- Bootloader: GRUB fallback text colors → yellow/brown (amber tone)
- CLI charts: all GPU metric series use single amber color (xterm-256 #214)
- Wallpaper: logo width scaled to 400 px dynamically, shadow scales with font size
- Support bundle: renamed to YYYY-MM-DD (BEE-SP vX.X) SRV_MODEL SRV_SN ToD.tar.gz
using dmidecode for server model (spaces→underscores) and serial number
- Remove display resolution feature (UI card, API routes, handlers, tests)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
HPL 2.3 from netlib compiled against OpenBLAS with a minimal
single-process MPI stub — no MPI package required in the ISO.
Matrix size is auto-sized to 80% of total RAM at runtime.
Build:
- VERSIONS: HPL_VERSION=2.3, HPL_SHA256=32c5c17d…
- build-hpl.sh: downloads HPL + OpenBLAS from Debian 12 repo,
compiles xhpl with a self-contained mpi_stub.c
- build.sh: step 80-hpl, injects xhpl + libopenblas into overlay
Runtime:
- bee-hpl: generates HPL.dat (N auto from /proc/meminfo, NB=256,
P=1 Q=1), runs xhpl, prints standard WR... Gflops output
- platform/hpl.go: RunHPL(), parses WR line → GFlops + PASSED/FAILED
- tasks.go: target "hpl"
- pages.go: LINPACK (HPL) card in validate/stress grid (stress-only)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Replace diag level 1-4 dropdown with Validate/Stress radio buttons
- Validate: dcgmi L2, 60s CPU, 256MB/1p memtester, SMART short
- Stress: dcgmi L3 + targeted_stress in Run All, 30min CPU, 1GB/3p memtester, SMART long/NVMe extended
- Parallel GPU mode: spawn single task for all GPUs instead of splitting per model
- Benchmark table: per-GPU columns for sequential runs, server-wide column for parallel
- Benchmark report converted to Markdown with server model, GPU model, version in header; only steady-state charts
- Fix IPMI power parsing in benchmark (was looking for 'Current Power', correct field is 'Instantaneous power reading')
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds a "Multi-GPU tests — use all GPUs" checkbox to the NVIDIA GPU
selector (checked by default). When enabled, PSU Pulse, NCCL, and
NVBandwidth tests run on every GPU in the system regardless of the
per-GPU selection above — which is required for correct PSU stress
testing (synchronous pulses across all GPUs create worst-case
transients). When unchecked, only the manually selected GPUs are used.
The same logic applies both to Run All (expandSATTarget) and to the
individual Run button on each multi-GPU test card.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
pulse_test is a PSU/power-delivery test, not a per-GPU compute test.
Its purpose is to synchronously pulse all GPUs between idle and full
load to create worst-case transient spikes on the power supply.
Running it one GPU at a time would produce a fraction of the PSU load
and miss any PSU-level failures.
- Move nvidia-pulse from nvidiaPerGPUTargets to nvidiaAllGPUTargets
(same dispatch path as NCCL and NVBandwidth)
- Change card onclick to runNvidiaFabricValidate (all selected GPUs at once)
- Update card title to "NVIDIA PSU Pulse Test" and description to
explain why synchronous multi-GPU execution is required
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
nvidia-interconnect (NCCL all_reduce_perf) and nvidia-bandwidth
(NVBandwidth) verify fabric connectivity and bandwidth — they are
not sustained burn loads. Move both from the Burn section to the
Validate section under the stress-mode toggle, alongside the other
DCGM diagnostic tests moved in the previous commit.
