot, so confirm it is sustained before treating it as a cooling fault. On a dev box running inference, an 80–90°C card under load is often expected, not a defect.

=== GPU ECC Errors ===

* Trigger: the lifetime corrected and uncorrectable ECC error counters from '''nvidia-smi'''.
* Why: uncorrectable ECC errors are a physical memory-cell failure — data the GPU computed may be wrong; a fast-growing corrected count is early degradation.
* Severity: critical for any uncorrectable error; warning above 1000 lifetime corrected.
* False positives: on consumer/prosumer cards that do not [[support]] ECC (for example many TITAN/GeForce-class GPUs), the counters read back as unavailable/zero — that is '''not a failure, it is simply not applicable''', and the check should be read as N/A on such cards.

=== GPU MIG ===

* Trigger: the per-GPU MIG mode ('''nvidia-smi --query-gpu=mig.mode.current''') plus the GPU-instance table ('''nvidia-smi mig -lgi''').
* Why: MIG is the datacenter multi-tenant primitive — it slices one GPU into isolated instances with dedicated memory and compute. Knowing whether a card is sliced tells you its intended tenancy model.
* Severity: info for MIG mode off (the default, whole-card use), MIG on with instances created, or MIG on but unsliced. '''Warning''' for a MIG-capable card that reports no MIG state at all (display mode).
* Display mode: a workstation Blackwell card (for example RTX PRO 4000/5000/6000) in '''display mode''' reports mig.mode of [N/A] — byte-for-byte identical to a genuinely non-MIG card. The check distinguishes the two by GPU model name: a known MIG-capable model reading all-N/A is flagged as a warning, because MIG profiles are not exposed until the card is switched to compute mode (displaymodeselector --gpumode compute --auto, then reboot).
* False positives: a truly non-MIG card (TITAN, GeForce) also reports [N/A], but it does not match the MIG-capable model list, so it produces '''no''' finding — the warning fires only when the model is known to support MIG but the mode is hidden behind display mode.
=== GPU Drop ===

* Trigger: the number of GPUs the driver reports ('''nvidia-smi''') is compared with the number physically present on the PCIe bus ('''lspci'''). A bus count greater than the driver count means a card is present but the driver failed to initialize it. The check also flags an adapter-init failure ("RmInitAdapter failed" in the kernel log) and an unreadable Video BIOS ("???") on a card that did initialize.
* Why: a GPU that is physically present but dropped by the driver is the classic "one of my GPUs vanished" symptom — usually a marginal card (silicon or power) that initializes only some of the time, most often after a cold boot. It also captures the driver-side cause: a hung [[Nvidia GPU Tips and Tricks#Xid 119|GSP firmware (Xid 119)]] or an adapter that fails to come up.
* Severity: critical for a count mismatch (a GPU is unavailable), an unreadable VBIOS, or an RmInitAdapter failure.
* Remediation (Portal): the report surfaces a "Reinstall latest driver" action for this check. Reinstalling the latest driver (Linux x86_64, proprietary mode) refreshes the GSP firmware and often resolves the drop. If the fault persists after a fresh driver, the card itself is the cause — reseat it, slot-swap it to separate a bad card from a bad board/slot, then RMA.
* False positives: none when every bus GPU initializes — a healthy multi-GPU host produces no finding. A card deliberately disabled (removed from use) can still be listed by lspci, so confirm intent before acting on a lone mismatch.
=== GPU XID Errors ===

* Trigger: NVRM "Xid" lines from the kernel log ('''dmesg''').
* Why: XID is NVIDIA's hardware fault code. Certain codes (48, 63, 64, 79) mean the GPU fell off the bus or hit a fatal fault; others (31, 43, 45, 68) are page-fault/ECC related.
* Severity: critical for the fatal set; warning for the page-fault set; otherwise info.
* [[Nvidia GPU Tips and Tricks#Xid 119|GSP (Xid 119)]]: Xid 119 is a GSP firmware timeout — a driver/firmware-level fault rather than a hardware one. The check escalates it to critical and recommends reinstalling the latest driver, which refreshes the GSP firmware. If Xid 119 recurs after a driver change, the fault is the card: reseat, slot-swap, then RMA.
* False positives: a Xid can be logged for a transient context error that does not recur. The check reports the recent set so an operator can judge whether it is recurring; a single benign Xid does not by itself mean the card must be replaced.

=== Network Links ===

* Trigger: the state of each network interface from '''ip -brief link''', excluding loopback and virtual ([[docker]]/veth/bridge) interfaces.
* Why: a physical NIC that is down means that path has no connectivity — a real problem for a server that is expected to be reachable on it.
* Severity: warning for each non-lo, non-virtual interface in the DOWN state.
* False positives: this is the check most often "wrong" for good reason. (a) A NIC that is '''intentionally not in use''' (a spare or a management port that is switched off) is legitimately down. (b) A '''USB NIC''' that is not plugged in shows as DOWN. (c) Some hosts run networking via a different stack (for example NetworkManager or a custom unit) where a legacy interface name is down by design. A DOWN flag here is a prompt to '''confirm whether that interface is supposed to be up''', not a confirmed fault.

=== NIC Speed ===

* Trigger: negotiated link speed from '''ethtool''' per interface.
* Why: a link that auto-negotiated below the port's rating (for example 1 GbE on a 10 GbE port, or a bad cable/SFP) silently caps throughput.
* Severity: warning if a real interface negotiates below 1000 Mbps.
* False positives: virtual/USB interfaces report "Unknown" or auto and are skipped; a 10 GbE link that negotiates 1 GbE may be an intentional patch for a [[test]] — confirm the intent.

=== Temperature Sensors ===

* Trigger: chassis/CPU temperature readings from '''sensors''' (lm-sensors).
* Why: a sensor reading in the dangerous band means the cooling loop (fans, pump, airflow) is failing to hold the package temperature.
* Severity: critical above 95°C; warning above 85°C.
* False positives: many bare-metal servers do not expose a usable sensor map to lm-sensors, in which case the check reports nothing rather than a fault. A high reading on a heavily loaded node is expected and should be read together with the GPU/CPU load.

=== Services ===

* Trigger: for each expected service (sshd, docker, NetworkManager/network), the diagnostic first checks the systemd unit state ('''systemctl is-active'''). If the unit is active the service is considered working. If the unit is '''not''' active, a '''function probe''' determines whether the service is actually serving: sshd → is port 22 listening?; docker → does '''docker info''' succeed?; network → is there a default route and a usable interface?
* Why: a service that is genuinely down is a common root cause of "the server does not respond the way it should". However, systemd unit state alone is not a reliable health signal.
* Severity: warning only when the unit is not active '''AND''' the function probe confirms the service is not actually serving. If the unit reads inactive but the function probe succeeds (port listening, default route present, docker responding), the service is considered working and no finding is raised.
* False positives: this is the check most often "wrong" for good reason. (a) On Ubuntu 24.04, sshd and NetworkManager are commonly '''socket-activated''' — the long-running service unit reads as inactive while the socket accepts connections and the service is fully functional. (b) On dev/test hosts, sshd may run as a '''non-systemd daemon''' (a service script or a container) — both sshd.service and sshd.socket read inactive while port 22 is serving and you are literally connected over SSH. (c) The legacy '''network''' unit is inactive on hosts managed by NetworkManager. In all these cases the function probe confirms the service is working, so no warning is raised. A finding is only produced when the service is genuinely down (unit inactive and function probe fails).

=== Kernel Errors ===

* Trigger: the most recent '''dmesg''' messages at error/critical/alert level.
* Why: the kernel log is the last word on hardware faults — machine-check exceptions (CPU/RAM), disk I/O errors, [[PCIe]] AER errors, and OOM kills all surface here.
* Severity: critical for machine-check and disk I/O errors; warning for PCIe AER and OOM events; otherwise info.
* False positives: a few stale error lines from a past event can linger in the ring buffer. The check reads the most recent window so old, already-resolved events do not keep flagging a healthy host; if a recurring pattern appears, treat it as real.

=== Log Hardware ===

* Trigger: a bounded scan of the persisted logs for hardware-failure signatures, bucketed into machine-check, disk I/O, NVMe, PCIe AER, memory/EDAC, GPU Xid, and NIC link-down. The '''primary''' sources are the /var/log files that survive a reboot (syslog, the kernel log kern.log, messages); the previous-boot journal ('''journalctl -b -1''') is appended as a '''complement''' when present so journald-only hosts are covered too. Matched lines are de-duplicated first — a host running rsyslog AND persistent journald writes the same kernel line to both, which would otherwise double-count.
* Why: dmesg and the current journal are volatile and gone after a reboot, but a persisted journal or syslog keeps the failures that PRECEDED the reboot — exactly the ones the customer is already living with. This complements the Kernel Errors check (current dmesg) by covering the last boot, and it names the component to replace.
* Severity: critical for machine-check events or memory/ECC faults (CPU/RAM failure) and for sustained disk I/O errors; warning for a recurring NVMe, PCIe-AER, NIC-link, or GPU-Xid pattern.
* Fault word required: every bucket demands an explicit error keyword so benign lines never count — for example "MCE: In-kernel MCE decoding enabled" (a benign boot line) and EDAC initialization lines (channel count, "Giving out device", chip selects) are NOT counted, and "Shutdown timeout" is not an NVMe fault; the ACPI line "_OSC: platform does not support [AER LTR DPC]" (a platform-limit note, not an error) is excluded from the PCIe-AER count.
* False positives: the patterns are deliberately conservative (verified to yield zero hits on a healthy box). If no persisted log source is found at all (journal not persisted and no syslog file), an info note suggests enabling persistent journald so pre-reboot faults can be reviewed remotely.
=== NUMA ===

* Trigger: the socket/node count from '''numactl --hardware'''.
* Why: multi-socket hosts get a large latency penalty when a process is scheduled on a socket different from the memory it touches; being [[NUMA]]-aware removes it.
* Severity: optimization (informational) when more than one NUMA node is present.
* False positives: none by design — a single-socket host simply produces no finding. On a multi-socket HPC node this is a genuine tuning opportunity, not a defect.

=== SSD Wear ===

* Trigger: for NVMe, the SMART/Health Log via '''sudo smartctl -a''' (Percentage Used, Data Units Written, Available Spare, Media/Data Integrity Errors, temperature); for SATA SSDs, the wear attributes.
* Why: every SSD has a finite write budget (TBW/dWPD). "Percentage Used" is the vendor's normalized wear gauge; crossing it means the remaining life can no longer guarantee the rated endurance. Data integrity errors are a direct signal of media failure.
* Severity: critical at 80% or more wear, or any media/integrity error; warning at 60% or more; otherwise info (reports wear, total written, spare capacity, and temperature).
* False positives: on a lightly used server a 2% gauge with low total-bytes-written is normal and healthy. The gauge is not "health" — a drive at 95% wear can still pass a short SMART test, which is exactly why the write-based gauge matters more than the pass/fail flag.

=== Network Health ===

* Trigger: per-interface error and drop counters from '''/proc/net/dev''' (receive/ transmit errors and drops), excluding loopback and virtual interfaces.
* Why: physical-layer problems (bad cable, dirty/failing SFP, a dying NIC) show up as a growing error/drop counter long before the link goes down. Catching this is what turns a "mystery slowness" into a replace-this-part answer.
* Severity: critical for sustained high error counts (thousands or more); warning for any non-zero errors or a high drop count; otherwise no finding.
* False positives: a small, stable drop count that does not grow is usually harmless (a brief boot-time burst). The counters are cumulative since boot, so the meaningful signal is whether they keep climbing — compare against a previous run before replacing hardware. Virtual interfaces (tap/tun/veth) are excluded because their counters carry different meanings.

=== Syslog Patterns ===

* Trigger: the last week of error-priority journal ('''journalctl --priority=err'''), categorized into patterns: machine-check, disk I/O, PCIe AER, OOM, GPU/NVIDIA, USB, NIC link, NVMe.
* Why: hardware faults repeat and leave a signature. A cluster of machine-check events points at CPU/RAM; I/O errors point at a specific disk; PCIe AER at a GPU or riser; OOM at memory pressure. Grouping the log tells you '''which component''' to replace, not just that something is wrong.
* Severity: critical for multiple machine-check or disk I/O events; warning for a recurring PCIe-AER, GPU, NIC, or OOM pattern; info only for high overall log volume with no dominant pattern.
* False positives: a large volume of unrelated error-level messages (for example repetitive protocol chatter) raises the "log volume" info note without being a fault. The check deliberately escalates only on a recognizable hardware pattern, not on raw message count.

=== Storage Space ===

* Trigger: '''df -h''' over all real filesystems (NFS included) plus an inode check ('''df -i''').
* Why: a full filesystem stops writes and can take services down; a full inode table stops file creation even when space remains. Both are capacity-planning signals that lead directly to an expansion or a cleanup.
* Severity: critical at 95% capacity; warning at 85% or at 90% of inodes; info at 70% (early capacity planning).
* False positives: a near-full tmpfs or a deliberately small dedicated partition can trip the capacity thresholds without being a real risk — the check scores the major real filesystems and reports the mount so the operator can judge intent.

=== Software Versions ===

A read-only snapshot of the software running on the box. It never changes anything — it only records versions so an operator can see what is installed and decide what to update or re-flash.

{| class="wikitable"
!! Item
!! What it reports
!! Typical value
|-
| OS
| Distribution and release, as the OS identifies itself
| Rocky [[Linux]] 9.4
|-
| Kernel
| The running kernel version
| 5.14.0-427.el9.x86_64
|-
| Uptime
| How long the host has been up
| up 3 days, 4 hours
|-
| GPU driver
| Installed [[NVIDIA driver]] version (first GPU)
| 550.54.15
|-
| NVMe firmware
| Firmware revision of each NVMe drive
| 0x114
|-
| Docker
| Container engine version
| 26.1.4
|-
| Python
| Python 3 interpreter version
| 3.11.2
|-
| Node.js
| Node.js runtime version
| v22.0.0
|-
| CUDA
| [[CUDA Toolkit|CUDA toolkit]] release
| 12.2
|-
| GCC
| Compiler version
| 12.3.0
|-
| smartctl
| Smartmontools version (disk health tooling)
| 7.3
|}

Notes:

* Severity: info — a reference snapshot, not a fault. It never changes the health verdict.
* Conditional fields: GPU driver, NVMe firmware, and the package rows appear only when that software is actually present on the host; a minimal box shows just OS, kernel, and uptime.
* In the Portal: the finding shows the versions as a compact key=value list (first eight); the full list is under the 'Raw diagnostic data' expander.
* False positives: an old-but-stable driver or firmware is not broken — cross-reference the recorded versions against vendor advisories before acting. This is a planning aid, not a pass/fail check.

=== Hardware Identity ===

* Trigger: system / baseboard / [[BIOS]] identity from '''dmidecode''' (with a '''/sys/class/dmi/id''' fallback when root is unavailable), BMC info from '''sudo ipmitool mc info''' ([[IPMI]]), full FRU inventory from '''sudo ipmitool fru''' (Board Mfg/Serial, Product Mfg/Serial, chassis), and every NIC MAC address (loopback excluded).
* Why: the board model, [[BIOS]] revision, BMC firmware, FRU serials, and NIC MACs are the asset identity used for tracking, [[warranty]]/RMA matching, and firmware upgrade planning. Recording them in the report turns the box into a known asset with a concrete upgrade target.
* Severity: info (a reference snapshot, never a fault).
* Summary line: the finding detail is a single pipe-joined identity, for example '''Supermicro H12SSL-NT v1.01 | BIOS 2.3 | BMC 1.00 | SN 0123456789 | 3 NICs''' — vendor + board model (+ board version), [[BIOS]] version, BMC firmware, the first available serial (DMID system serial, falling back to FRU product/board serial), and the count of '''physical''' NICs.
* Physical NIC count: virtual interfaces (docker bridge, veth pairs, tap/tun, kube/cni/calico/flannel, wireguard) are excluded from the NIC count so a Docker host does not read as "17 NICs" — only real onboard/attached NICs are counted.
* Privilege: the BMC and FRU reads require root (local-bus [[IPMI]] is root-only), so the check runs them with '''sudo'''. On a host where the SSH user has no passwordless sudo (factory provisioning missing), those fields are simply absent — the check degrades to dmidecode/sysfs identity rather than faulting.
* False positives: none by design. If '''dmidecode''' cannot be run, the check falls back to '''/sys/class/dmi/id''' fields; if there is no BMC (consumer/workstation board) the BMC/FRU fields are empty and skipped.

=== Power Health ===

* Trigger: [[IPMI]] '''chassis status''' (power overload / main power fault / cooling-fan fault / drive fault), voltage sensors from the SDR, PSU sensors (used to detect PMBus availability), and power-related events from the SEL.
* Why: a power-subsystem fault (overload, main power fault, a failing cooling loop) is an immediate physical risk, and the SEL is the historical record of power events — together they name whether the power delivery is healthy.
* Severity: '''critical''' for any chassis power fault; '''warning''' for an unexpected system power state or for power-related SEL events; '''info''' for voltage / PSU telemetry.
* PMBus limitation: many boards (for example the MSI D3060GB2N) do not wire a PMBus/IPMB PSU telemetry bus, so the PSU sensors read "No Reading". The check reports this as '''info, not a fault''' — power is still verified through the voltage sensors and the chassis status. For long-term PSU wear monitoring, use a board whose PSU exposes PMBus.
* False positives: "0 PSU No Reading (no PMBus)" is a board limitation and is never escalated to a fault. The critical path requires the BMC to explicitly report a power fault flag set to true.
=== Upgrade Opportunities ===

* Trigger: a separate layer of checks that fires when a component has headroom a customer could buy more of — not when something is broken. Eight checks: storage (any real mount, LVM volume, or ZFS pool at 80% or more), RAM (80% or more in use), GPU VRAM (80% or more on any card) or MIG disabled on a MIG-capable datacenter GPU, NIC (fastest physical link at 10 GbE or below), CPU (16 cores or fewer, or a previous-generation part), cooling (sustained GPU temperature 75 degrees C or above), PSU (a single unit with no N+1 redundancy), and software (a bare OS with no GPU driver, [[CUDA]], or Python environment).
* Why: these are revenue-relevant. The Portal surfaces them to the customer as an "Upgrade Opportunities" banner — a card per item, each carrying an HPCMATE service offer and a contact-sales call-to-action — so the same diagnostic that reassures a healthy machine also points the customer at the next thing worth buying.
* Severity: all eight are '''optimization''' (weight 1) and are tagged '''category Upgrade'''. They never raise the overall verdict and are excluded from the fault table; they render only in their own banner above the findings.
* False positives: the GPU MIG check is gated on a model-name allowlist (A100/A30/A40, H100/H200/H800, L40/L40S, B100/B200 and similar) because consumer cards (for example the TITAN RTX) have no MIG capability and must never be offered an upgrade they cannot take. The CPU check uses a conservative generation heuristic, so a well-chosen modern low-core part is the main case it can over-flag. A 1 GbE NIC on a deliberately single-tenant inference box is a legitimate configuration even though the check still surfaces the 25/100 GbE option.

== Limitations and Common False Positives ==

The diagnostic is a snapshot, not a continuous monitor. Read a single run that way: a transient (a load burst, a one-off Xid, a momentarily hot GPU) can show up even on a healthy machine. The checks are tuned to reduce the noisiest false positives: memory uses the '''available''' column (not "used") so page cache is not misread as pressure; CPU load uses the 15-minute average (not 1-minute) so parallel-job bursts do not trigger; missing sensors or tools are reported as "not applicable" rather than "failed"; the Services check uses a '''function probe''' (port listening, default route, docker responding) so socket-activated or non-systemd services are not flagged as down; and the Services and Network checks together report a network condition only once. The remaining areas that most benefit from a human look are the '''GPU ECC''' check (N/A on cards without ECC support) and the '''SSD Wear''' gauge (a low gauge on a lightly used drive is healthy, not a fault). When a finding is not obvious, open the Why? link for that check and confirm with the suggested command before treating the host as faulty.

== References ==

* HPCMATE Portal — Diagnostic view (POST/GET /api/servers/:id/diagnostic)
* diagnose_server.py — the read-only check implementation
* NVIDIA nvidia-smi documentation (temperature, ECC, XID)
* smartmontools (smartctl) documentation
* lm-sensors documentation

== Related Pages ==

* [[Server]]
* [[System Update and Optimization]]
* [[Pre-Ship Validation]]

[[Category:Server]]
[[Category:Reference]]

Nvidia GPU Tips and Tricks

2026-10-14T05:50:17Z

Admin: /* Detailed Explanation */

== Overview ==

Nvidia GPU Tips and Tricks에 대한 기술 문서입니다.

=== Summary ===

* 무엇인가? - Nvidia GPU Tips and Tricks
* 왜 필요한가? - HPC 및 서버 환경에서 필수 개념
* 언제 사용하는가? - 서버 구성, 성능 튜닝, 문제 해결 시

== Purpose ==

이 문서가 존재하는 이유

* Goal: Nvidia GPU Tips and Tricks에 대한 기술 정보 제공
* Scope: Nvidia GPU Tips and Tricks의 개념, 사용법, 설정
* Non-goals: 다른 주제로의 확장

== Key Concepts ==

{| class="wikitable"
! Concept
! Description
! Related
|-
| Nvidia GPU Tips and Tricks
| HPC/서버 환경에서 중요한 기술 개념
| [[Linux]], [[Server]]
|}

== Xid 119 ==
Based on many cases, XID 119 error caused by many reasons. one of that is the incompatibility between Nvidia GSP firmware and [[PCIe ASPM (Active State Power Management)]], especially if you can see GPU device on PCI list ($lspci | grep -i vga) while nvidia-msi does not show the GPU.

From [[ASPM]] perspective, host driver is trying to wakeup GPU in deep-P-state, GPU's GSP response has been timeout then host disconnected [[PCIe]] bus slot to disconnect the GPU (Fallen off the bus / Xid 79). simply idle deep-P-state GSP disconnection.

<syntaxhighlight lang="bash">
#Xid 119 error example
$ sudo cat /var/log/syslog | grep -i NVRM
Oct 12 00:32:17 Venus kernel: [ 1074.666075] NVRM: GPU at PCI:0000:81:00: GPU-760dab3f-a851-fc4c-dba8-c400d7db4b17
Oct 12 00:32:17 Venus kernel: [ 1074.666086] NVRM: GPU Board Serial Number: 1714923003394
Oct 12 00:32:17 Venus kernel: [ 1074.666089] NVRM: Xid (PCI:0000:81:00): 119, pid=2816, name=nvidia-smi, Timeout after 6s of waiting for RPC response from GPU1 GSP! Expected function 103 (GSP_RM_ALLOC) (0x0 0x78).
Oct 12 00:32:17 Venus kernel: [ 1074.666114] NVRM: GPU1 GSP RPC buffer contains function 103 (GSP_RM_ALLOC) and data 0x0000000000000000 0x0000000000000078.
Oct 12 00:32:17 Venus kernel: [ 1074.666122] NVRM: GPU1 RPC history (CPU -> GSP):
Oct 12 00:32:17 Venus kernel: [ 1074.666126] NVRM: entry function data0 data1 ts_start ts_end duration actively_polling
Oct 12 00:32:17 Venus kernel: [ 1074.666130] NVRM: 0 103 GSP_RM_ALLOC 0x0000000000000000 0x0000000000000078 0x00065d99d4274f2e 0x0000000000000000 y
Oct 12 00:32:17 Venus kernel: [ 1074.666141] NVRM: -1 76 GSP_RM_CONTROL 0x0000000020800aff 0x0000000000000010 0x00065d999699a9c2 0x00065d999699ad04 834us
Oct 12 00:32:17 Venus kernel: [ 1074.666152] NVRM: -2 103 GSP_RM_ALLOC 0x0000000000002080 0x0000000000000004 0x00065d9996999382 0x00065d999699946b 233us
Oct 12 00:32:17 Venus kernel: [ 1074.666161] NVRM: -3 103 GSP_RM_ALLOC 0x0000000000000080 0x0000000000000038 0x00065d99969991fa 0x00065d999699929e 164us
Oct 12 00:32:17 Venus kernel: [ 1074.666169] NVRM: -4 103 GSP_RM_ALLOC 0x0000000000000000 0x0000000000000078 0x00065d999699910f 0x00065d99969991f6 231us
Oct 12 00:32:17 Venus kernel: [ 1074.666177] NVRM: -5 76 GSP_RM_CONTROL 0x000000000073028b 0x0000000000000038 0x00065d9996951efc 0x00065d9996951f73 119us
Oct 12 00:32:17 Venus kernel: [ 1074.666185] NVRM: -6 76 GSP_RM_CONTROL 0x0000000000730211 0x0000000000000010 0x00065d9996951e4d 0x00065d9996951ef3 166us
Oct 12 00:32:17 Venus kernel: [ 1074.666194] NVRM: -7 76 GSP_RM_CONTROL 0x000000000073028b 0x0000000000000038 0x00065d9996951d96 0x00065d9996951e41 171us
Oct 12 00:32:17 Venus kernel: [ 1074.666202] NVRM: GPU1 RPC event history (CPU <- GSP):
Oct 12 00:32:17 Venus kernel: [ 1074.666206] NVRM: entry function data0 data1 ts_start ts_end duration during_incomplete_rpc
Oct 12 00:32:17 Venus kernel: [ 1074.666211] NVRM: 0 4108 UCODE_LIBOS_PRINT 0x0000000000000000 0x0000000000000000 0x00065d99968fe0bd 0x00065d99968fe0bd
Oct 12 00:32:17 Venus kernel: [ 1074.666220] NVRM: -1 4108 UCODE_LIBOS_PRINT 0x0000000000000000 0x0000000000000000 0x00065d99968fdfc9 0x00065d99968fdfcb 2us
Oct 12 00:32:17 Venus kernel: [ 1074.666229] NVRM: -2 4098 GSP_RUN_CPU_SEQUENCER 0x000000000000061c 0x0000000000003fe2 0x00065d99968ec520 0x00065d99968ed9b7 5271us
Oct 12 00:32:23 Venus kernel: [ 1080.710822] NVRM: Xid (PCI:0000:81:00): 119, pid=2816, name=nvidia-smi, Timeout after 6s of waiting for RPC response from GPU1 GSP! Expected function 103 (GSP_RM_ALLOC) (0x80 0x38).
Oct 12 00:32:29 Venus kernel: [ 1086.715510] NVRM: Xid (PCI:0000:81:00): 119, pid=2816, name=nvidia-smi, Timeout after 6s of waiting for RPC response from GPU1 GSP! Expected function 103 (GSP_RM_ALLOC) (0x2080 0x4).
Oct 12 00:32:35 Venus kernel: [ 1092.732051] NVRM: Rate limiting GSP RPC error prints for GPU at PCI:0000:81:00 (printing 1 of every 30). The GPU likely needs to be reset.
Oct 12 00:39:36 Venus kernel: [ 1514.082939] NVRM: Xid (PCI:0000:81:00): 119, pid=3611, name=nvidia-smi, Timeout after 6s of waiting for RPC response from GPU1 GSP! Expected function 103 (GSP_RM_ALLOC) (0x80 0x38).
Oct 12 00:48:55 Venus kernel: [ 2073.280826] NVRM: Xid (PCI:0000:81:00): 119, pid=4684, name=nvidia-smi, Timeout after 6s of waiting for RPC response from GPU1 GSP! Expected function 76 (GSP_RM_CONTROL) (0x20801819 0x20).
Oct 12 01:09:34 Venus kernel: [ 10.489443] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 570.169 Thu Jun 12 20:04:34 UTC 2025
Oct 12 01:09:44 Venus kernel: [ 24.019501] NVRM: GPU 0000:81:00.0: RmInitAdapter failed! (0x25:0xffff:1545)
Oct 12 01:09:44 Venus kernel: [ 24.019545] NVRM: GPU 0000:81:00.0: rm_init_adapter failed, device minor number 0
Oct 12 01:09:58 Venus kernel: [ 37.212612] NVRM: GPU 0000:81:00.0: RmInitAdapter failed! (0x25:0xffff:1545)
Oct 12 01:09:58 Venus kernel: [ 37.212667] NVRM: GPU 0000:81:00.0: rm_init_adapter failed, device minor number 0
Oct 12 01:10:05 Venus kernel: [ 45.089470] NVRM: GPU 0000:81:00.0: RmInitAdapter failed! (0x25:0xffff:1545)
Oct 12 01:10:05 Venus kernel: [ 45.089530] NVRM: GPU 0000:81:00.0: rm_init_adapter failed, device minor number 0
Oct 12 01:29:15 Venus kernel: [ 9.350719] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 570.169 Thu Jun 12 20:04:34 UTC 2025
Oct 12 01:40:32 Venus kernel: [ 8.597394] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 570.169 Thu Jun 12 20:04:34 UTC 2025
Oct 12 01:40:41 Venus kernel: [ 20.848347] NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x25:0xffff:1545)
Oct 12 01:40:41 Venus kernel: [ 20.848419] NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0
Oct 12 01:41:36 Venus kernel: [ 75.328697] NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x25:0xffff:1545)
Oct 12 01:41:36 Venus kernel: [ 75.328742] NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0

</syntaxhighlight>

The solution would be deactivate GSP - ASPM signaling using one of following way

1. Set GPU persistence mode (Most simple and recommended) along with custom systemd service to PM on after system bootup.

<nowiki>#</nowiki>sudo nvidia-smi -pm 1

2. Disable PCIe ASPM with [[kernel]] param

Add pcie-aspm=off in GRUB_CMDLINE_LINUX_DEFAULT then update-grub and reboot

3. Disable GSP feature (depending on version of driver, proprieatry vs open-kernel driver or GPU generation)

Create /etc/modprobe.d/gsp.conf 파일에 options nvidia NVreg_EnableGpuFirmware=0

== Best Practices ==

* 최신 버전 사용 권장
* 공식 문서 참고
* 테스트 환경에서 먼저 검증

== References ==

* [https://wiki.hpcmate.com Nvidia GPU Tips and Tricks]

== Related Pages ==

* [[Linux]]
* [[Server]]
* [[Hardware]]
* [[Network]]
* [[NVIDIA GPUs]]

[[Category:GPU]]
== Knowledge Graph ==

Related

→ [[CUDA]]
→ [[GPU]]
→ [[NVIDIA GPU]]
→ [[H100]]
→ [[NVLink]]
→ [[AMD GPUs]]
→ [[A100]]
→ [[MIG]]

[[Category:Guide]]

AIS (AI Station)

2026-10-02T04:20:00Z

Admin: Created page with "== Overview == == References == <references />"

== Overview ==

== References ==
<references />

AI Agents

2026-10-02T04:19:05Z

Admin:

== Overview ==

AI Agent 는 인공지능을 활용하여 환경을 인지하고, 판단하며, 자율적으로 행동을 수행하는 소프트웨어 시스템입니다.

=== Summary ===

* 무엇인가? AI 모델을 핵심 지능으로 활용하며 메모리, 도구, 의사결정 로직을 패키징한 자율 소프트웨어 시스템
* 왜 필요한가? 인간 개입 없이 사전에 계획된 작업을 자동 실행하여 운영 효율성 향상
* 언제 사용하는가? 디지털 워크어시스턴트, 스마트 홈 자동화, 반복적 IT 운영 작업

== Purpose ==

이 문서가 존재하는 이유

* Goal: AI Agent의 정의, AI Model과의 차이, 활용 사례를 명확히 설명
* Scope: AI Agent 개념, 주요 예시(Hermes, Claude Code, Codex 등)
* Non-goals: 특정 AI Agent 구현 방법, 프롬프트 엔지니어링 상세

== Key Concepts ==

{| class="wikitable"
! Concept
! Description
! Related
|-
| AI Agent
| 자율적으로 환경을 인지하고 판단하여 행동을 수행하는 소프트웨어 시스템
| [[AI Model]]
|-
| AI Model
| 텍스트 생성, 문맥 이해, 질문 답변에 특화된 핵심 지능 엔진
| [[AI Agent]]
|-
| Digital Work Assistant
| AI Agent의 대표적 활용 사례 — 문서 처리, 일정 관리, 정보 검색 자동화
| [[AI Agent]]
|-
| Smart Home & Automation
| AI Agent의 또 다른 활용 사례 — 가전 제어, 환경 모니터링, 예측적 자동화
| [[AI Agent]]
|}

== Architecture ==

AI Agent는 다음 구성 요소로 이루어집니다:

<syntaxhighlight lang="mermaid">
graph TD
A[AI Agent] --> B[AI Model - 핵심 지능]
A --> C[메모리 - 과거 상호작용 저장]
A --> D[도구 - 외부 시스템 연동]
A --> E[의사결정 로직 - 목표 기반 행동 선택]
B --> F[Qwen, GPT, Gemini 등]
C --> G[대화 기록, 사용자 선호도]
D --> H[API 호출, 파일 시스템, 웹 검색]
E --> I[목표 달성 위한 단계별 계획]
</syntaxhighlight>

== Workflow ==

AI Agent의 일반적인 작동 흐름:

# 환경 인지 — 사용자 입력 또는 센서 데이터 수집
# 판단 — AI 모델을 통해 상황 분석 및 옵션 생성
# 의사결정 — 메모리, 도구 가용성, 목표에 기반하여 최적 행동 선택
# 실행 — 선택된 행동을 외부 시스템에 적용
# 학습 — 결과 피드백을 메모리에 반영하여 향후 성능 향상

{| class="wikitable"
! Stage
! Input
! Output
|-
| 인지
| 사용자 요청, 센서 데이터, 시스템 상태
| 상황 분석 결과
|-
| 판단
| 분석 결과, AI 모델
| 가능한 행동 옵션
|-
| 의사결정
| 옵션, 메모리, 도구 가용성
| 선택된 행동
|-
| 실행
| 선택된 행동, 외부 시스템
| 실행 결과
|-
| 학습
| 실행 결과, 목표 달성도
| 메모리 업데이트
|}

== Detailed Explanation ==

=== AI Model vs. AI Agent ===

* AI Models (Qwen, GPT, Gemini 등)는 "두뇌" 역할을 합니다. 텍스트 생성, 문맥 이해, 질문 답변에 탁월하지만, 직접적인 단계별 인간 프롬프트가 필요합니다.
* AI Agents는 전체 소프트웨어 시스템입니다. AI 모델을 핵심 지능으로 활용하되, 메모리, 도구, 의사결정 로직을 패키징하여 실제로 작업을 수행합니다.

=== AI Agent 활용 사례 ===

* Digital Work Assistants — 문서 처리, 이메일 관리, 일정 조정, 정보 검색 자동화
* Smart Home & Automation — 가전 제어, 환경 모니터링, 예측적 자동화
* IT Operations — 서버 모니터링, 로그 분석, 자동 복구, 성능 최적화

== Best Practices ==

* AI Agent 설계 시 명확한 목표와 범위 정의
* 메모리 관리 전략 수립 — 중요 정보 보존, 불필요한 데이터 정리
* 도구 선택 시 보안과 가용성 고려
* 인간 감독(Human-in-the-loop) 메커니즘 구축 — 중요한 결정은 인간 확인
* 피드백 루프 구축 — 실행 결과 기반 지속적 개선

== Limitations ==

* AI 모델의 정확도에 의존 — 잘못된 판단 시 잘못된 행동 실행 가능
* 메모리 제한 — 장기 기억 관리 어려움
* 도구 가용성 — 사용 가능한 도구에 따라 기능 제한
* 보안 리스크 — 자율적 행동으로 인한 예상치 못한 결과 발생 가능
* 윤리적 고려사항 — 자동화 결정의 투명성과 책임 소재 명확화 필요

== References ==

* Hermes — Nous Research의 AI Agent 프레임워크
* Claude Code — Anthropic의 코딩 에이전트
* Codex — OpenAI의 코딩 에이전트
* Opencraw — 오픈소스 코딩 에이전트

== Related Pages ==

* [[AI Coding Tool]]
* [[AIS (AI Station)]]
* [[AI Model]]
* [[Digital Work Assistant]]
* [[Smart Home Automation]]

[[Category:AI]]
== Knowledge Graph ==

Related

→ [[Linux]]
→ [[Server]]
→ [[Hardware]]
→ [[Network]]

[[Category:Reference]]

U4008

2026-10-02T04:04:27Z

Admin:

== Overview ==

U4008은 HPCMATE 가 독자 설계한 다재 다능한 시스템을 지원할 수 있는 공냉식/수냉식 서버 시스템 입니다.

== Key Concepts ==

{| class="wikitable"
! Concept
! Description
! Related
|-
| U4008
| HPC/서버 환경에서 중요한 기술 개념
| [[Linux]], [[Server]]
|}

== U4008 Series ==

U4008 has been designed to provide a flexible system configuration high performance computer server for AI/ML/LLM.
[[File:U4008.png|center|frameless|446x446px]]
In brief : U4008 has been designed to provide a flexible system configuration high performance computer server for AI/ML/LLM

== Features ==

* 4U rack mount server, (W441mm x H177mm x D866mm, D930mm with external Fan)
* Single Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor, TDP up to 350W
* Default support 4 x double width 600W GPUs (Optional configuration +1 x double width 600W GPU, Total 5 x 600W double width GPU** )
* 8 x DDR5 DIMM Slots, 8 channels(1DPC), RDIMM/RIMM-3DS/MCRDIMM (RDIMM: 6400MT/s, MCRDIMM: 8000MT/s)
* 4 x U.2/U.3 hot-swap front drive (Optional configuration +8 x U.2/U.3 or +8 x NVMe or +8 x SATA SSD **)
* NVMe RAID 0/1/5/10 support (Intel® VROC RAID key required)
* 2 x 1000 Mbps RJ45, 1 x 1000 Mbps RJ45 Dedicated IPMI 2.0 LAN
* Button (1 x Power, 1 x Reset), LED(1 x Power, 1 x HDD Activity, 2 x Network Activity)
* 2 x USB 3.2 Gen 1 Type-A ports, 1 x VGA D-Sub port, 1 x COM port, 1 x UID LED button
* 2 x 2200W Platinum PSU (automatic mode : 1+1 2200W or 2+0 4400W depending on the workload)
* Introduction
** <nowiki>http://www.hpcmate.com/src/video/U4008.mp4</nowiki>
** <nowiki>http://www.hpcmate.com/src/video/U4008_PSU_Options.mp4</nowiki>

== Technical specification ==
{| class="wikitable"
|+
!
!Default
!Max Configuration
!Options
|-
|Power
|2000W Platinum Modular PSU
|Configurable up to 5600W
|Dual 2200W '''Titanium''' CRPS
Automatic switching between 1+1 and 2+0 based on [[workload]]
|-
|Storage
|3 x 5.15" dummy cage
|Any type of 5.25" hard rack
|[[Category:Server]]
Depending on models
[[Category:Reference]]
|-
|Dimension
| colspan="3" |4U rack mount server, (441mm x 177mm x 866mm)
|-
|Cooling
|Air Cooling
Cooling 4 x middle / 2 x rear fan (Max. 4,500 RPM/168.2CFM/54.5dbA), 1 x rear fan (Max. 4800 RPM/60.8 CFM/44.5dbA) / up to 500W active CPU cooler (Adjustable noise / cooling ratio)
|
|Optional [[Liquid cooling|Liquid Cooling]] with [[CoolUX]]
|-
|Front I/O
| colspan="3" |Button(1 x Power, 1 x Reset), LED(1 x Power, 1 x HDD Activity, 2 x [[Network]] Activity)
|-
|SKU
| colspan="2" |HM-U48-22K5G ( Air cooling)
|Configuration
|-
|
|CPU
|Single Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor, TDP up to 350W
|
|-
|
|Memory
|8 x DDR5 DIMM Slots, 8 channels(1DPC), RDIMM/RIMM-3DS/MCRDIMM (RDIMM: 6400MT/s, MCRDIMM: 8000MT/s)
|
|-
|
|Storage
|4 x U.2/U.3 hot-swap front drive (Optional configuration +8 x U.2/U.3 or +8 x NVMe or +8 x SATA SSD **)
|
|-
|
|I/O
|<nowiki>- 2 x 2280/22110 PCIe 5.0 x2 NVMe M.2 ports, 3 x SlimlineSAS 4i for 12 SATA3.0 drives (NVMe RAID 0/1/5/10 support)</nowiki>

- 4 x [[PCIe]] 5.0 x16 slots (one slot only functional with R1S CPU), 6 x PCIe 5.0 x8 MCIO

- 2x USB 3.2 Type A, 1x DB9, 1 COM port, UID LED button
|
|-
|
|PCIe
|11 x Full-Height/Full-Length expansion slots, 7 x PCIe 5.0 x16 (supports 4 x CXL2.0* x16), 1 x PCIe5.0 x8
|
|-
|
|GPU
|Default [[support]] 4 x double width 600W GPUs ( Optional configuration +1 x double width 600W GPU** )
|5 x double width 600W GPUs on PCIe Gen5 x16
|-
|
|Network
|2 x 1000 Mbps RJ45, 1 x 1000 Mbps RJ45 Dedicated [[IPMI]] LAN
|
|-
|
|VGA
|VGA 1 x D-Sub(AST2600 )
|
|-
|
|Software
| colspan="2" |HPCMATE system level optimized Linux based Ubuntu LTS (Default)
Optional Windows/Windows Server or other Linux distribution

Optional HPCMATE DLS (Deep Learning Station) Framework.

