Verify Pmc Metrics

Mock-Based PMC Validation

Deterministic validation of AMD SMI Performance Monitoring Counter (PMC) metrics using the mock system (AMD_SMI_MOCK_CONFIG). Tests sentinel filtering, expected value ranges, positive presence of configured metrics, and negative absence of unsupported metrics across all output formats.

Sentinel Value Reference

GPU firmware returns these values for unsupported/unavailable metrics. They must NEVER appear in output.

Perfetto stores counter values as doubles, so UINT64_MAX appears as 1.8446744073709552e+19. Both exact and float representations must be checked.

Known Sentinel Filtering Pipeline

Known leakage risks:

1. RocPD VCN/JPEG array values (no explicit sentinel filter)
2. RocPD PCIe values (relies on device.hpp zeroing only)
3. NIC metrics in all outputs (no sentinel filtering anywhere)

Mock Configurations

Two architectures test complementary metric subsets. Located in ~/worktree/rocm-systems/amd-smi-mocks/.

MI300X (mock_mi300x.yaml)

Data-center GPU with per-XCP metrics, HBM, and NIC.

Navi 48 (mock_navi48.yaml)

Consumer GPU with device-level metrics, no XCP, no NIC.

Output Formats

Perfetto Track Name Differences

Legacy also uses abbreviated forms: Pkts vs Packets.

Test Matrix

Common env for ALL runs:

• AMD_SMI_MOCK_CONFIG=<path-to-yaml>
• ROCPROFSYS_AMD_SMI_METRICS="all"

Workload: ./transpose (mock generates all metric values regardless of workload type).

Execution

Phase 1: Setup

1. Auto-discover build directory (build/debug, build/release, or build/)
2. cd into build dir, source share/rocprofiler-systems/setup-env.sh
3. Verify mock support: ldd bin/rocprof-sys-run | grep yaml (yaml-cpp must be linked)
4. Locate mock configs: ~/worktree/rocm-systems/amd-smi-mocks/mock_mi300x.yaml and mock_navi48.yaml
5. Verify transpose workload exists

Phase 2: Execute Test Matrix

Run 6 profiling sessions. Use /tmp/pmc-validate-<run-id>/ output paths.

source share/rocprofiler-systems/setup-env.sh

# MI300X — Perfetto Standard
AMD_SMI_MOCK_CONFIG=~/worktree/rocm-systems/amd-smi-mocks/mock_mi300x.yaml \
ROCPROFSYS_OUTPUT_PATH=/tmp/pmc-validate-mi300-perfetto \
ROCPROFSYS_TRACE=1 \
ROCPROFSYS_AMD_SMI_METRICS="all" \
ROCPROFSYS_SAMPLING_AINICS="all" \
bin/rocprof-sys-run -- ./transpose

# MI300X — Perfetto Legacy
AMD_SMI_MOCK_CONFIG=~/worktree/rocm-systems/amd-smi-mocks/mock_mi300x.yaml \
ROCPROFSYS_OUTPUT_PATH=/tmp/pmc-validate-mi300-legacy \
ROCPROFSYS_TRACE=1 \
ROCPROFSYS_TRACE_LEGACY=true \
ROCPROFSYS_AMD_SMI_METRICS="all" \
ROCPROFSYS_SAMPLING_AINICS="all" \
bin/rocprof-sys-run -- ./transpose

# MI300X — RocPD
AMD_SMI_MOCK_CONFIG=~/worktree/rocm-systems/amd-smi-mocks/mock_mi300x.yaml \
ROCPROFSYS_OUTPUT_PATH=/tmp/pmc-validate-mi300-rocpd \
ROCPROFSYS_USE_ROCPD=1 \
ROCPROFSYS_AMD_SMI_METRICS="all" \
ROCPROFSYS_SAMPLING_AINICS="all" \
bin/rocprof-sys-run -- ./transpose

# Navi 48 — Perfetto Standard
AMD_SMI_MOCK_CONFIG=~/worktree/rocm-systems/amd-smi-mocks/mock_navi48.yaml \
ROCPROFSYS_OUTPUT_PATH=/tmp/pmc-validate-navi-perfetto \
ROCPROFSYS_TRACE=1 \
ROCPROFSYS_AMD_SMI_METRICS="all" \
bin/rocprof-sys-run -- ./transpose