- Add sat-card-nvidia-interconnect and sat-card-nvidia-bandwidth
validate cards (stress-only, all selected GPUs at once)
- Add runNvidiaFabricValidate() for all-GPU-at-once dispatch
- Add nvidiaAllGPUTargets handling in expandSATTarget/runAllSAT
- Remove Interconnect / Bandwidth card from Burn section
- Remove nvidia-interconnect and nvidia-bandwidth from runAllBurnTasks
and the gpu/tools availability map
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- bee-gpu-stress.c: remove per-wave cuCtxSynchronize barrier in both
cuBLASLt and PTX hot loops; sync at most once/sec so the GPU queue
stays continuously full — eliminates the CPU↔GPU ping-pong that
prevented reaching full TDP
- sat_fan_stress.go: default SizeMB 0 (auto = 95% VRAM) instead of
hardcoded 64 MB; tiny matrices caused <0.1 ms kernels where CPU
re-queue overhead dominated
- pages.go: move nvidia-targeted-power and nvidia-pulse from Burn →
Validate stress section alongside nvidia-targeted-stress; these are
DCGM pass/fail diagnostics, not sustained burn loads; remove the
Power Delivery / Power Budget card from Burn entirely
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
PowerSustainScore now uses DefaultPowerLimitW as reference so a
manually reduced power limit does not inflate the score. Falls back
to enforced limit if default is unavailable.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Sample server power (IPMI dcmi) during baseline+steady phases in parallel;
compute delta vs GPU-reported sum; flag ratio < 0.75 as unreliable reporting
- Collect base_graphics_clock_mhz, multiprocessor_count, default_power_limit_w
from nvidia-smi alongside existing GPU info
- Add tops_per_sm_per_ghz efficiency metric (model-agnostic silicon quality signal)
- Flag when enforced power limit is below default TDP by >5%
- Add fp64 profile to bee-gpu-burn worker (CUDA_R_64F, CUBLAS_COMPUTE_64F, min cc 8.0)
- Improve Executive Summary: overall pass count, FAILED GPU finding
- Throttle counters now shown as % of steady window instead of raw microseconds
- bible-local: clock calibration research, H100/H200 spec, real-world GEMM baselines
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
losetup --replace --direct-io=on fails with EINVAL when the target file
is on tmpfs (/dev/shm), because tmpfs does not support O_DIRECT.
Strip the --direct-io flag from the replace call and downgrade the
verification failure to a warning so boot continues.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Servers with NVIDIA compute GPUs (H100 etc.) have no display output,
so KMS blanks the console. nomodeset disables kernel modesetting and
lets the NVIDIA proprietary driver handle display via Xorg.
KMS variant moved to advanced submenu for cases where it is needed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- bee-nvidia-load: run insmod in background, poll /proc/devices for
nvidiactl; if GSP init doesn't complete in 90s, kill insmod and retry
with NVreg_EnableGpuFirmware=0. Handles EBUSY case with clear error.
- Write /run/bee-nvidia-mode (gsp-on/gsp-off/gsp-stuck) for audit layer
- Show GSP mode badge in sidebar: yellow for gsp-off, red for gsp-stuck
- Report NvidiaGSPMode in RuntimeHealth with issue entries
- Simplify GRUB menu: default (KMS+GSP), advanced submenu (GSP=off,
nomodeset, fail-safe), remove load-to-RAM entry
- Add pcmanfm, ristretto, mupdf, mousepad to desktop packages
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- services.go: use sudo systemctl so bee user can control system services
- api.go: always return 200 with output field even on error, so the
frontend shows the actual systemctl message instead of "exit status 1"
- pages.go: button shows "..." while pending then restores label;
output panel is full-width under the table with ✓/✗ status indicator;
output auto-scrolls to bottom
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
bee-gpu-burn now receives --seconds <LoadSec> so it exits naturally
when the cycle ends, rather than relying solely on context cancellation
to kill it. Process group kill (Setpgid+Cancel) is kept as a safety net
for early cancellation (user stop, context timeout). Same fix for AMD
RVS which now gets duration_ms = LoadSec * 1000.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
nvvs (DCGM validation suite) survives when dcgmi is killed mid-run,
leaving the GPU occupied. The next dcgmi diag invocation then fails
with "affected resource is in use".
Two-part fix:
- Add nvvs and dcgmi to KillTestWorkers patterns so they are cleaned
up by the global cancel handler
- Call KillTestWorkers at the start of RunNvidiaTargetedStressValidatePack
to clear any stale processes before dcgmi diag runs
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
bee-gpu-burn is a shell script that spawns bee-gpu-burn-worker children.
exec.CommandContext default cancel only kills the shell parent; the worker
processes survive and keep loading the GPU indefinitely.