Optional HPCMATE SmartLLM (HMSmartLLM)Framework.

Various optional on-demand custom software solutions
|}

== Contact US ==
Mail us to sales@hpcmate.com

U4008

2026-10-02T04:03:33Z

Admin:

== Overview ==

U4008은 HPCMATE 가 독자개발한 다양한 기능을 내포할 수 있는 시스템 입니다.

== Key Concepts ==

{| class="wikitable"
! Concept
! Description
! Related
|-
| U4008
| HPC/서버 환경에서 중요한 기술 개념
| [[Linux]], [[Server]]
|}

== U4008 Series ==

U4008 has been designed to provide a flexible system configuration high performance computer server for AI/ML/LLM.
[[File:U4008.png|center|frameless|446x446px]]
In brief : U4008 has been designed to provide a flexible system configuration high performance computer server for AI/ML/LLM

== Features ==

* 4U rack mount server, (W441mm x H177mm x D866mm, D930mm with external Fan)
* Single Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor, TDP up to 350W
* Default support 4 x double width 600W GPUs (Optional configuration +1 x double width 600W GPU, Total 5 x 600W double width GPU** )
* 8 x DDR5 DIMM Slots, 8 channels(1DPC), RDIMM/RIMM-3DS/MCRDIMM (RDIMM: 6400MT/s, MCRDIMM: 8000MT/s)
* 4 x U.2/U.3 hot-swap front drive (Optional configuration +8 x U.2/U.3 or +8 x NVMe or +8 x SATA SSD **)
* NVMe RAID 0/1/5/10 support (Intel® VROC RAID key required)
* 2 x 1000 Mbps RJ45, 1 x 1000 Mbps RJ45 Dedicated IPMI 2.0 LAN
* Button (1 x Power, 1 x Reset), LED(1 x Power, 1 x HDD Activity, 2 x Network Activity)
* 2 x USB 3.2 Gen 1 Type-A ports, 1 x VGA D-Sub port, 1 x COM port, 1 x UID LED button
* 2 x 2200W Platinum PSU (automatic mode : 1+1 2200W or 2+0 4400W depending on the workload)
* Introduction
** <nowiki>http://www.hpcmate.com/src/video/U4008.mp4</nowiki>
** <nowiki>http://www.hpcmate.com/src/video/U4008_PSU_Options.mp4</nowiki>

== Technical specification ==
{| class="wikitable"
|+
!
!Default
!Max Configuration
!Options
|-
|Power
|2000W Platinum Modular PSU
|Configurable up to 5600W
|Dual 2200W '''Titanium''' CRPS
Automatic switching between 1+1 and 2+0 based on [[workload]]
|-
|Storage
|3 x 5.15" dummy cage
|Any type of 5.25" hard rack
|[[Category:Server]]
Depending on models
[[Category:Reference]]
|-
|Dimension
| colspan="3" |4U rack mount server, (441mm x 177mm x 866mm)
|-
|Cooling
|Air Cooling
Cooling 4 x middle / 2 x rear fan (Max. 4,500 RPM/168.2CFM/54.5dbA), 1 x rear fan (Max. 4800 RPM/60.8 CFM/44.5dbA) / up to 500W active CPU cooler (Adjustable noise / cooling ratio)
|
|Optional [[Liquid cooling|Liquid Cooling]] with [[CoolUX]]
|-
|Front I/O
| colspan="3" |Button(1 x Power, 1 x Reset), LED(1 x Power, 1 x HDD Activity, 2 x [[Network]] Activity)
|-
|SKU
| colspan="2" |HM-U48-22K5G ( Air cooling)
|Configuration
|-
|
|CPU
|Single Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor, TDP up to 350W
|
|-
|
|Memory
|8 x DDR5 DIMM Slots, 8 channels(1DPC), RDIMM/RIMM-3DS/MCRDIMM (RDIMM: 6400MT/s, MCRDIMM: 8000MT/s)
|
|-
|
|Storage
|4 x U.2/U.3 hot-swap front drive (Optional configuration +8 x U.2/U.3 or +8 x NVMe or +8 x SATA SSD **)
|
|-
|
|I/O
|<nowiki>- 2 x 2280/22110 PCIe 5.0 x2 NVMe M.2 ports, 3 x SlimlineSAS 4i for 12 SATA3.0 drives (NVMe RAID 0/1/5/10 support)</nowiki>

- 4 x [[PCIe]] 5.0 x16 slots (one slot only functional with R1S CPU), 6 x PCIe 5.0 x8 MCIO

- 2x USB 3.2 Type A, 1x DB9, 1 COM port, UID LED button
|
|-
|
|PCIe
|11 x Full-Height/Full-Length expansion slots, 7 x PCIe 5.0 x16 (supports 4 x CXL2.0* x16), 1 x PCIe5.0 x8
|
|-
|
|GPU
|Default [[support]] 4 x double width 600W GPUs ( Optional configuration +1 x double width 600W GPU** )
|5 x double width 600W GPUs on PCIe Gen5 x16
|-
|
|Network
|2 x 1000 Mbps RJ45, 1 x 1000 Mbps RJ45 Dedicated [[IPMI]] LAN
|
|-
|
|VGA
|VGA 1 x D-Sub(AST2600 )
|
|-
|
|Software
| colspan="2" |HPCMATE system level optimized Linux based Ubuntu LTS (Default)
Optional Windows/Windows Server or other Linux distribution

Optional HPCMATE DLS (Deep Learning Station) Framework.

Optional HPCMATE SmartLLM (HMSmartLLM)Framework.

Various optional on-demand custom software solutions
|}

== Contact US ==
Mail us to sales@hpcmate.com

U4008

2026-10-02T04:03:02Z

Admin:

{{Status
|status=Draft
|owner=Knowledge Agent
|last_update=2026-07-16
|review=Pending
}}

{{TOC}}

== Overview ==

U4008은 HPCMATE 가 독자개발한 다양한 기능을 내포할 수 있는 시스템 입니다.

== Key Concepts ==

{| class="wikitable"
! Concept
! Description
! Related
|-
| U4008
| HPC/서버 환경에서 중요한 기술 개념
| [[Linux]], [[Server]]
|}

== U4008 Series ==

U4008 has been designed to provide a flexible system configuration high performance computer server for AI/ML/LLM.
[[File:U4008.png|center|frameless|446x446px]]
In brief : U4008 has been designed to provide a flexible system configuration high performance computer server for AI/ML/LLM

== Features ==

* 4U rack mount server, (W441mm x H177mm x D866mm, D930mm with external Fan)
* Single Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor, TDP up to 350W
* Default support 4 x double width 600W GPUs (Optional configuration +1 x double width 600W GPU, Total 5 x 600W double width GPU** )
* 8 x DDR5 DIMM Slots, 8 channels(1DPC), RDIMM/RIMM-3DS/MCRDIMM (RDIMM: 6400MT/s, MCRDIMM: 8000MT/s)
* 4 x U.2/U.3 hot-swap front drive (Optional configuration +8 x U.2/U.3 or +8 x NVMe or +8 x SATA SSD **)
* NVMe RAID 0/1/5/10 support (Intel® VROC RAID key required)
* 2 x 1000 Mbps RJ45, 1 x 1000 Mbps RJ45 Dedicated IPMI 2.0 LAN
* Button (1 x Power, 1 x Reset), LED(1 x Power, 1 x HDD Activity, 2 x Network Activity)
* 2 x USB 3.2 Gen 1 Type-A ports, 1 x VGA D-Sub port, 1 x COM port, 1 x UID LED button
* 2 x 2200W Platinum PSU (automatic mode : 1+1 2200W or 2+0 4400W depending on the workload)
* Introduction
** <nowiki>http://www.hpcmate.com/src/video/U4008.mp4</nowiki>
** <nowiki>http://www.hpcmate.com/src/video/U4008_PSU_Options.mp4</nowiki>

== Technical specification ==
{| class="wikitable"
|+
!
!Default
!Max Configuration
!Options
|-
|Power
|2000W Platinum Modular PSU
|Configurable up to 5600W
|Dual 2200W '''Titanium''' CRPS
Automatic switching between 1+1 and 2+0 based on [[workload]]
|-
|Storage
|3 x 5.15" dummy cage
|Any type of 5.25" hard rack
|[[Category:Server]]
Depending on models
[[Category:Reference]]
|-
|Dimension
| colspan="3" |4U rack mount server, (441mm x 177mm x 866mm)
|-
|Cooling
|Air Cooling
Cooling 4 x middle / 2 x rear fan (Max. 4,500 RPM/168.2CFM/54.5dbA), 1 x rear fan (Max. 4800 RPM/60.8 CFM/44.5dbA) / up to 500W active CPU cooler (Adjustable noise / cooling ratio)
|
|Optional [[Liquid cooling|Liquid Cooling]] with [[CoolUX]]
|-
|Front I/O
| colspan="3" |Button(1 x Power, 1 x Reset), LED(1 x Power, 1 x HDD Activity, 2 x [[Network]] Activity)
|-
|SKU
| colspan="2" |HM-U48-22K5G ( Air cooling)
|Configuration
|-
|
|CPU
|Single Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor, TDP up to 350W
|
|-
|
|Memory
|8 x DDR5 DIMM Slots, 8 channels(1DPC), RDIMM/RIMM-3DS/MCRDIMM (RDIMM: 6400MT/s, MCRDIMM: 8000MT/s)
|
|-
|
|Storage
|4 x U.2/U.3 hot-swap front drive (Optional configuration +8 x U.2/U.3 or +8 x NVMe or +8 x SATA SSD **)
|
|-
|
|I/O
|<nowiki>- 2 x 2280/22110 PCIe 5.0 x2 NVMe M.2 ports, 3 x SlimlineSAS 4i for 12 SATA3.0 drives (NVMe RAID 0/1/5/10 support)</nowiki>

- 4 x [[PCIe]] 5.0 x16 slots (one slot only functional with R1S CPU), 6 x PCIe 5.0 x8 MCIO

- 2x USB 3.2 Type A, 1x DB9, 1 COM port, UID LED button
|
|-
|
|PCIe
|11 x Full-Height/Full-Length expansion slots, 7 x PCIe 5.0 x16 (supports 4 x CXL2.0* x16), 1 x PCIe5.0 x8
|
|-
|
|GPU
|Default [[support]] 4 x double width 600W GPUs ( Optional configuration +1 x double width 600W GPU** )
|5 x double width 600W GPUs on PCIe Gen5 x16
|-
|
|Network
|2 x 1000 Mbps RJ45, 1 x 1000 Mbps RJ45 Dedicated [[IPMI]] LAN
|
|-
|
|VGA
|VGA 1 x D-Sub(AST2600 )
|
|-
|
|Software
| colspan="2" |HPCMATE system level optimized Linux based Ubuntu LTS (Default)
Optional Windows/Windows Server or other Linux distribution

Optional HPCMATE DLS (Deep Learning Station) Framework.

Optional HPCMATE SmartLLM (HMSmartLLM)Framework.

Various optional on-demand custom software solutions
|}

== Contact US ==
Mail us to sales@hpcmate.com

U4008

2026-09-04T02:16:43Z

Admin:

대문

2026-08-19T11:34:41Z

Admin:

Wiki for those who are enthusiastic for High Performance Computing ![[File:Galileo Galilei.png|center|frameless|539x539px]]

[[Category:Server]]

[[Category:Reference]]

JS tips and tricks

2026-07-14T05:32:04Z

Admin: Undo revision 2343 by Clara (talk)

== Display JavaScript datetime in 12 hour AM/PM format<ref>https://stackoverflow.com/questions/8888491/how-do-you-display-javascript-datetime-in-12-hour-am-pm-format</ref> ==
<code>only toLocaleTimeString() may behave differently based on region / location</code>

<syntaxhighlight lang="javascript">
var time = new Date();
time
Thu May 30 2024 11:56:46 GMT+0900 (한국 표준시)
time.toLocaleString('en-US', { hour: 'numeric', minute: 'numeric', hour12: true })
'11:56 AM'
time.toLocaleString('ko-KR', { hour: 'numeric', minute: 'numeric', hour12: true })
'오전 11:56'
time.toLocaleString('ko-KR')
'2024. 5. 30. 오전 11:56:46'
</syntaxhighlight>

== References ==
<references />

AI Agents

2026-07-03T01:00:22Z

Admin: /* Motivation */

== Motivation ==
Year 2026 would be remembered in human history as a year to use AI for small and large operation.

To help customers, we starts to provide live updated AI related topics as easy as possible based on [[AIS (AI Station)]] production.

== What is AI Agent ==
An AI agent is an autonomous software system that uses artificial intelligence to perceive its environment, make decisions, reason, and take action to achieve specific goals. for example, Digital Work Assistants, Smart Home & Automation and execute pre-planned actions without needing your trigger like press a button to start/resume.

== AI Model vs. AI Agent ==

* '''AI Models''' (like Qwen, GPT or Gemini) are the "brains." They are excellent at generating text, understanding context, and answering questions, but they require direct, step-by-step human prompts.
* '''AI Agents''' are the full software systems. They use the AI model as their core intelligence but package it with memory, tools, and decision-making logic so they can actually ''do'' the work for you. for example, Hermes, Claude Code, Codex, Opencraw etc

== References ==
<references />

AI Agents

2026-07-03T00:59:49Z

Admin: /* AI Model vs. AI Agent */

== Motivation ==
Year 2026 would be remembered in human history as a year to use AI for small and large operation.

To help customers, we starts to provide live updated AI related topics as easy as possible based on AIS (AI Station) production.

== What is AI Agent ==
An AI agent is an autonomous software system that uses artificial intelligence to perceive its environment, make decisions, reason, and take action to achieve specific goals. for example, Digital Work Assistants, Smart Home & Automation and execute pre-planned actions without needing your trigger like press a button to start/resume.

== AI Model vs. AI Agent ==

* '''AI Models''' (like Qwen, GPT or Gemini) are the "brains." They are excellent at generating text, understanding context, and answering questions, but they require direct, step-by-step human prompts.
* '''AI Agents''' are the full software systems. They use the AI model as their core intelligence but package it with memory, tools, and decision-making logic so they can actually ''do'' the work for you. for example, Hermes, Claude Code, Codex, Opencraw etc

== References ==
<references />

AI Agents

2026-07-03T00:56:58Z

Admin: /* AI Agents */

== Motivation ==
Year 2026 would be remembered in human history as a year to use AI for small and large operation.

To help customers, we starts to provide live updated AI related topics as easy as possible based on AIS (AI Station) production.

== What is AI Agent ==
An AI agent is an autonomous software system that uses artificial intelligence to perceive its environment, make decisions, reason, and take action to achieve specific goals

== AI Model vs. AI Agent ==

* '''AI Models''' (like Qwen, GPT or Gemini) are the "brains." They are excellent at generating text, understanding context, and answering questions, but they require direct, step-by-step human prompts.
* '''AI Agents''' are the full software systems. They use the AI model as their core intelligence but package it with memory, tools, and decision-making logic so they can actually ''do'' the work for you. for example, Hermes, Claude Code, Codex, Opencraw etc

== References ==
<references />

AI Agents

2026-07-03T00:43:04Z

Admin: Created page with "== AI Agents == Year 2026 would be remembered in human history as a year to use AI for small and large operation. To help customers, we starts to provide live updated AI related topics as easy as possible based on AIS (AI Station) production. == References == <references />"

== AI Agents ==
Year 2026 would be remembered in human history as a year to use AI for small and large operation.

To help customers, we starts to provide live updated AI related topics as easy as possible based on AIS (AI Station) production.

== References ==
<references />

AI Coding Tool

2026-07-03T00:34:34Z

Admin: /* AI Coling Tools */

== AI Coding Tools ==
{| class="wikitable sortable" style="text-align: left;"
|+ AI Coding Tools Comparison Matrix (2026)
! Scope !! [https://github.com/anthropics/claude-code Claude Code] !! [https://github.com/anomalyco/opencode OpenCode] !! [https://cursor.com/ Cursor] !! [https://kilo.ai/ Kilo Code] !! [https://github.com Aider]
|-
! Core Nature
| Official Anthropic terminal agent
| Model-agnostic open-source agent
| VS Code fork-based AI native IDE
| Enterprise-focused hybrid agent
| Git-integrated CLI coding assistant
|-
! Primary UI
| Terminal CLI
| Terminal TUI / Desktop App / Web UI
| Standalone Desktop IDE
| VS Code & JetBrains Plugins / CLI
| Terminal CLI
|-
! Supported Models
| Claude ecosystem exclusively
| 75+ providers (GPT, Gemini, Local LLMs)
| Multi-model [[support]] + custom finetunes
| 500+ (Local and Cloud LLMs)
| Multi-model support via API keys
|-
! Pricing Model
| Paid subscription or usage-based API
| 100% Free tool (BYOK / Local)
| Free tier / $20/month Pro tier
| Enterprise plans / Usage-based
| 100% Free tool (BYOK)
|-
! License Type
| Proprietary (Closed-Source)
| Open-Source (MIT License)
| Proprietary (Closed-Source)
| Hybrid / Commercial
| Open-Source (Apache 2.0)
|-
! Key Strength
| Lightning-fast agentic [[feedback]] loops
| Rigorous full [[test]] suite validation
| Seamless tab-completion & low friction
| Multi-IDE support & remote environment
| Flawless git integration & auto-commits
|}

== LLM Serving Framework ==
{| class="wikitable sortable" style="text-align: left;"
|+ LLM Service & Development Frameworks Comparison
! Feature !! [https://langchain.com LangChain] !! [https://llamaindex.ai LlamaIndex] !! [https://langchain.comlanggraph LangGraph] !! [https://crewai.com CrewAI] !! [https://github.com vLLM] !! [https://github.com llama.cpp]
|-
! Core Focus
| General-purpose LLM orchestration
| Data ingestion, indexing & RAG
| Complex cyclical & stateful agents
| Multi-agent role-playing & tasks
| High-performance enterprise serving
| Ultra-lightweight local deployment & quantization
|-
! Architecture
| Chain-based sequential pipelines
| Hierarchical data structures & indexes
| Graph-based state machines (DAGs/Cyclic)
| Role-based autonomous agent crews
| C++ / Python optimized inference engines
| Pure C/C++ implementation with no dependencies
|-
! Primary Use Case
| Quick prototyping of simple LLM apps
| Advanced Search, QA, and Enterprise RAG
| Enterprise-grade complex agent workflows
| Process automation & multi-agent debate
| Scaling open-source models on cloud GPUs
| Running LLMs on consumer hardware, MacBooks, and edge devices
|-
! State Management
| Basic memory components
| Stateless query engines (mostly)
| Rich, persistent, multi-actor state
| Shared memory & task-state tracking
| Stateless token generation (KV caching)
| Direct memory-mapped file loading (mmap)
|-
! Learning Curve
| Moderate (Highly abstracted)
| Moderate (Data-focused)
| Steep (Requires graph-thinking)
| Low to Moderate (Intuitive design)
| High (Requires infrastructure & cloud GPU [[optimization]])
| Moderate (Requires command-line and build knowledge)
|-
! Key Strength
| Massive ecosystem & integrations
| Unmatched data retrieval efficiency
| Deterministic control over chaotic agents
| Easy human-in-the-loop setup
| Maximum throughput via PagedAttention
| Incredible CPU/GPU hybrid execution & portability
|}

== References ==
<references />

AI Coding Tool

2026-06-25T23:27:44Z

Admin: /* AI Coling Tools */

== AI Coling Tools ==
{| class="wikitable sortable" style="text-align: left;"
|+ AI Coding Tools Comparison Matrix (2026)
! Scope !! [https://github.com/anthropics/claude-code Claude Code] !! [https://github.com/anomalyco/opencode OpenCode] !! [https://cursor.com/ Cursor] !! [https://kilo.ai/ Kilo Code] !! [https://github.com Aider]
|-
! Core Nature
| Official Anthropic terminal agent
| Model-agnostic open-source agent
| VS Code fork-based AI native IDE
| Enterprise-focused hybrid agent
| Git-integrated CLI coding assistant
|-
! Primary UI
| Terminal CLI
| Terminal TUI / Desktop App / Web UI
| Standalone Desktop IDE
| VS Code & JetBrains Plugins / CLI
| Terminal CLI
|-
! Supported Models
| Claude ecosystem exclusively
| 75+ providers (GPT, Gemini, Local LLMs)
| Multi-model [[support]] + custom finetunes
| 500+ (Local and Cloud LLMs)
| Multi-model support via API keys
|-
! Pricing Model
| Paid subscription or usage-based API
| 100% Free tool (BYOK / Local)
| Free tier / $20/month Pro tier
| Enterprise plans / Usage-based
| 100% Free tool (BYOK)
|-
! License Type
| Proprietary (Closed-Source)
| Open-Source (MIT License)
| Proprietary (Closed-Source)
| Hybrid / Commercial
| Open-Source (Apache 2.0)
|-
! Key Strength
| Lightning-fast agentic [[feedback]] loops
| Rigorous full [[test]] suite validation
| Seamless tab-completion & low friction
| Multi-IDE support & remote environment
| Flawless git integration & auto-commits
|}

== LLM Serving Framework ==
{| class="wikitable sortable" style="text-align: left;"
|+ LLM Service & Development Frameworks Comparison
! Feature !! [https://langchain.com LangChain] !! [https://llamaindex.ai LlamaIndex] !! [https://langchain.comlanggraph LangGraph] !! [https://crewai.com CrewAI] !! [https://github.com vLLM] !! [https://github.com llama.cpp]
|-
! Core Focus
| General-purpose LLM orchestration
| Data ingestion, indexing & RAG
| Complex cyclical & stateful agents
| Multi-agent role-playing & tasks
| High-performance enterprise serving
| Ultra-lightweight local deployment & quantization
|-
! Architecture
| Chain-based sequential pipelines
| Hierarchical data structures & indexes
| Graph-based state machines (DAGs/Cyclic)
| Role-based autonomous agent crews
| C++ / Python optimized inference engines
| Pure C/C++ implementation with no dependencies
|-
! Primary Use Case
| Quick prototyping of simple LLM apps
| Advanced Search, QA, and Enterprise RAG
| Enterprise-grade complex agent workflows
| Process automation & multi-agent debate
| Scaling open-source models on cloud GPUs
| Running LLMs on consumer hardware, MacBooks, and edge devices
|-
! State Management
| Basic memory components
| Stateless query engines (mostly)
| Rich, persistent, multi-actor state
| Shared memory & task-state tracking
| Stateless token generation (KV caching)
| Direct memory-mapped file loading (mmap)
|-
! Learning Curve
| Moderate (Highly abstracted)
| Moderate (Data-focused)
| Steep (Requires graph-thinking)
| Low to Moderate (Intuitive design)
| High (Requires infrastructure & cloud GPU [[optimization]])
| Moderate (Requires command-line and build knowledge)
|-
! Key Strength
| Massive ecosystem & integrations
| Unmatched data retrieval efficiency
| Deterministic control over chaotic agents
| Easy human-in-the-loop setup
| Maximum throughput via PagedAttention
| Incredible CPU/GPU hybrid execution & portability
|}

== References ==
<references />

AI Coding Tool

2026-06-25T23:27:24Z

Admin: /* LLM Serving Framework */

== AI Coling Tools ==
{| class="wikitable sortable" style="text-align: left;"
|+ AI Coding Tools Comparison Matrix (2026)
! Scope !! Claude Code (Anthropic) !! OpenCode
(Open-Source)
! Cursor
(Anysphere)
! Kilo Code
(Kilo AI)
! Aider
|-
! Core Nature
| Official Anthropic terminal agent
| Model-agnostic open-source agent
| VS Code fork-based AI native IDE
| Enterprise-focused hybrid agent
| Git-integrated CLI coding assistant
|-
! Primary UI
| Terminal CLI
| Terminal TUI / Desktop App / Web UI
| Standalone Desktop IDE
| VS Code & JetBrains Plugins / CLI
| Terminal CLI
|-
! Supported Models
| Claude ecosystem exclusively
| 75+ providers (GPT, Gemini, Local LLMs)
| Multi-model [[support]] + custom finetunes
| 500+ (Local and Cloud LLMs)
| Multi-model support via API keys
|-
! Pricing Model
| Paid subscription or usage-based API
| 100% Free tool (BYOK / Local)
| Free tier / $20/month Pro tier
| Enterprise plans / Usage-based
| 100% Free tool (BYOK)
|-
! License Type
| Proprietary (Closed-Source)
| Open-Source (MIT License)
| Proprietary (Closed-Source)
| Hybrid / Commercial
| Open-Source (Apache 2.0)
|-
! Key Strength
| Lightning-fast agentic [[feedback]] loops
| Rigorous full [[test]] suite validation
| Seamless tab-completion & low friction
| Multi-IDE support & remote environment
| Flawless git integration & auto-commits
|}

== LLM Serving Framework ==
{| class="wikitable sortable" style="text-align: left;"
|+ LLM Service & Development Frameworks Comparison
! Feature !! [https://langchain.com LangChain] !! [https://llamaindex.ai LlamaIndex] !! [https://langchain.comlanggraph LangGraph] !! [https://crewai.com CrewAI] !! [https://github.com vLLM] !! [https://github.com llama.cpp]
|-
! Core Focus
| General-purpose LLM orchestration
| Data ingestion, indexing & RAG
| Complex cyclical & stateful agents
| Multi-agent role-playing & tasks
| High-performance enterprise serving
| Ultra-lightweight local deployment & quantization
|-
! Architecture
| Chain-based sequential pipelines
| Hierarchical data structures & indexes
| Graph-based state machines (DAGs/Cyclic)
| Role-based autonomous agent crews
| C++ / Python optimized inference engines
| Pure C/C++ implementation with no dependencies
|-
! Primary Use Case
| Quick prototyping of simple LLM apps
| Advanced Search, QA, and Enterprise RAG
| Enterprise-grade complex agent workflows
| Process automation & multi-agent debate
| Scaling open-source models on cloud GPUs
| Running LLMs on consumer hardware, MacBooks, and edge devices
|-
! State Management
| Basic memory components
| Stateless query engines (mostly)
| Rich, persistent, multi-actor state
| Shared memory & task-state tracking
| Stateless token generation (KV caching)
| Direct memory-mapped file loading (mmap)
|-
! Learning Curve
| Moderate (Highly abstracted)
| Moderate (Data-focused)
| Steep (Requires graph-thinking)
| Low to Moderate (Intuitive design)
| High (Requires infrastructure & cloud GPU [[optimization]])
| Moderate (Requires command-line and build knowledge)
|-
! Key Strength
| Massive ecosystem & integrations
| Unmatched data retrieval efficiency
| Deterministic control over chaotic agents
| Easy human-in-the-loop setup
| Maximum throughput via PagedAttention
| Incredible CPU/GPU hybrid execution & portability
|}

