[ALMIOPEN-1856] Simplify dnn-benchmarking engine comparison CLI

[SamuelReeder]

Summary

This PR replaces the old A/B-specific dnn-benchmarking workflow with an explicit multi-engine comparison path. It keeps engine selection on --engine, allows comma-delimited plugin paths, and makes comparison output opt-in through --compare-engines.

Risk Assessment

Medium risk. This changes CLI behavior and removes legacy A/B flags in a benchmark tool, but the new path is covered by parser, config, execution, reporting, non-GPU tests, and gfx90a hardware smoke validation.

Testing Summary

• Static Python compile validation covered the changed source and tests.
• Focused unit tests covered CLI parsing, config validation, engine/plugin ordering, comparison deltas, output shape, and serialization.
• Full non-GPU test suite covered existing CPU-safe dnn-benchmarking behavior.
• CLI help was manually checked for new flags and absence of old A/B flags.
• Alola gfx90a validation built hipDNN/MIOpen provider through setup.sh --force-build -y and ran dnn-benchmarking smoke tests for no-comparison, comparison, and comma-delimited plugin-path flows.

Testing Checklist

• Python compile - python3 -m compileall -q projects/hipdnn/tools/dnn-benchmarking/src/dnn_benchmarking projects/hipdnn/tools/dnn-benchmarking/tests - Status: Passed
• Focused dnn-benchmarking tests - .venv/bin/python -m pytest tests/unit/cli/test_internal_profiling.py tests/unit/cli/test_suite_cli.py tests/unit/config/test_benchmark_config.py tests/unit/reporting/test_reporter.py tests/unit/reporting/test_suite_results.py tests/unit/execution/test_suite_runner.py - Status: Passed
• Non-GPU dnn-benchmarking suite - .venv/bin/python -m pytest -m 'not gpu' - Status: Passed
• CLI help check - .venv/bin/python -m dnn_benchmarking --help - Status: Passed
• Alola setup - bash setup.sh --force-build -y - ASICs: gfx90a - Status: Passed
• Alola single-engine smoke - python -m dnn_benchmarking --graph graphs/sample_conv_fwd.json --warmup 1 --iters 2 --engine 1563989756945604898 --plugin-path /opt/rocm/lib/hipdnn_plugins/engines - ASICs: gfx90a - Status: Passed
• Alola comparison smoke - python -m dnn_benchmarking --graph graphs/sample_conv_fwd.json --warmup 1 --iters 2 --engine 1563989756945604898,-6748551569128940061 --plugin-path /opt/rocm/lib/hipdnn_plugins/engines --compare-engines - ASICs: gfx90a - Status: Passed
• Alola comma-delimited plugin-path smoke - python -m dnn_benchmarking --graph graphs/sample_conv_fwd.json --warmup 1 --iters 1 --engine 1563989756945604898,-6748551569128940061 --plugin-path /opt/rocm/lib/hipdnn_plugins/engines,/opt/rocm/lib/hipdnn_plugins/engines --compare-engines - ASICs: gfx90a - Status: Passed
• PR CI - GitHub PR checks - Status: Pending

Technical Changes

• Removes legacy A/B runner/configuration code and old --AId, --BId, --APath, and --BPath flags.
• Adds --compare-engines and comma-delimited --plugin-path parsing with validation for one shared plugin path or one path per selected engine.
• Runs explicit --engine IDs in caller order so the first selected engine is the comparison baseline.
• Adds per-engine summary rows with raw kernel/E2E means and medians, plus opt-in percent delta columns for kernel mean, kernel median, E2E mean, and E2E median.
• Serializes median timing data and comparison-to-baseline metadata for JSON consumers.
• Keeps tarball graph extraction compatible with older Python tarfile APIs while preserving path traversal protections.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.

❌ Your project status has failed because the head coverage (77.83%) is below the target coverage (80.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #7866      +/-   ##
===========================================
+ Coverage    62.06%   62.08%   +0.02%     
===========================================
  Files         2085     2087       +2     
  Lines       357573   357886     +313     
  Branches     54060    54097      +37     
===========================================
+ Hits        221918   222176     +258     
- Misses      116866   116897      +31     
- Partials     18789    18813      +24     

Flag	Coverage Δ		*Carryforward flag
TensileLite	27.29% <ø> (ø)		Carriedforward from 5dbd0d3
hipBLAS	90.65% <ø> (ø)		Carriedforward from 5dbd0d3
hipBLASLt	41.27% <ø> (ø)		Carriedforward from 5dbd0d3
hipCUB	82.21% <ø> (ø)		Carriedforward from 5dbd0d3
hipDNN	86.58% <ø> (-0.03%)	⬇️
hipFFT	50.00% <ø> (ø)		Carriedforward from 5dbd0d3
hipRAND	76.12% <ø> (ø)		Carriedforward from 5dbd0d3
hipSOLVER	69.24% <ø> (ø)		Carriedforward from 5dbd0d3
hipSPARSE	85.42% <ø> (ø)		Carriedforward from 5dbd0d3
rocBLAS	48.09% <ø> (ø)		Carriedforward from 5dbd0d3
rocFFT	52.07% <ø> (ø)		Carriedforward from 5dbd0d3
rocRAND	57.04% <ø> (ø)		Carriedforward from 5dbd0d3
rocSOLVER	77.83% <ø> (ø)		Carriedforward from 5dbd0d3
rocSPARSE	72.68% <ø> (ø)		Carriedforward from 5dbd0d3

*This pull request uses carry forward flags. Click here to find out more.
see 8 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.