Fix: set Setpgid=true and a custom Cancel that sends SIGKILL to the
entire process group (-pid), same pattern already used in runSATCommandCtx.
Applied to Nvidia, AMD, and CPU stress commands for consistency.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add feh and python3-pil to package list
- Add chroot hook that generates /usr/share/bee/wallpaper.png using PIL:
black background, EASY-BEE box-drawing logo in amber (#f6c90e),
"Hardware Audit LiveCD" subtitle in dim amber — matches motd exactly
- bee-openbox-session: set wallpaper with feh --bg-fill, fall back to
xsetroot -solid black if wallpaper not found
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Set xsetroot solid background (#12100a, dark amber) so openbox
doesn't show bare black before Chromium opens
- Re-add healthz wait loop before launching Chromium: without it
Chromium opens localhost/loading before bee-web is up and gets
connection-refused which renders as a blank white page
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add TTYReset=yes and TTYVHangup=yes so systemd restores the terminal
to a clean state before handing tty1 to getty. Without this the screen
went black with no cursor after the status display finished.
Also remove DefaultDependencies=no which was too aggressive.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Command substitution in sh strips trailing newlines, so accumulating
output in a variable via $(...) lost all line breaks. Reverted to
direct printf calls which work correctly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Refresh every 3s instead of 1s to reduce flicker
- Show ssh, bee-sshsetup in service list
- Show failure reason for failed services
- Show last journal line for activating services
- Show IP addresses and web UI URL when network is up
- Render frame to variable before printing to reduce flicker
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
loglevel=6 floods the screen with mpt3sas/scsi/sd informational
messages, hiding systemd service status and bee-boot-status display.
loglevel=3 shows only kernel errors; all messages still go to serial.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add bee-boot-status service: shows live service status on tty1 with
ASCII logo before getty, exits when all bee services settle
- Remove lightdm dependency on bee-preflight so GUI starts immediately
without waiting for NVIDIA driver load
- Replace Chromium blank-page problem with /loading spinner page that
polls /api/services and auto-redirects when services are ready; add
"Open app now" override button; use fresh --user-data-dir=/tmp/bee-chrome
- Unify branding: add "Hardware Audit LiveCD" subtitle to GRUB menu,
bee-boot-status (with yellow ASCII logo), and web spinner
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
exec.CommandContext only kills the direct child (the shell script), leaving
grandchildren (john, gpu-burn, etc.) as orphans. Set Setpgid so each SAT
job runs in its own process group, then send SIGKILL to the whole group
(-pgid) in the Cancel hook.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When 8 john processes start simultaneously they race for GPU memory during
OpenCL GWS auto-tuning. Slower devices settle on a smaller work size (~594MiB
vs 762MiB) and run at 40% instead of 100% load. Add 3s sleep between launches
so each instance finishes memory allocation before the next one starts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- bee-john-gpu-stress: spawn one john process per OpenCL device in parallel
so all GPUs are stressed simultaneously instead of only device 1
- bee-openbox-session: --start-fullscreen → --start-maximized to fix blank
white page on first render in fbdev environment
- storage collector: skip Virtual HDisk* devices reported by BMC/iDRAC
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- nvidia collector: read pcie.link.gen.current/max from nvidia-smi instead
of sysfs to avoid false Gen1 readings when GPU is in ASPM idle state
- build: remove bee-nccl-gpu-stress from rm -f list so shell script from
overlay is not silently dropped from the ISO
- smoketest: add explicit checks for bee-gpu-burn, bee-john-gpu-stress,
bee-nccl-gpu-stress, all_reduce_perf
- netconf: re-exec via sudo when not root to fix RTNETLINK/resolv.conf errors
- auto/config: reduce loglevel 7→3 to show clean systemd output on boot
- auto/config: blacklist snd_hda_intel and related audio modules (unused on servers)
- package-lists: remove firmware-intel-sound and firmware-amd-graphics from
base list; move firmware-amd-graphics to bee-amd variant only
- bible-local: mark memtest ADR resolved, document working solution
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
ISO 9660 volume labels allow only A-Z, 0-9, and underscore.
Dashes cause xorriso WARNING on every build.