== References ==
<references />

AI Coding Tool

2026-06-25T23:25:11Z

Admin: /* LLM Serving Framework */

== AI Coling Tools ==
{| class="wikitable sortable" style="text-align: left;"
|+ AI Coding Tools Comparison Matrix (2026)
! Scope !! Claude Code (Anthropic) !! OpenCode
(Open-Source)
! Cursor
(Anysphere)
! Kilo Code
(Kilo AI)
! Aider
|-
! Core Nature
| Official Anthropic terminal agent
| Model-agnostic open-source agent
| VS Code fork-based AI native IDE
| Enterprise-focused hybrid agent
| Git-integrated CLI coding assistant
|-
! Primary UI
| Terminal CLI
| Terminal TUI / Desktop App / Web UI
| Standalone Desktop IDE
| VS Code & JetBrains Plugins / CLI
| Terminal CLI
|-
! Supported Models
| Claude ecosystem exclusively
| 75+ providers (GPT, Gemini, Local LLMs)
| Multi-model [[support]] + custom finetunes
| 500+ (Local and Cloud LLMs)
| Multi-model support via API keys
|-
! Pricing Model
| Paid subscription or usage-based API
| 100% Free tool (BYOK / Local)
| Free tier / $20/month Pro tier
| Enterprise plans / Usage-based
| 100% Free tool (BYOK)
|-
! License Type
| Proprietary (Closed-Source)
| Open-Source (MIT License)
| Proprietary (Closed-Source)
| Hybrid / Commercial
| Open-Source (Apache 2.0)
|-
! Key Strength
| Lightning-fast agentic [[feedback]] loops
| Rigorous full [[test]] suite validation
| Seamless tab-completion & low friction
| Multi-IDE support & remote environment
| Flawless git integration & auto-commits
|}

== LLM Serving Framework ==
{| class="wikitable sortable" style="text-align: left;"
|+ LLM Service & Development Frameworks Comparison (2026)
! Feature !! LangChain !! LlamaIndex !! LangGraph !! CrewAI !! vLLM !! llama.cpp
|-
! Core Focus
| General-purpose LLM orchestration
| Data ingestion, indexing & RAG
| Complex cyclical & stateful agents
| Multi-agent role-playing & tasks
| High-performance enterprise serving
| Ultra-lightweight local deployment & quantization
|-
! Architecture
| Chain-based sequential pipelines
| Hierarchical data structures & indexes
| Graph-based state machines (DAGs/Cyclic)
| Role-based autonomous agent crews
| C++ / Python optimized inference engines
| Pure C/C++ implementation with no dependencies
|-
! Primary Use Case
| Quick prototyping of simple LLM apps
| Advanced Search, QA, and Enterprise RAG
| Enterprise-grade complex agent workflows
| Process automation & multi-agent debate
| Scaling open-source models on cloud GPUs
| Running LLMs on consumer hardware, MacBooks, and edge devices
|-
! State Management
| Basic memory components
| Stateless query engines (mostly)
| Rich, persistent, multi-actor state
| Shared memory & task-state tracking
| Stateless token generation (KV caching)
| Direct memory-mapped file loading (mmap)
|-
! Learning Curve
| Moderate (Highly abstracted)
| Moderate (Data-focused)
| Steep (Requires graph-thinking)
| Low to Moderate (Intuitive design)
| High (Requires infrastructure & cloud GPU [[optimization]])
| Moderate (Requires command-line and build knowledge)
|-
! Key Strength
| Massive ecosystem & integrations
| Unmatched data retrieval efficiency
| Deterministic control over chaotic agents
| Easy human-in-the-loop setup
| Maximum throughput via PagedAttention
| Incredible CPU/GPU hybrid execution & portability
|}

== References ==
<references />

AI Coding Tool

2026-06-25T23:24:56Z

Admin: /* AI Coling Tools */

== AI Coling Tools ==
{| class="wikitable sortable" style="text-align: left;"
|+ AI Coding Tools Comparison Matrix (2026)
! Scope !! Claude Code (Anthropic) !! OpenCode
(Open-Source)
! Cursor
(Anysphere)
! Kilo Code
(Kilo AI)
! Aider
|-
! Core Nature
| Official Anthropic terminal agent
| Model-agnostic open-source agent
| VS Code fork-based AI native IDE
| Enterprise-focused hybrid agent
| Git-integrated CLI coding assistant
|-
! Primary UI
| Terminal CLI
| Terminal TUI / Desktop App / Web UI
| Standalone Desktop IDE
| VS Code & JetBrains Plugins / CLI
| Terminal CLI
|-
! Supported Models
| Claude ecosystem exclusively
| 75+ providers (GPT, Gemini, Local LLMs)
| Multi-model [[support]] + custom finetunes
| 500+ (Local and Cloud LLMs)
| Multi-model support via API keys
|-
! Pricing Model
| Paid subscription or usage-based API
| 100% Free tool (BYOK / Local)
| Free tier / $20/month Pro tier
| Enterprise plans / Usage-based
| 100% Free tool (BYOK)
|-
! License Type
| Proprietary (Closed-Source)
| Open-Source (MIT License)
| Proprietary (Closed-Source)
| Hybrid / Commercial
| Open-Source (Apache 2.0)
|-
! Key Strength
| Lightning-fast agentic [[feedback]] loops
| Rigorous full [[test]] suite validation
| Seamless tab-completion & low friction
| Multi-IDE support & remote environment
| Flawless git integration & auto-commits
|}

== LLM Serving Framework ==
{| class="wikitable sortable" style="text-align: left;"
|+ LLM Service & Development Frameworks Comparison
! Feature !! LangChain !! LlamaIndex !! LangGraph !! CrewAI !! vLLM !! llama.cpp
|-
! Core Focus
| General-purpose LLM orchestration
| Data ingestion, indexing & RAG
| Complex cyclical & stateful agents
| Multi-agent role-playing & tasks
| High-performance enterprise serving
| Ultra-lightweight local deployment & quantization
|-
! Architecture
| Chain-based sequential pipelines
| Hierarchical data structures & indexes
| Graph-based state machines (DAGs/Cyclic)
| Role-based autonomous agent crews
| C++ / Python optimized inference engines
| Pure C/C++ implementation with no dependencies
|-
! Primary Use Case
| Quick prototyping of simple LLM apps
| Advanced Search, QA, and Enterprise RAG
| Enterprise-grade complex agent workflows
| Process automation & multi-agent debate
| Scaling open-source models on cloud GPUs
| Running LLMs on consumer hardware, MacBooks, and edge devices
|-
! State Management
| Basic memory components
| Stateless query engines (mostly)
| Rich, persistent, multi-actor state
| Shared memory & task-state tracking
| Stateless token generation (KV caching)
| Direct memory-mapped file loading (mmap)
|-
! Learning Curve
| Moderate (Highly abstracted)
| Moderate (Data-focused)
| Steep (Requires graph-thinking)
| Low to Moderate (Intuitive design)
| High (Requires infrastructure & cloud GPU [[optimization]])
| Moderate (Requires command-line and build knowledge)
|-
! Key Strength
| Massive ecosystem & integrations
| Unmatched data retrieval efficiency
| Deterministic control over chaotic agents
| Easy human-in-the-loop setup
| Maximum throughput via PagedAttention
| Incredible CPU/GPU hybrid execution & portability
|}

== References ==
<references />

AI Coding Tool

2026-06-25T23:23:26Z

Admin:

== AI Coling Tools ==
{| class="wikitable sortable" style="text-align: left;"
|+ AI Coding Tools Comparison Matrix (2026)
! Scope !! Claude Code (Anthropic) !! OpenCode
(Open-Source)
! Cursor
(Anysphere)
! Kilo Code
(Kilo AI)
! Aider
|-
! Core Nature
| Official Anthropic terminal agent
| Model-agnostic open-source agent
| VS Code fork-based AI native IDE
| Enterprise-focused hybrid agent
| Git-integrated CLI coding assistant
|-
! Primary UI
| Terminal CLI
| Terminal TUI / Desktop App / Web UI
| Standalone Desktop IDE
| VS Code & JetBrains Plugins / CLI
| Terminal CLI
|-
! Supported Models
| Claude ecosystem exclusively
| 75+ providers (GPT, Gemini, Local LLMs)
| Multi-model [[support]] + custom finetunes
| 500+ (Local and Cloud LLMs)
| Multi-model support via API keys
|-
! Pricing Model
| Paid subscription or usage-based API
| 100% Free tool (BYOK / Local)
| Free tier / $20/month Pro tier
| Enterprise plans / Usage-based
| 100% Free tool (BYOK)
|-
! License Type
| Proprietary (Closed-Source)
| Open-Source (MIT License)
| Proprietary (Closed-Source)
| Hybrid / Commercial
| Open-Source (Apache 2.0)
|-
! Key Strength
| Lightning-fast agentic [[feedback]] loops
| Rigorous full [[test]] suite validation
| Seamless tab-completion & low friction
| Multi-IDE support & remote environment
| Flawless git integration & auto-commits
|}

LLM Serving Framework
{| class="wikitable sortable" style="text-align: left;"
|+ LLM Service & Development Frameworks Comparison
! Feature !! LangChain !! LlamaIndex !! LangGraph !! CrewAI !! vLLM / Ollama
|-
! Core Focus
| General-purpose LLM orchestration
| Data ingestion, indexing & RAG
| Complex cyclical & stateful agents
| Multi-agent role-playing & tasks
| High-performance LLM serving & inference
|-
! Architecture
| Chain-based sequential pipelines
| Hierarchical data structures & indexes
| Graph-based state machines (DAGs/Cyclic)
| Role-based autonomous agent crews
| C++ / Python optimized inference engines
|-
! Primary Use Case
| Quick prototyping of simple LLM apps
| Advanced Search, QA, and Enterprise RAG
| Enterprise-grade complex agent workflows
| Process automation & multi-agent debate
| Self-hosting & serving open-source models
|-
! State Management
| Basic memory components
| Stateless query engines (mostly)
| Rich, persistent, multi-actor state
| Shared memory & task-state tracking
| Stateless token generation (KV caching)
|-
! Learning Curve
| Moderate (Highly abstracted)
| Moderate (Data-focused)
| Steep (Requires graph-thinking)
| Low to Moderate (Intuitive design)
| Low (API setup) / High (Infrastructure [[optimization]])
|-
! Key Strength
| Massive ecosystem & integrations
| Unmatched data retrieval efficiency
| Deterministic control over chaotic agents
| Easy human-in-the-loop setup
| High throughput & structural memory savings
|}

== References ==
<references />

AI Coding Tool

2026-06-25T23:22:11Z

Admin: /* AI Coling Tools */

AI Coding Tool

2026-06-25T23:21:34Z

Admin: /* AI Coling Tools */

AI Coding Tool

2026-06-25T23:20:45Z

Admin:

AI Coding Tool

2026-06-25T23:20:23Z

Admin: Created page with " == References == <references />"

== References ==
<references />

Openclaw Agents

2026-06-03T04:09:19Z

Admin: /* HPCMATE Openclaw Agents */

== Type of Agents (Openclaw) ==

{| class="wikitable sort0"
|+ Comparison of Multiple Agent Configuration Strategies in OpenClaw<ref>Generated by HPCMATE local AI</ref>
! Strategy !! Description !! Advantages !! Best Use Case
|-
| '''Strategy ①: Specialized Individual Agents ([[Library]])''' || Create separate .md files for each role, such as coder.md, tester.md, or reviewer.md. || Simple to implement and very easy to manage individual roles. || Quick, single-purpose tasks or simple automation.
|-
| '''Strategy ②: Orchestrator Pattern (Manager-Worker)''' || A "Manager" agent analyzes the user's request and uses sessions_spawn to delegate sub-tasks to specific agents. || High autonomy; can handle complex, multi-step reasoning without human intervention. || Complex software development or multi-stage project management.
|-
| '''Strategy ③: Pipeline Configuration (Sequential)''' || A chain of agents where the output of one agent serves as the input for the next (e.g., Scraper -> Analyzer -> Notifier). || Provides a robust, automated workflow for continuous data processing. || Automated monitoring, data extraction, and reporting pipelines.
|}

== HPCMATE Openclaw Agents ==

{| class="wikitable"
|+
!Agent Group
!Type
!Description
!Details
|-
|[[Private:Baremetal System Configuration|Baremetal System Configuration]]
|Orchestrator
|We use this agent to setup general purpose bare metal system configuration
Such as DLS(Deep Learning Station), AIS (AI Station) for each node
|TBD
|-
|[[Private:Cluster Configuration|Cluster Configuration]]
|
|We use this agent to setup a cluster master and slave for micro datacenter
|TBD
|-
|
|
|
|
|}

== Reference ==
<references />

Openclaw Agents

2026-06-03T04:06:43Z

Admin: /* HPCMATE Openclaw Agents */

== Type of Agents (Openclaw) ==

{| class="wikitable sort0"
|+ Comparison of Multiple Agent Configuration Strategies in OpenClaw<ref>Generated by HPCMATE local AI</ref>
! Strategy !! Description !! Advantages !! Best Use Case
|-
| '''Strategy ①: Specialized Individual Agents ([[Library]])''' || Create separate .md files for each role, such as coder.md, tester.md, or reviewer.md. || Simple to implement and very easy to manage individual roles. || Quick, single-purpose tasks or simple automation.
|-
| '''Strategy ②: Orchestrator Pattern (Manager-Worker)''' || A "Manager" agent analyzes the user's request and uses sessions_spawn to delegate sub-tasks to specific agents. || High autonomy; can handle complex, multi-step reasoning without human intervention. || Complex software development or multi-stage project management.
|-
| '''Strategy ③: Pipeline Configuration (Sequential)''' || A chain of agents where the output of one agent serves as the input for the next (e.g., Scraper -> Analyzer -> Notifier). || Provides a robust, automated workflow for continuous data processing. || Automated monitoring, data extraction, and reporting pipelines.
|}

== HPCMATE Openclaw Agents ==

{| class="wikitable"
|+
!Agent Group
!Type
!Description
!Details
|-
|[[Baremetal System Configuration]]
|Orchestrator
|We use this agent to setup general purpose bare metal system configuration
Such as DLS(Deep Learning Station), AIS (AI Station) for each node
|TBD
|-
|[[Cluster Configuration]]
|
|We use this agent to setup a cluster master and slave for micro datacenter
|TBD
|-
|
|
|
|
|}

== Reference ==
<references />

Openclaw Agents

2026-06-03T04:06:01Z

Admin: /* HPCMATE Openclaw Agents */

== Type of Agents (Openclaw) ==

{| class="wikitable sort0"
|+ Comparison of Multiple Agent Configuration Strategies in OpenClaw<ref>Generated by HPCMATE local AI</ref>
! Strategy !! Description !! Advantages !! Best Use Case
|-
| '''Strategy ①: Specialized Individual Agents ([[Library]])''' || Create separate .md files for each role, such as coder.md, tester.md, or reviewer.md. || Simple to implement and very easy to manage individual roles. || Quick, single-purpose tasks or simple automation.
|-
| '''Strategy ②: Orchestrator Pattern (Manager-Worker)''' || A "Manager" agent analyzes the user's request and uses sessions_spawn to delegate sub-tasks to specific agents. || High autonomy; can handle complex, multi-step reasoning without human intervention. || Complex software development or multi-stage project management.
|-
| '''Strategy ③: Pipeline Configuration (Sequential)''' || A chain of agents where the output of one agent serves as the input for the next (e.g., Scraper -> Analyzer -> Notifier). || Provides a robust, automated workflow for continuous data processing. || Automated monitoring, data extraction, and reporting pipelines.
|}

== HPCMATE Openclaw Agents ==

{| class="wikitable"
|+
!Agent Group
!Type
!Description
!Details
|-
|Baremetal System Configuration
|Orchestrator
|We use this agent to setup general purpose bare metal system configuration
Such as DLS(Deep Learning Station), AIS (AI Station) for each node
|TBD
|-
|Cluster Configuration
|
|We use this agent to setup a cluster master and slave for micro datacenter
|TBD
|-
|
|
|
|
|}

== Reference ==
<references />

Openclaw Agents

2026-06-03T04:00:36Z

Admin: /* Type of Agents (Openclaw) */

Openclaw Agents

2026-06-03T03:59:36Z

Admin: /* Type of Agents (Openclaw) */

== Type of Agents (Openclaw) ==

{| class="wikitable sort0"
|+ [[Generated by HPCMATE local AI|Comparison of Multiple Agent Configuration Strategies in OpenClaw]]
! Strategy !! Description !! Advantages !! Best Use Case
|-
| '''Strategy ①: Specialized Individual Agents ([[Library]])''' || Create separate .md files for each role, such as coder.md, tester.md, or reviewer.md. || Simple to implement and very easy to manage individual roles. || Quick, single-purpose tasks or simple automation.
|-
| '''Strategy ②: Orchestrator Pattern (Manager-Worker)''' || A "Manager" agent analyzes the user's request and uses sessions_spawn to delegate sub-tasks to specific agents. || High autonomy; can handle complex, multi-step reasoning without human intervention. || Complex software development or multi-stage project management.
|-
| '''Strategy ③: Pipeline Configuration (Sequential)''' || A chain of agents where the output of one agent serves as the input for the next (e.g., Scraper -> Analyzer -> Notifier). || Provides a robust, automated workflow for continuous data processing. || Automated monitoring, data extraction, and reporting pipelines.
|}

== HPCMATE Openclaw Agents ==

== Reference ==
<references />

Openclaw Agents

2026-06-03T03:58:28Z

Admin: /* Type of Agents (Openclaw) */

== Type of Agents (Openclaw) ==

{| class="wikitable sort0"
|+ Comparison of Multiple Agent Configuration Strategies in OpenClaw
! Strategy !! Description !! Advantages !! Best Use Case
|-
| '''Strategy ①: Specialized Individual Agents ([[Library]])''' || Create separate .md files for each role, such as coder.md, tester.md, or reviewer.md. || Simple to implement and very easy to manage individual roles. || Quick, single-purpose tasks or simple automation.
|-
| '''Strategy ②: Orchestrator Pattern (Manager-Worker)''' || A "Manager" agent analyzes the user's request and uses sessions_spawn to delegate sub-tasks to specific agents. || High autonomy; can handle complex, multi-step reasoning without human intervention. || Complex software development or multi-stage project management.
|-
| '''Strategy ③: Pipeline Configuration (Sequential)''' || A chain of agents where the output of one agent serves as the input for the next (e.g., Scraper -> Analyzer -> Notifier). || Provides a robust, automated workflow for continuous data processing. || Automated monitoring, data extraction, and reporting pipelines.
|}

== HPCMATE Openclaw Agents ==

== Reference ==
<references />

Openclaw Agents

2026-06-03T03:57:18Z

Admin: /* Type of Agents (Openclaw) */

== Type of Agents (Openclaw) ==

{| class="wikitable sortable"
|+ Comparison of Multiple Agent Configuration Strategies in OpenClaw
! Strategy !! Description !! Advantages !! Best Use Case
|-
| '''Strategy ①: Specialized Individual Agents ([[Library]])''' Create separate .md files for each role, such as coder.md, tester.md, or reviewer.md. Simple to implement and very easy to manage individual roles. Quick, single-purpose tasks or simple automation.
|-
| '''Strategy ②: Orchestrator Pattern (Manager-Worker)''' A "Manager" agent analyzes the user's request and uses sessions_spawn to delegate sub-tasks to specific agents. High autonomy; can handle complex, multi-step reasoning without human intervention. Complex software development or multi-stage project management.
|-
| '''Strategy ③: Pipeline Configuration (Sequential)''' A chain of agents where the output of one agent serves as the input for the next (e.g., Scraper $\rightarrow$ Analyzer $\rightarrow$ Notifier). Provides a robust, automated workflow for continuous data processing. || Automated monitoring, data extraction, and reporting pipelines.
|}

== HPCMATE Openclaw Agents ==

== Reference ==
<references />

Openclaw Agents

2026-06-03T03:55:05Z

Admin:

== Type of Agents (Openclaw) ==
| 구성 방식 | 난이도 | 특징 | 추천 상황 |
| ------- | --- | -------------------- | -------------------------- |
| 개별 에이전트 | 낮음 | 단순 호출, 독립적 작업 | 특정 기능(코드 작성 등)만 필요할 때 |
| 오케스트레이터 | 높음 | 자율적 업무 배분, 자동화된 팀 구성 | 복잡한 소프트웨어 개발 프로젝트 전체를 맡길 때 |
| 파이프라인 | 중간 | 데이터 흐름 중심, 정해진 순서 | 주기적인 데이터 처리(Batch) 작업 시 |

== HPCMATE Openclaw Agents ==

== Reference ==
<references />

Openclaw Agents

2026-06-03T03:54:12Z

Admin: Created page with "== Type of Agents (Openclaw) == == HPCMATE Openclaw Agents == == Reference == <references />"

== Type of Agents (Openclaw) ==

== HPCMATE Openclaw Agents ==

== Reference ==
<references />

U4008

2026-05-14T05:33:10Z

Admin: /* U4008 Highlight */

== HPCMATE U4008 ==

U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server.

[[File:U4008.png|center|frameless|446x446px]]

In brief : U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server

== U4008 Highlight ==
{| class="wikitable"
|+
!
!Default
!Max Configuration
!Options
|-
|Power
|2000W Platinum Modular PSU
|Configurable up to 5600W
|Dual 2200W '''Titanium''' CRPS
Automatic switching between 1+1 and 2+0
|-
|Storage
|3 x 5.15" dummy cage
|Any type of 5.25" hard rack
|Depending on models
|-
|Mother Board
|From Mini-ITX up to EEB (13"x14")
|
|Optional [[PCIe]] 5.0 x16 extend board
|-
| rowspan="2" |Cooling
|Air cooling
|[[Support]] up to dual 500W CPUs
Support up to 5 x DP GPUs
|
|-
|[[Liquid cooling|Liquid Cooling]]
|Support up to dual 500W CPUs
Support up to 8 x SP GPUs
|Compatible with [[CoolUX]]
|}

== U4008 Models ==
{| class="wikitable"
|+
!Model
!Socket
!DIMM slots
!M.2 slots
!Max GPU
(double width)
!Cooling
!Power
!Others
|-
|2KHR0
|Single Socket SP5 (LGA 6096), AMD EPYC™ 9004 (TDP up to 500W)

|8 DIMM slots
|2
|4
|Air
|2000W
|7x PCIe5.0 x16,
1 x PCIe5.0 x8

3 x MCIO (PCIe5.0 x4 or 4 SATA 6Gb/s)

2 x RJ45 (10GbE)

1 x [[IPMI]]

4 x hot swap 2.5" SATA drives
|-
|3KIHR1
|Single Intel LGA 4710(Socket E2), Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor (TDP up to 350W)
|8 DIMM slots
|2
|5
|Air
|4400W
|5 x PCIe5.0 x16,
6 x MCIO (PCIe5.0 x8 R1S CPU)

2 x RJ45 (1GbE)

1 x [[IPMI]]

4 x hot swap 2.5" SATA drives
|-
|
|TBD more models
|
|
|
|
|
|
|}

== Contact ==
Mail us for more details sales@hpcmate.com

== References ==
<references />
[[Category:Article]]
[[Category:Hpcmate]]
[[Category:Tech]]
[[Category:Products]]

Node.js tips and tricks

2026-04-14T05:11:31Z

Admin:

=== Update nodejs package.json ===
<syntaxhighlight lang="bash">
# Install npm-check-updates globally
$ npm i -g npm-check-updates

# Inside of nodejs project folder where package.json is located
$ ncu -u
Upgrading /opt/imp/home/prj/package.json
[====================] 39/39 100%
...
async ^2.6.3 → ^3.2.6
cheerio ^1.0.0-rc.2 → ^1.0.0
cookie-parser ^1.4.4 → ^1.4.7
....

# Run npm install to install new versions.
$ npm install
</syntaxhighlight>

== Install NVM ==
<syntaxhighlight lang="bash">
curl https://raw.githubusercontent.com/creationix/nvm/master/install.sh | bash
source ~/.bashrc

# List up node version
nvm list-remote

# Install specific version
nvm install 18.0.0

# List of installed versions
nvm ls

# Change version
nvm use 16.15.0

# Update npm
npm install -g npm@latest

</syntaxhighlight>

== Install latest Node.js<ref>https://deb.nodesource.com/</ref> ==
<syntaxhighlight lang="bash">
#Install Node.js 20.x LTS
$curl -fsSL https://deb.nodesource.com/setup_20.x | sudo bash -
$sudo apt-get install -y nodejs

</syntaxhighlight>

== Change node version ==
<syntaxhighlight lang="bash">
# Get available version list using nvm
$ nvm ls
v4.9.1
v6.17.1
v8.8.1
v8.9.4
v8.17.0
v10.24.1
-> v14.15.1
v16.20.0
v18.16.0
system
default -> v14.15.1

# change version v18 temporary and check the changing version
$ nvm use v18.16.0
$ nvm ls
v4.9.1
v6.17.1
v8.8.1
v8.9.4
v8.17.0
v10.24.1
v14.15.1
v16.20.0
-> v18.16.0
system
default -> v14.15.1 << default version will not be changed

# change default version v18 sytemwide
$ nvm alias default v18
default -> v18 (-> v18.16.0)

$ nvm ls
v4.9.1
v6.17.1
v8.8.1
v8.9.4
v8.17.0
v10.24.1
v14.15.1
v16.20.0
-> v18.16.0
system
default -> v18 (-> v18.16.0) << default version has been changed
node -> stable (-> v18.16.0) (default)
stable -> 18.16 (-> v18.16.0) (default)

# Install new version of node
$ nvm install <version you want to>
</syntaxhighlight>

== References ==
<references />

Intel Xeon 6 processor

2026-03-25T03:13:55Z

Admin: Created page with " == Xeon 6 R1S features == Intel Xeon 6 R1S is a specialized, single-socket processor within the Granite Rapids family uses the LGA 4710 socket, optimized for high-I/O workloads like AI inference, edge computing, and storage. It offers up to 136 PCIe Gen5 lanes - significantly more than the typical 88-96 lanes on other models - providing superior connectivity compared to standard. == References == <references />"

== Xeon 6 R1S features ==
Intel Xeon 6 R1S is a specialized, single-socket processor within the Granite Rapids family uses the LGA 4710 socket, optimized for high-I/O workloads like AI inference, edge computing, and storage. It offers up to 136 [[PCIe]] Gen5 lanes - significantly more than the typical 88-96 lanes on other models - providing superior connectivity compared to standard.

== References ==
<references />

U4008

2026-03-17T03:45:22Z

Admin: /* U4008 Highlight */

== HPCMATE U4008 ==

U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server.

[[File:U4008.png|center|frameless|446x446px]]

In brief : U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server

== U4008 Highlight ==
{| class="wikitable"
|+
!
!Default
!Max Configuration
!Options
|-
|Power
|2000W Platinum Modular PSU
|Configurable up to 5600W
|2+0 CPS hot plug PSU (Max. 4000W)
|-
|Storage
|3 x 5.15" dummy cage
|Any type of 5.25" hard rack
|Depending on models
|-
|Mother Board
|From Mini-ITX up to EEB (13"x14")
|
|Optional [[PCIe]] 5.0 x16 extend board
|-
| rowspan="2" |Cooling
|Air cooling
|[[Support]] up to dual 500W CPUs
Support up to 5 x DP GPUs
|
|-
|[[Liquid cooling|Liquid Cooling]]
|Support up to dual 500W CPUs
Support up to 8 x SP GPUs
|Compatible with [[CoolUX]]
|}

== U4008 Models ==
{| class="wikitable"
|+
!Model
!Socket
!DIMM slots
!M.2 slots
!Max GPU
!Cooling
!Power
!Others
|-
|2KHR0
|Single Socket SP5 (LGA 6096), AMD EPYC™ 9004 (TDP up to 500W)

|8 DIMM slots
|2
|4 x double width GPUs
|Air
|2000W
|7x PCIe5.0 x16,
1 x PCIe5.0 x8

3 x MCIO (PCIe5.0 x4 or 4 SATA 6Gb/s)

2 x RJ45 (10GbE)

1 x [[IPMI]]

4 x hot swap 2.5" SATA drives
|-
|3KIHR1
|Single Intel LGA 4710(Socket E2), Intel® Xeon® 6700E-series, 6500P-series and 6700P-series processor (TDP up to 350W)
|8 DIMM slots
|2
|5 x double width GPUs
|Air
|3000W
|5 x PCIe5.0 x16,
6 x MCIO (PCIe5.0 x8 R1S CPU)

2 x RJ45 (1GbE)

1 x [[IPMI]]

4 x hot swap 2.5" SATA drives
|}

== Contact ==
Mail us for more details sales@hpcmate.com

== References ==
<references />
[[Category:Article]]
[[Category:Hpcmate]]
[[Category:Tech]]
[[Category:Products]]

RIMA

2026-02-27T00:36:36Z

Admin: /* Settings > Device */

== RIMA ==
RIMA (Rack Information Management Archive) is an appliance (Hardware plus Software) product by HPCMATE. RIMA provides real-time data monitoring for multiple RIM (DS7) devices and report features. ([[Private:RIMA|Internal RIMA document]])

== Setup guide ==

# Install RIMA 1U system at target place
# Setup network depending on customer's network environment - refer to below System Admin > System IP configuration

== Access to RIMA ==

=== Over VGA ===
VGA interface of RIMA system provides RIMA Web pages on system boot-up.