# Navi 48 — Perfetto Legacy
AMD_SMI_MOCK_CONFIG=~/worktree/rocm-systems/amd-smi-mocks/mock_navi48.yaml \
ROCPROFSYS_OUTPUT_PATH=/tmp/pmc-validate-navi-legacy \
ROCPROFSYS_TRACE=1 \
ROCPROFSYS_TRACE_LEGACY=true \
ROCPROFSYS_AMD_SMI_METRICS="all" \
bin/rocprof-sys-run -- ./transpose

# Navi 48 — RocPD
AMD_SMI_MOCK_CONFIG=~/worktree/rocm-systems/amd-smi-mocks/mock_navi48.yaml \
ROCPROFSYS_OUTPUT_PATH=/tmp/pmc-validate-navi-rocpd \
ROCPROFSYS_USE_ROCPD=1 \
ROCPROFSYS_AMD_SMI_METRICS="all" \
bin/rocprof-sys-run -- ./transpose

Phase 3: Sentinel Detection (CRITICAL)

This is the most important validation phase. A single leaked sentinel value is a FAIL.

Perfetto Traces (Standard and Legacy)

Load each trace with MCP tools, then run:

-- SENTINEL CHECK: Find ANY sentinel value in PMC counters
SELECT t.name, c.value, c.ts
FROM counter c JOIN counter_track t ON c.track_id = t.id
WHERE t.name LIKE '%(S)%'
  AND (c.value = 65535
    OR c.value = 4294967295
    OR c.value = 18446744073709551615
    OR c.value = 1.8446744073709552e+19
    OR c.value >= 1.844e+19)
LIMIT 20;
-- Expected: 0 rows. ANY result is a FAIL.

RocPD Databases

sqlite3 <db-file> "
SELECT p.name, e.value
FROM rocpd_pmc_event e JOIN rocpd_info_pmc p ON e.pmc_id = p.id
WHERE e.value IN (65535, 4294967295, 18446744073709551615)
   OR e.value > 1e18
LIMIT 20;
"
# Expected: 0 rows. ANY result is a FAIL.

Phase 4: Positive Validation (Configured Metrics Present)

Verify that all configured metrics appear in output with expected value ranges.

Perfetto SQL

-- ALL PMC TRACKS: Get every metric with stats
SELECT t.name,
       COUNT(*) as samples,
       MIN(c.value) as min_val,
       MAX(c.value) as max_val,
       ROUND(AVG(c.value), 2) as avg_val
FROM counter c JOIN counter_track t ON c.track_id = t.id
WHERE t.name LIKE '%(S)%'
GROUP BY t.name
ORDER BY t.name;

RocPD SQL

sqlite3 <db-file> "
SELECT p.name, COUNT(*) as samples,
       MIN(e.value) as min_val, MAX(e.value) as max_val,
       ROUND(AVG(e.value), 2) as avg_val
FROM rocpd_pmc_event e JOIN rocpd_info_pmc p ON e.pmc_id = p.id
GROUP BY p.name ORDER BY p.name;
"

MI300X Expected Metrics

Navi 48 Expected Metrics

Range Validation SQL

For each metric with a known range, verify no values fall outside:

-- Example: MI300 power should be [300, 500]
SELECT t.name, c.value
FROM counter c JOIN counter_track t ON c.track_id = t.id
WHERE t.name LIKE '%Current Power%' AND t.name LIKE '%(S)%'
  AND (c.value < 300 OR c.value > 500)
LIMIT 10;
-- Expected: 0 rows

Repeat for each metric with known range bounds from the mock config.

Phase 5: Negative Validation (Unsupported Metrics Absent)

Verify that metrics NOT configured in the mock do NOT appear in output.

MI300X Should NOT Have

Navi 48 Should NOT Have

Perfetto Negative Check SQL

-- MI300X: Verify absent metrics
SELECT t.name FROM counter_track t
WHERE t.name LIKE '%(S)%'
  AND (t.name LIKE '%Temperature Edge%'
    OR t.name LIKE '%MM Busy%'
    OR t.name LIKE '%VCN Activity%'
    OR t.name LIKE '%JPEG Activity%'
    OR t.name LIKE '%XGMI%'
    OR t.name LIKE '%Voltage%'
    OR t.name LIKE '%Fan%');
-- Expected: 0 rows

-- Navi 48: Verify absent metrics
SELECT t.name FROM counter_track t
WHERE t.name LIKE '%(S)%'
  AND (t.name LIKE '%HBM Temp%'
    OR t.name LIKE '%VCN Busy%'
    OR t.name LIKE '%JPEG Busy%'
    OR t.name LIKE '%JPEG Activity%'
    OR t.name LIKE '%XGMI%'
    OR t.name LIKE '%NIC%');
-- Expected: 0 rows

RocPD Negative Check

# MI300X
sqlite3 <db-file> "
SELECT DISTINCT p.name FROM rocpd_info_pmc p
WHERE p.name IN ('device_busy_mm', 'device_temp_edge')
   OR p.name LIKE 'device_vcn_activity%'
   OR p.name LIKE 'device_jpeg_activity%'
   OR p.name LIKE 'device_xgmi%'
   OR p.name LIKE 'device_voltage%'
   OR p.name LIKE 'device_fan%';
"
# Expected: 0 rows

# Navi 48
sqlite3 <db-file> "
SELECT DISTINCT p.name FROM rocpd_info_pmc p
WHERE p.name LIKE 'device_temp_hbm%'
   OR p.name LIKE 'device_jpeg_activity%'
   OR p.name LIKE 'device_xgmi%'
   OR p.name LIKE 'nic_%';
"
# Expected: 0 rows

Phase 6: Cross-Format Consistency

For each architecture, verify that the same set of metrics appears across all 3 formats (accounting for naming differences):

-- Compare metric counts between Standard and Legacy Perfetto
-- Both should have the same number of unique PMC track names
SELECT COUNT(DISTINCT t.name) FROM counter_track t WHERE t.name LIKE '%(S)%';

For RocPD, compare the set of distinct rocpd_info_pmc.name values against Perfetto track names.

Discrepancies between formats indicate a bug in one of the output processors.

Phase 7: Cleanup

rm -rf /tmp/pmc-validate-mi300-perfetto /tmp/pmc-validate-mi300-legacy /tmp/pmc-validate-mi300-rocpd
rm -rf /tmp/pmc-validate-navi-perfetto /tmp/pmc-validate-navi-legacy /tmp/pmc-validate-navi-rocpd

Report Format

# PMC Validation Report (Mock-Based)

**Date**: YYYY-MM-DD
**Build**: <build_dir>
**Branch**: <branch>

## Sentinel Check Results

| Config | Format | Sentinel Values Found | Status |
|--------|--------|--------------------|--------|
| MI300X | Perfetto | 0 | PASS |
| MI300X | Legacy | 0 | PASS |
| MI300X | RocPD | 0 | PASS |
| Navi 48 | Perfetto | 0 | PASS |
| Navi 48 | Legacy | 0 | PASS |
| Navi 48 | RocPD | 0 | PASS |

## Positive Validation (Configured Metrics)

| Config | Metric | Perfetto | Legacy | RocPD | Range OK |
|--------|--------|----------|--------|-------|----------|
| MI300X | Power | PASS (N, [300-500]) | PASS | PASS | YES |
| MI300X | Temperature | PASS (N, [55-75]) | PASS | PASS | YES |
| ... | ... | ... | ... | ... | ... |
| Navi 48 | Power | PASS (N, [20-30]) | PASS | PASS | YES |
| ... | ... | ... | ... | ... | ... |

## Negative Validation (Unsupported Metrics)

| Config | Metric Should Be Absent | Perfetto | Legacy | RocPD |
|--------|------------------------|----------|--------|-------|
| MI300X | temperature_edge | PASS (absent) | PASS | PASS |
| MI300X | MM Busy | PASS (absent) | PASS | PASS |
| MI300X | XGMI | PASS (absent) | PASS | PASS |
| Navi 48 | HBM Temperature | PASS (absent) | PASS | PASS |
| Navi 48 | JPEG Activity | PASS (absent) | PASS | PASS |
| Navi 48 | NIC | PASS (absent) | PASS | PASS |
| ... | ... | ... | ... | ... |

## Cross-Format Consistency

| Config | Perfetto Metrics | Legacy Metrics | RocPD Metrics | Match |
|--------|-----------------|----------------|---------------|-------|
| MI300X | N | N | N | YES |
| Navi 48 | N | N | N | YES |

## Issues Found
[List with severity: CRITICAL / HIGH / MEDIUM / LOW]

## Conclusion
[Overall assessment]

Save report to: planning/pmc-validation-report.md in the project worktree.

Common Issues & Known Limitations

Integration

This skill is also agent-level knowledge for the orchestrator's Tester agent.