EASY-BEE-NVIDIA → EASY_BEE_NVIDIA (iso-application keeps dashes, it's UDF).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
memtest86+ postinst does not place files in /boot in a live-build chroot
without grub triggers. Added fallback: extract directly from the cached
.deb via dpkg-deb -x, with verbose logging throughout.
Also remove "NVIDIA no MSI-X" from boot menu (premature — root cause unknown).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
live-build picks up ALL .list.chroot files in config/package-lists/.
After rsync, bee-nvidia.list.chroot, bee-amd.list.chroot, and
bee-nogpu.list.chroot all end up in BUILD_WORK_DIR — causing lb to
try installing packages from every variant (and leaving version
placeholders unsubstituted in the unused lists).
Fix: after copying bee-${BEE_GPU_VENDOR}.list.chroot → bee-gpu.list.chroot,
delete all other bee-{nvidia,amd,nogpu}.list.chroot from BUILD_WORK_DIR.
Also includes nomsi boot mode changes (bee-nvidia-load + grub.cfg).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- build.sh: add --variant nvidia|amd; separate work dirs per variant
(live-build-work-nvidia / live-build-work-amd); GPU-specific steps
(modules, NCCL, cuBLAS, nccl-tests) run only for nvidia; deb package
cache synced back to shared location after each lb build so second
variant reuses downloaded packages; ISO output named
easy-bee-{variant}-v{ver}-amd64.iso
- build-in-container.sh: add --variant nvidia|amd|all (default: all);
runs build.sh twice in one container for 'all'; --clean-build wipes
both variant work dirs
- package-lists: remove GPU packages from bee.list.chroot; add
bee-nvidia.list.chroot (DCGM) and bee-amd.list.chroot (ROCm)
- 9000-bee-setup hook: read /etc/bee-gpu-vendor; enable bee-nvidia.service
and DCGM only for nvidia; set up ROCm symlinks only for amd
- auto/config: --iso-volume uses BEE_GPU_VENDOR_UPPER env var
- grub.cfg: add nomodeset to EASY-BEE and EASY-BEE (load to RAM) entries
— fixes X/lightdm on BMC KVM (ASPEED AST chip requires nomodeset for
fbdev to work; NVIDIA H100 compute does not need KMS)
- bee.sh / smoketest.sh: add /usr/sbin to PATH so dmidecode, smartctl,
nvme are found
- 9100-memtest hook: add diagnostic listing of chroot/boot/memtest* files
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
ROCM_BANDWIDTH_TEST_VERSION, ROCM_VALIDATION_SUITE_VERSION, ROCBLAS,
ROCRAND, HIP_RUNTIME_AMD, HIPBLASLT, COMGR were defined in VERSIONS and
in bee.list.chroot but the sed substitution block only covered 3 of them.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Runs CPU (stressapptest) + GPU stress simultaneously across multiple
load/idle cycles with varying idle durations (120s/60s/30s) to detect
cooling systems that fail to recover under repeated load.
Presets: smoke (~5 min), acceptance (~25 min), overnight (~100 min).
Outputs metrics.csv + summary.txt with per-cycle throttle and fan
spindown analysis, packed as tar.gz.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add modernc.org/sqlite dependency; write every sample to
/appdata/bee/metrics.db (WAL mode, prune to 24h on startup)
- Pre-fill ring buffers from last 120 DB rows on startup so charts
survive service restarts
- Ticker changed 3s→1s; chart JS refresh will be set to 2s (lag ≤3s)
- Add GET /api/metrics/export.csv for full history download
- Chart rendering: SymbolNone (no dots), right padding=80px so peak
mark line label is not clipped, min/avg/max appended to chart title
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Finds the series with the highest value across all datasets and adds
a SeriesMarkTypeMax dashed mark line to it. Since all series share the
same Y axis this effectively shows a single "global peak" line for the
whole chart with a label on the right.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
rvs was not in PATH so the stress job exited immediately (UNSUPPORTED).
Now resolveRVSCommand searches /opt/rocm-*/bin/rvs before failing.