=== Over network ===
RIMA provides http Web interface through '''''<host-ip>:8099''''' port.

== Dashboard ==
The dashboard provides a simple summary for the registered RIM devices.

Dashboard information list are

* Total power consumption at now
* Cumulative total power consumption so far
* Maximum power consumption RIM and it's name
* Highest temperature RIM and it's name
* Time series data graph for current power consumption and highest temperature
* Registered RIM wheel and selected RIM information

=== Automatic wheel rotation ===
To show multiple RIM devices effectively, each registered RIM device wheel rotates in every 5 seconds. the automatic rotation to stop when RIMA detect user action on the web session such as mouse move, click, key press and resume automatic rotation after 3 seconds of non user action.

(** The local VGA display will not affect to other user session)
[[File:Dashboard-rima.png|center|frameless|676x676px]]

=== System Reboot or Shutdown ===
RIMA dashboard supports system reboot or shutdown using the power icon on top right.
[[File:Rebootorshutdown.png|center|frameless|394x394px]]

=== Admin login / logout ===
System/Settings/Reports menu is only available after successful Admin login.

The factory default Admin password is 'Admin'. Admin password can be changed via Settings > Admin tab.

<big>'''''*** There is no way to recover Admin password once it has been changed. So please do not forget once it has been changed !!!'''''</big>

[[File:Rima-loginout.png|center|frameless|475x475px]]

== System ==

=== System tab ===
[[File:System-system.png|frameless|943x943px]]

System tab provides real time system resources information in every 3 seconds as well as provided license information including maximum possible RIM(DS7) devices depending on the license type. license section describes RIMA license type and allowable maximum devices.

=== Update tab ===
RIMA system package can be updated through http://update.hpcmate.com. to do so, RIMA system must be able to connect internet.

Update button to check and show the latest available RIMA system package.

[[File:Rima system update.png|center|frameless|645x645px]]

== Settings ==

=== Admin tab ===
[[File:Settings-admin.png|frameless|932x932px]]

=== Admin Password ===
RIMA supports only one administrator account for simple and operation ownership and responsibility.

Factory default Admin's ID and password are "Admin" and "Admin". we recommend to change more secure password in production operation.

=== System IP ===

* RIMA system always supports default IP - 192.168.1.253/24 - to ensure system configuration by Admin at any IP environment
** For the initial setup, please use the default IP with direct connect to your notebook by changing 192.168.1.0/24 networks
* To Set up IP address for your environment, use the System IP menu
** Admin could set comma seperated multiple IP address using CIDR format. the first CIDR information will be used for RIMA system and others to point RIM(DS-7) devices

=== Notification Mail ===
RIMA supports real-time notification message using gmail app password to registered email list. To do that Admin needs to make Gmail's App Password. Following is an overview how to create App password and register

# Log in to your Google account or use this link <nowiki>https://security.google.com/settings/security/apppasswords</nowiki>
# Go to My Account > Sign-in & Security > App Passwords
# Put App name whatever you want then create app password
# Copy the password then put it as sendmail password along with your gmail account
# [[Test]] send mail and verify mail receiving from the RIMA

[[File:RIMA-gmail-app-passowrd.png|center|frameless]]

=== Settings > Device ===
Settings > Device tap to be used to register single or multiple RIM device into the RIMA appliance.

=== Add single RIM ===
Put IP address of RIM then click test action button, then RIMA try to find give RIM information through the network and display RIM's basic information with connected status (blue circle) with blue MAC address on success or (magenta circle) on fail.

'''''<big>*** Because RIMA use only RIM name, each RIM device should have different name regardless of its IP configuration. (Please refer to RIM manual to change RIM device name).</big>'''''

<big>*** MAC Address should be detected by the test button otherwise RIMA does not care the given row information</big>

[[File:RIMA-devices.png|center|frameless|787x787px]]

=== Delete device ===
the remove button handle GUI row only, to delete device from the database, Admin must also apply by "Save DB"

=== Device > Settings > import ===
To add multiple RIM (DS-7) device into the device table, Admin is able to put csv format excel file using this button.

The CSV format must be the same with following example, All string is lower case and name and ip must have fields while mac and status does not used in current version of RIMA
{| class="wikitable"
|name
|ip
|mac
|status
|-
|RIM 1
|10.230.200.1/24
|
|
|-
|RIM 2
|10.230.200.2/24
|
|
|-
|RIM 3
|10.230.200.3/24
|
|
|-
|RIM 4
|10.230.200.4/24
|
|
|-
|
|
|
|
|}

=== Device > Settings > Export ===
To save current device list, Admin can export current device list using this menu.

'''''*** RIMA exports only validated IP and MAC devices'''''
{| class="wikitable"
|name
|ip
|mac
|status
|-
|RID-1234 (192.168.1.4/24, 10.230.200.1/24)
|192.168.1.4
|00:15:5d:01:df:1a
|
|-
|
|
|
|
|}

=== Device > Settings > Add(DS7) ===
Admin manually add RIM(DS 7) devices and verify device connection using the test button.

'''''*** Only licensed number of device can be added'''''

=== Settings > Threshold ===
RIMA supports automatic notification email using above notification mail to monitor each RIM's power consumption, temperature and Humidity.

Admin enable or disable real-time notification by setting real-time mail notification toggle button.
[[File:Settings threshold.png|center|frameless|876x876px]]

== Reports ==
RIMA supports overview reports up to 1 years from at the time of click this menu.

[[File:Rima-reports.png|center|frameless|577x577px]]

== License Type ==
As an appliance type of product, RIMA product has its own license type. following are the license type and maximum allowable device supports
{| class="wikitable"
|+
!License Type
!Max. supported RIM (DS7) devices
!Notes
!data retension policy
!Features
|-
|L1
|[[support]] up to 12 RIM devices
|1TB x 4ea for data storage, usable 3TB over RAID5
|1 year
|Default
|-
|L2
|support up to 24 RIM devices
|2TB x 4ea for data storage, usable 5TB over RAID5
|1 year
|Default
|-
|L3
|support up to 48 RIM devices
|4TB x 4ea for data storage, usable 8TB over RAID6
|1 year
|Adjust default action
|-
|L4
|support up to 100 RIM devices
|8TB x 4ea for data storage, usable 16TB over RAID6
|1 year
|Adjust default action
|}

== References ==
<references />

U4008

2026-02-25T05:12:58Z

Admin: /* U4008 Models */

== HPCMATE U4008 ==

U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server.

[[File:U4008.png|center|frameless|446x446px]]

In brief : U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server

== U4008 Highlight ==
{| class="wikitable"
|+
!
!Default
!Option
!Notices
|-
| rowspan="2" |Power
|Single ATX 3.1 modular PSU (2000W)
|Max. 5600W by using 2+0 dual 2800W PSU
|
|-
|
|2+0 modular hot plug PSU
|Max. 4000W
|-
|Storage
|3 x 5.15" dummy cage
|Any type of 5.25" hard rack
|Partner with ICY Dock
|-
|Mother Board
|[[Support]] from Mini-ITX up to EEB (13"x14")
|Support extended [[PCIe]] 5.0 x 16 board
|Use 2 x MCIO 8i
|-
| rowspan="2" |Cooling
|Air cooling
|
|Support up to dual 500W CPUs
|-
|
|[[Liquid cooling|Liquid Cooling]]
|Compatible with [[CoolUX]]
|}

== U4008 Models ==
{| class="wikitable"
|+
!Model
!Socket
!DIMM slots
!M.2 slots
!Max GPU
!Cooling
!Power
!Others
|-
|2KHR0
|Single Socket SP5 (LGA 6096), AMD EPYC™ 9004 (TDP up to 500W)

|8 DIMM slots
|2
|4 x double width GPUs
|Air
|2000W
|4 x PCIe5.0 / CXL2.0 x16,
3 x PCIe5.0 x16
1 x PCIe5.0 x8
3 x MCIO (PCIe5.0 x4 or 4 SATA 6Gb/s)

1 x MCIO* (PCIe5.0 x4)

2 x RJ45 (10GbE)

1 x [[IPMI]]

4 x hot swap 2.5" SATA drives
|-
|TBA
|
|
|
|
|
|
|
|}

== Contact ==
Mail us for more details sales@hpcmate.com

== References ==
<references />
[[Category:Article]]
[[Category:Hpcmate]]
[[Category:Tech]]
[[Category:Products]]

U4008

2026-02-25T05:11:47Z

Admin: /* U4008 Models */

== HPCMATE U4008 ==

U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server.

[[File:U4008.png|center|frameless|446x446px]]

In brief : U4008 has been designed to provide a flexible system configuration for multi purpose high performance computer server

== U4008 Highlight ==
{| class="wikitable"
|+
!
!Default
!Option
!Notices
|-
| rowspan="2" |Power
|Single ATX 3.1 modular PSU (2000W)
|Max. 5600W by using 2+0 dual 2800W PSU
|
|-
|
|2+0 modular hot plug PSU
|Max. 4000W
|-
|Storage
|3 x 5.15" dummy cage
|Any type of 5.25" hard rack
|Partner with ICY Dock
|-
|Mother Board
|[[Support]] from Mini-ITX up to EEB (13"x14")
|Support extended [[PCIe]] 5.0 x 16 board
|Use 2 x MCIO 8i
|-
| rowspan="2" |Cooling
|Air cooling
|
|Support up to dual 500W CPUs
|-
|
|[[Liquid cooling|Liquid Cooling]]
|Compatible with [[CoolUX]]
|}

== U4008 Models ==
{| class="wikitable"
|+
!Model
!Socket
!DIMM slots
!M.2 slots
!Max GPU
!Cooling
!Power
!Others
|-
|2KHR0
|Single Socket SP5 (LGA 6096), AMD EPYC™ 9004 (TDP up to 500W)

|8 DIMM slots
|2
|Supports 4 x double width GPUs
|Air
|2000W
|4 PCIe5.0 / CXL2.0 x16*, 3 PCIe5.0 x16, 1 PCIe5.0 x8
3 MCIO (PCIe5.0 x4 or 4 SATA 6Gb/s), 1 MCIO* (PCIe5.0 x4)
2 RJ45 (10GbE), 1 [[IPMI]]
4 hot swap 2.5" SATA drives
|-
|TBA
|
|
|
|
|
|
|
|}

== Contact ==
Mail us for more details sales@hpcmate.com

== References ==
<references />
[[Category:Article]]
[[Category:Hpcmate]]
[[Category:Tech]]
[[Category:Products]]

Bifurcation

2026-02-25T05:07:11Z

Admin: Created page with " == PCIe Bifurcation == PCIe bifurcation is a motherboard and CPU feature that splits a single physical PCIe slot (usually x16) into multiple, smaller, independent logical lanes (e.g., two x8, four x4). It enables running multiple devices, such as NVMe SSDs or GPUs, from one slot. It's not one of the things that can boost your frame rates, affect your CPU or GPU performance<ref>https://www.xda-developers.com/pcie-bifurcation-most-underrated-pc-feature-nobody-checks-for/<..."

== PCIe Bifurcation ==
[[PCIe]] bifurcation is a [[motherboard]] and CPU feature that splits a single physical PCIe slot (usually x16) into multiple, smaller, independent logical lanes (e.g., two x8, four x4). It enables running multiple devices, such as NVMe SSDs or GPUs, from one slot. It's not one of the things that can boost your frame rates, affect your CPU or GPU performance<ref>https://www.xda-developers.com/pcie-bifurcation-most-underrated-pc-feature-nobody-checks-for/</ref>

For example, you can divide a PCIe x16 slot into multiples of eight or four, which means it can [[support]] the following configurations:

* x8 / x8
* x8 / x4 / x4
* x4 / x4 / x4 / x4

Each lane 'split' can behave like an independent PCIe connection if your board supports it. Primarily used to install multiple M.2 NVMe drives via an adapter card, or to split a GPU slot to accommodate other peripheral cards.

Requires support from both the motherboard ([[BIOS]]/UEFI) and the CPU. '''Bifurcation''' divides lanes in the BIOS/UEFI, allowing the motherboard/CPU(actual bifurcation occurs at the CPU and chipset levels) to recognize and manage each split group as a separate device. That's why not all motherboards or BIOS/UEFI versions support PCIe Bifurcation. common in most high-end boards, it's not unusual for mid-range boards.

With two MCIO to PCIe adapter, like 2x MCIO 8i (74-pin) to PCIe x16 adapters to enable connecting high-speed PCIe 5.0 devices (GPUs/NVMe) via two SFF-9402 MCIO cables, providing full x16 bandwidth (x8+x8), These adapters often feature a 6-pin PCIe power connector, with the features of motherboard bifurcation support (x8x8)

== Examples ==
This [https://global.icydock.com/resources/icy_tips_1462.html ICY Dock URL] shows an example how to setup PCIe Bifurcation in BIOS

== References ==
<references />