Skip to content

Add benchmark methodology to profiling skill - #276

Merged
justinchuby merged 2 commits into
mainfrom
skill-benchmark-dimensions
May 6, 2026
Merged

Add benchmark methodology to profiling skill#276
justinchuby merged 2 commits into
mainfrom
skill-benchmark-dimensions

Conversation

@justinchuby

Copy link
Copy Markdown
Member

Add a comprehensive benchmark methodology section to the profiling skill.

Benchmark dimensions covered

  • 5 pipelines: ORT CUDA, GenAI CUDA/CPU, HF PyTorch CUDA/CPU
  • 5 dtypes: F16, BF16, F32, Q4_K_M, NF4
  • 3 EP variants: default (MEA), cuda (GQA rewrite), onnx-standard
  • 4 input modalities: text, image+text, audio+text, multimodal
  • 4 core metrics: decode tok/s, TTFT, VRAM, model size
  • 5 additional dimensions: cache type, prompt length, batch size, padding strategy, KV buffer
  • 6 best practices for reproducible benchmarks

Add comprehensive benchmark dimensions covering:
- 5 pipeline options (ORT CUDA, GenAI CUDA/CPU, HF PyTorch)
- 5 dtype options (F16, BF16, F32, Q4_K_M, NF4)
- 3 EP variants with attention kernel mapping
- 4 input modalities (text, image+text, audio+text, all)
- 4 core metrics (decode tok/s, TTFT, VRAM, model size)
- 5 additional dimensions (cache type, prompt length, batch size,
  padding strategy, KV buffer sharing)
- 6 best practices for reproducible benchmarks

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Signed-off-by: Justin Chu <justinchu@microsoft.com>
@github-actions

github-actions Bot commented May 6, 2026

Copy link
Copy Markdown

Performance Comparison

Comparing e5eb9666030f30

Model Metric Baseline Current Delta
bert (feature-extraction) model_size_bytes 359 KB 359 KB +0.0%
bert (feature-extraction) num_nodes 60 60 +0.0%
falcon model_size_bytes 364 KB 364 KB +0.0%
falcon num_nodes 66 66 +0.0%
gemma2 model_size_bytes 428 KB 428 KB +0.0%
gemma2 num_nodes 107 107 +0.0%
gpt2 model_size_bytes 388 KB 388 KB +0.0%
gpt2 num_nodes 53 53 +0.0%
llama model_size_bytes 425 KB 425 KB +0.0%
llama num_nodes 61 61 +0.0%
llama (static-cache) model_size_bytes 425 KB 425 KB +0.0%
llama (static-cache) num_nodes 58 58 +0.0%
mamba (ssm-text-generation) model_size_bytes 296 KB 296 KB +0.0%
mamba (ssm-text-generation) num_nodes 98 98 +0.0%
phi3 model_size_bytes 421 KB 421 KB +0.0%
phi3 num_nodes 59 59 +0.0%
phi3 (static-cache) model_size_bytes 421 KB 421 KB +0.0%
phi3 (static-cache) num_nodes 56 56 +0.0%
qwen2 model_size_bytes 425 KB 425 KB +0.0%
qwen2 num_nodes 61 61 +0.0%
qwen2 (static-cache) model_size_bytes 425 KB 425 KB +0.0%
qwen2 (static-cache) num_nodes 58 58 +0.0%
qwen3_5_moe (hybrid-text-generation) model_size_bytes 506 KB 506 KB +0.0%
qwen3_5_moe (hybrid-text-generation) num_nodes 275 275 +0.0%
qwen3_5_text (hybrid-text-generation) model_size_bytes 458 KB 458 KB +0.0%
qwen3_5_text (hybrid-text-generation) num_nodes 129 129 +0.0%
qwen3_5_vl (hybrid-qwen-vl) model_size_bytes 977 KB 977 KB +0.0%
qwen3_5_vl (hybrid-qwen-vl) num_nodes 408 408 +0.0%
t5 (seq2seq) model_size_bytes 836 KB 836 KB +0.0%
t5 (seq2seq) num_nodes 166 166 +0.0%
whisper (speech-to-text) model_size_bytes 1008 KB 1008 KB +0.0%
whisper (speech-to-text) num_nodes 128 128 +0.0%

No performance regressions.

@codecov

codecov Bot commented May 6, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

@justinchuby

Copy link
Copy Markdown
Member Author

Lowering opset is not the right thing to do. We should update onnxruntime kernel registration

Address reviewer feedback: lowering opset is not the right approach.
The correct fix for opset version mismatches is to update ORT kernel
registration for the missing opset versions.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Signed-off-by: Justin Chu <justinchu@microsoft.com>
@justinchuby
justinchuby merged commit 8542e9f into main May 6, 2026
19 checks passed
@justinchuby
justinchuby deleted the skill-benchmark-dimensions branch May 6, 2026 21:22
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