Also add a Copy button overlay on all .terminal elements and set
user-select:text so logs can be copied from the web UI.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Legend names were "GPU 0 %" — remove unit suffix since chart title already
conveys it. Fan parsing now handles the 5-field IPMI SDR format where the
value+unit ("4340 RPM") are combined in the last column rather than split
across separate fields.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
MI250X outputs 7 temperature columns before power/use%; positional parsing
read junction temp (~40°C) as GPU utilisation. Switch to header-based
colIdx() lookup so the correct fields are read regardless of column order
or rocm-smi version.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add udev rule: /dev/ipmi0 readable by 'ipmi' group (no sudo needed)
- Add 'ipmi' group creation and bee user membership in chroot hook
- Remove legend from all charts (data shown in GPU table below)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Without these modules /dev/ipmi0 doesn't exist and ipmitool can't
read fan RPM, PSU fans, or IPMI temperature sensors.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Tccd1-8 (AMD CCD die temps) now classified as 'cpu' group,
appear on CPU Temperature chart instead of ambient
- Fan RPM card hidden when no fans detected
- Remove CPU Load/Mem Load/Power from fan table (have dedicated charts)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add rocm-validation-suite, rocblas, rocrand, hip-runtime-amd,
hipblaslt, comgr to ISO (~700MB, needed for HIP compute)
- RunAMDStressPack: run RVS GST (SGEMM ~31 TFLOPS/GPU) + bandwidth test
- Add rvs symlink in chroot setup hook
- Pin all new package versions in VERSIONS
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add rocm-bandwidth-test package to ISO
- Add bee user to 'render' group (/dev/kfd, /dev/dri/renderD* access)
- Add rocm-bandwidth-test symlink alongside rocm-smi
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
ip route show includes state flags like 'linkdown' that ip route add
does not accept, causing restore to fail.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace per-GPU cards with 4 combined charts showing all GPUs as
separate series. Add gpu-all-load/memload/power/temp endpoints.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- bee-web.service: remove After=bee-audit so Go starts immediately
- Go serves loading page from / when audit JSON not yet present;
JS polls /api/ready (503 until file exists, 200 when ready)
then redirects to dashboard
- bee-openbox-session: wait for /healthz (Go binds fast <2s),
open http://localhost/ directly — no file:// cross-origin issues
- Remove loading.html static file
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace 15s blocking wait with instant Chromium launch showing a
dark loading page that polls /healthz every 500ms and auto-redirects
to the app when ready.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
memtest files live in chroot /boot (inside squashfs) but GRUB needs
them on the ISO filesystem. Binary hook copies them out at build time.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Switch to performance governor when task queue starts processing,
back to powersave when queue drains. Removes bee-cpuperf.service.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Task queue: all SAT/audit jobs enqueue and run one-at-a-time;
tasks persist past page navigation; new Tasks page with cancel/priority/log stream
- UI: consolidate nav (Validate, Burn, Tasks, Tools); Audit becomes modal;
Dashboard hardware summary badges + split metrics charts (load/temp/power);
Tools page consolidates network, services, install, support bundle
- AMD GPU: acceptance test and stress burn cards; GPU presence API greys
out irrelevant SAT cards automatically
- Burn tests: Memory Stress (stress-ng --vm), SAT Stress (stressapptest)
- Install to RAM: copies squashfs to /dev/shm, re-associates loop devices
via LOOP_CHANGE_FD ioctl so live media can be ejected
- Charts: relative time axis (0 = now, negative left)
- memtester: LimitMEMLOCK=infinity in bee-web.service; empty output → UNSUPPORTED
- SAT overlay applied dynamically on every /audit.json serve
- MIME panic guard for LiveCD ramdisk I/O errors
- ISO: add memtest86+, stressapptest packages; memtest86+ GRUB entry;
disable screensaver/DPMS in bee-openbox-session
- Unknown SAT status severity = 1 (does not override OK)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Move datacenter-gpu-manager and rocm-smi-lib from dynamic chroot hooks
into live-build's config/archives mechanism so lb caches the .deb files
in cache/packages.chroot/ between builds, eliminating repeated 900+ MB
downloads. Versions pinned via VERSIONS and substituted into package
lists at build time.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
mchusandClaude Sonnet 4.6