reanimator/bee
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

Split bee-bench into perf and power workflows

Browse Source
This commit is contained in:
Mikhail Chusavitin
2026-04-14 17:33:13 +03:00
parent 54338dbae5
commit 95124d228f
17 changed files with 718 additions and 259 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
audit/internal/platform/benchmark_report.go
View File

@@ -48,7 +48,7 @@ func renderBenchmarkReportWithCharts(result NvidiaBenchmarkResult) string {
fmt.Fprintf(&b, "**GPU(s):** %s \n", strings.Join(parts, ", "))
}
fmt.Fprintf(&b, "**Profile:** %s \n", result.BenchmarkProfile)
fmt.Fprintf(&b, "**App version:** %s \n", result.BenchmarkVersion)
fmt.Fprintf(&b, "**Benchmark version:** %s \n", result.BenchmarkVersion)
fmt.Fprintf(&b, "**Generated:** %s \n", result.GeneratedAt.Format("2006-01-02 15:04:05 UTC"))
if result.RampStep > 0 && result.RampTotal > 0 {
fmt.Fprintf(&b, "**Ramp-up step:** %d of %d \n", result.RampStep, result.RampTotal)
@@ -83,7 +83,7 @@ func renderBenchmarkReportWithCharts(result NvidiaBenchmarkResult) string {
// ── Methodology ───────────────────────────────────────────────────────────
b.WriteString("## Methodology\n\n")
fmt.Fprintf(&b, "- Profile `%s` uses standardized baseline -> warmup -> steady-state -> interconnect -> cooldown phases.\n", result.BenchmarkProfile)
fmt.Fprintf(&b, "- Profile `%s` uses standardized baseline -> warmup -> steady-state -> interconnect phases.\n", result.BenchmarkProfile)
b.WriteString("- Single-GPU compute score comes from `bee-gpu-burn` on the cuBLASLt path when available.\n")
b.WriteString("- Thermal and power limits are inferred from NVIDIA clock-event counters plus sustained telemetry.\n")
b.WriteString("- `result.json` is the canonical machine-readable source for the run.\n\n")
@@ -170,6 +170,16 @@ func renderBenchmarkReportWithCharts(result NvidiaBenchmarkResult) string {
if gpu.PowerLimitW > 0 {
fmt.Fprintf(&b, "- **Power limit:** %.0f W (default %.0f W)\n", gpu.PowerLimitW, gpu.DefaultPowerLimitW)
}
if gpu.PowerLimitDerated {
fmt.Fprintf(&b, "- **Power limit derating:** active after %d targeted_power attempt(s)\n", gpu.PowerCalibrationTries)
}
if gpu.CalibratedPeakPowerW > 0 {
if gpu.CalibratedPeakTempC > 0 {
fmt.Fprintf(&b, "- **Power calibration (`dcgmi targeted_power`):** %.0f W p95 at %.1f °C p95\n", gpu.CalibratedPeakPowerW, gpu.CalibratedPeakTempC)
} else {
fmt.Fprintf(&b, "- **Power calibration (`dcgmi targeted_power`):** %.0f W p95\n", gpu.CalibratedPeakPowerW)
}
}
if gpu.LockedGraphicsClockMHz > 0 {
fmt.Fprintf(&b, "- **Locked clocks:** GPU %.0f MHz / Mem %.0f MHz\n", gpu.LockedGraphicsClockMHz, gpu.LockedMemoryClockMHz)
}
@@ -188,7 +198,7 @@ func renderBenchmarkReportWithCharts(result NvidiaBenchmarkResult) string {
// Per-precision stability phases.
if len(gpu.PrecisionSteady) > 0 {
b.WriteString("**Per-precision stability:**\n\n")
b.WriteString("| Precision | Clock CV | Power CV | Clock Drift | ECC corr | ECC uncorr |\n|-----------|----------|----------|-------------|----------|------------|\n")
b.WriteString("| Precision | Status | Clock CV | Power CV | Clock Drift | ECC corr | ECC uncorr |\n|-----------|--------|----------|----------|-------------|----------|------------|\n")
for _, p := range gpu.PrecisionSteady {
eccCorr := "—"
eccUncorr := "—"
@@ -196,8 +206,12 @@ func renderBenchmarkReportWithCharts(result NvidiaBenchmarkResult) string {
eccCorr = fmt.Sprintf("%d", p.ECC.Corrected)
eccUncorr = fmt.Sprintf("%d", p.ECC.Uncorrected)
}
fmt.Fprintf(&b, "| %s | %.1f%% | %.1f%% | %.1f%% | %s | %s |\n",
p.Precision, p.Steady.ClockCVPct, p.Steady.PowerCVPct, p.Steady.ClockDriftPct,
status := p.Status
if strings.TrimSpace(status) == "" {
status = "OK"
}
fmt.Fprintf(&b, "| %s | %s | %.1f%% | %.1f%% | %.1f%% | %s | %s |\n",
p.Precision, status, p.Steady.ClockCVPct, p.Steady.PowerCVPct, p.Steady.ClockDriftPct,
eccCorr, eccUncorr)
}
b.WriteString("\n")
@@ -364,6 +378,7 @@ func formatThrottleLine(t BenchmarkThrottleCounters, steadyDurationSec float64)
func renderBenchmarkSummary(result NvidiaBenchmarkResult) string {
var b strings.Builder
fmt.Fprintf(&b, "run_at_utc=%s\n", result.GeneratedAt.Format(time.RFC3339))
fmt.Fprintf(&b, "benchmark_version=%s\n", result.BenchmarkVersion)
fmt.Fprintf(&b, "benchmark_profile=%s\n", result.BenchmarkProfile)
fmt.Fprintf(&b, "overall_status=%s\n", result.OverallStatus)
fmt.Fprintf(&b, "gpu_count=%d\n", len(result.GPUs))