Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Benchmarks: Unify metric names of benchmarks #252

Merged
merged 16 commits into from Dec 9, 2021
Merged

Conversation

yukirora
Copy link
Contributor

@yukirora yukirora commented Dec 1, 2021

Description
Revise metric names of benchmarks.

@yukirora yukirora added the documentation Improvements or additions to documentation label Dec 1, 2021
@yukirora yukirora requested a review from abuccts December 1, 2021 12:04
@yukirora yukirora requested a review from cp5555 as a code owner December 1, 2021 12:04
@codecov
Copy link

codecov bot commented Dec 1, 2021

Codecov Report

Merging #252 (5a176fa) into main (c13ed2a) will increase coverage by 0.06%.
The diff coverage is 98.48%.

Impacted file tree graph

@@            Coverage Diff             @@
##             main     #252      +/-   ##
==========================================
+ Coverage   87.61%   87.67%   +0.06%     
==========================================
  Files          70       70              
  Lines        3947     3967      +20     
==========================================
+ Hits         3458     3478      +20     
  Misses        489      489              
Flag Coverage Δ
cpu-unit-test 72.19% <89.39%> (+0.14%) ⬆️
cuda-unit-test 87.62% <98.48%> (+0.06%) ⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files Coverage Δ
...arks/micro_benchmarks/ib_validation_performance.py 86.97% <95.00%> (+0.64%) ⬆️
.../docker_benchmarks/rocm_onnxruntime_performance.py 90.90% <100.00%> (ø)
...ro_benchmarks/computation_communication_overlap.py 83.01% <100.00%> (ø)
...ks/micro_benchmarks/cuda_gemm_flops_performance.py 71.69% <100.00%> (ø)
...ch/benchmarks/micro_benchmarks/disk_performance.py 98.96% <100.00%> (ø)
...ks/micro_benchmarks/gemm_flops_performance_base.py 96.77% <100.00%> (+0.10%) ⬆️
.../benchmarks/micro_benchmarks/gpcnet_performance.py 93.10% <100.00%> (+1.10%) ⬆️
...hmarks/micro_benchmarks/gpu_copy_bw_performance.py 97.56% <100.00%> (ø)
...hmarks/micro_benchmarks/ib_loopback_performance.py 88.88% <100.00%> (ø)
...chmarks/micro_benchmarks/kernel_launch_overhead.py 85.00% <100.00%> (ø)
... and 7 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c13ed2a...5a176fa. Read the comment docs.

Copy link
Member

@abuccts abuccts left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

two general questions:

  1. do we need to have a unified metrics format, e.g., benchmark/description_type_statistic? type can be time/bw/flops/count, etc., statistic can be min/max/avg, etc.
  2. is it necessary to include the unit in the metrics? like time_ms, etc.

docs/user-tutorial/benchmarks/micro-benchmarks.md Outdated Show resolved Hide resolved
docs/user-tutorial/benchmarks/micro-benchmarks.md Outdated Show resolved Hide resolved
docs/user-tutorial/benchmarks/micro-benchmarks.md Outdated Show resolved Hide resolved
docs/user-tutorial/benchmarks/micro-benchmarks.md Outdated Show resolved Hide resolved
docs/user-tutorial/benchmarks/micro-benchmarks.md Outdated Show resolved Hide resolved
@yukirora yukirora requested a review from a team as a code owner December 7, 2021 15:08
'throughput-train-fp32': [[step1_time, ..., stepK_time], ..., […]],
'throughput-train-fp16': [[step1_time, ..., stepK_time], ..., […]],
'throughput-inference-fp32': [[step1_time, ..., stepK_time], ..., […]],
'throughput-inference-fp16': [[step1_time, ..., stepK_time], ..., […]],
Copy link
Contributor

@cp5555 cp5555 Dec 8, 2021

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Change to avg_throughput

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

@yukirora yukirora changed the title Docs - Revise metric names of benchmarks Benchmarks: Unify metric names of benchmarks Dec 9, 2021
@abuccts abuccts enabled auto-merge (squash) December 9, 2021 04:29
@abuccts abuccts merged commit 9f56b21 into main Dec 9, 2021
@abuccts abuccts deleted the v-yujiang/doc-metric branch December 9, 2021 04:48
@cp5555 cp5555 mentioned this pull request Dec 21, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
documentation Improvements or additions to documentation
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants