Skip to content

node-info 0.28.0: Axera AX650N rows, R620+ GPU temperature, stable DXGI out-params - #41

Merged
dmmdea merged 3 commits into
mainfrom
feat/node-info-axera
Nov 8, 2026
Merged

dmmdea merged 3 commits into
mainfrom
feat/node-info-axera

Conversation

@dmmdea

@dmmdea dmmdea commented Nov 8, 2026

Copy link
Copy Markdown
Owner

node-info 0.28.0: three node-info changes, each measured on real hardware.

1. Axera AX650N accelerator rows (Linux)

An AX650N-based card (e.g. the M.2 "AX8850"; PCI 1f4b:0650) is listed in GPUs[] as an accelerator row (kind: "npu", inference_ready: false, never counted in telemetryValid or scheduling pressure), the same way Coral and RKNPU rows are.

  • Source: one axcl-smi table call every 5 s (5 s timeout, never overlapping; last good table used for at most 20 s). The table's NPU column is the truthful utilization source (info --npu prints -- at idle and a different line shape under load); a -- in the column is no reading, never 0.
  • Memory = the CMM pool (the NPU/media pool), not the card OS RAM. Temperature = the SoC sensor. Power: the card reports none (ceiling, documented).
  • Without the vendor CLI the card still appears from the PCI scan, by name only.
  • Measured (AXCL 3.16.0, YOLO11s ax650 model in axcl_run_model): idle 0 % → under load 77 / 93 / 91 / 93 % → idle; CMM used 18 → 34 MiB; temperature 35 → 37 °C, matching axcl-smi info --temp to the degree; the host's other rows were byte-identical to the running service.

2. GPU temperature on R620+ NVIDIA drivers (Windows + Linux)

NVIDIA-SMI 621.14 prints temperature.gpu with decimals (78.76). The integer-only parse rejected every row, so every NVIDIA card on such a host showed no temperature while power and utilization still worked. parseCelsius (beside parseWatts) accepts whole and fractional degrees and rounds.

  • Before: three cards, no temperature_celsius on any. After (same cards, same minute): 49 / 60 / 79 °C vs nvidia-smi 49.8 / 60.3 / 77.7 °C.

3. Stable DXGI out-parameters (Windows)

vtableCall takes out-parameters as uintptr(unsafe.Pointer(&local)). Without //go:uintptrescapes the local stays on the goroutine stack; when the stack grows inside the call the runtime moves it and DXGI writes to the old address, so the caller sees S_OK with a nil adapter: the recurring warning IDXGIFactory1::EnumAdapters1 failed; stopping enumeration idx=0 hr=0x00000000, which empties that pass's adapter list and temperature join.

  • go build -gcflags=-m: before adapter does not escape, after moved to heap: adapter (same directive as syscall.(*LazyProc).Call).
  • Same 8-minute window, same host: running 0.27.1 logged the warning 6 times, the fixed build 0.

Gates

  • go test . native Windows: ok. go vet linux/amd64, linux/arm64, darwin/arm64, windows/amd64: ok.
  • Linux amd64 test binary (cross-compiled) run on an AX650N + 2-GPU host: PASS.
  • No new tests (fork policy); the checks above are real runs.

🤖 Generated with Claude Code

dmmdea and others added 3 commits November 7, 2026 22:48
Detect PCI vendor 0x1f4b cards and list each as a kind "npu" accelerator row,
not inference-ready and outside GPU telemetry, like the Coral and RKNPU rows.
A single axcl-smi table call every 5 s (5 s timeout, never overlapping) feeds
the card name, CMM pool capacity and usage, NPU utilization and temperature;
the last good table is published for at most 20 s. A "--" NPU reading is
reported as utilization_unavailable, never as 0. Without the CLI the card is
still listed from PCI. Power is not reported by the card and is omitted.

gpuStat gains VRAMTotal so a sampler can supply a capacity the static
inventory does not know.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Signed-off-by: dmmdea <4965968+dmmdea@users.noreply.github.com>
…arams

Two Windows node-info defects seen on a three-card host:

- No GPU temperature at all. NVIDIA-SMI 621.14 prints temperature.gpu
  with decimals ("78.76"); the integer-only parse rejected every row, so
  every card lost its temperature while power and utilization (parsed as
  floats/ints) stayed. parseCelsius (beside parseWatts) accepts whole and
  fractional degrees and rounds; the Linux collector uses it too.

- Intermittent "IDXGIFactory1::EnumAdapters1 failed; stopping enumeration
  idx=0 hr=0x00000000" (hundreds per day). vtableCall receives
  out-parameters as uintptr(unsafe.Pointer(&local)); without
  //go:uintptrescapes the local stays on the goroutine stack, and a stack
  move inside the call makes DXGI write to the old address, so the caller
  sees S_OK with a nil adapter and the pass reports no adapters (the
  temperature join and inventory refresh then work from an empty list).
  go build -gcflags=-m: before "adapter does not escape", after "moved to
  heap: adapter". Same directive as syscall.(*LazyProc).Call.

Signed-off-by: dmmdea <4965968+dmmdea@users.noreply.github.com>
…re fix

Signed-off-by: dmmdea <4965968+dmmdea@users.noreply.github.com>
@dmmdea
dmmdea merged commit e3fca8d into main Nov 8, 2026
12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant