8000 [inductor][cpu]pyhpc_equation_of_state performance regression · Issue #114489 · pytorch/pytorch · GitHub
[go: up one dir, main page]

Skip to content
[inductor][cpu]pyhpc_equation_of_state performance regression #114489
@zxd1997066

Description

@zxd1997066

pyhpc_equation_of_state single thread performance regression in 2023-11-15

shape name batch_size_new speed_up_new inductor_new eager_new batch_size_old speed_up_old inductor_old eager_old Ratio Speedup(New/old) Eager Ratio(old/new) Inductor Ratio(old/new)
static pyhpc_equation_of_state 1 14.52048 0.000068334 0.000992242480319999 1 19.168862 0.000059365 0.00113795949263 0.76 1.15 0.87
dynamic pyhpc_equation_of_state 1 12.973086 0.0000785899999999999 0.00101955482874 1 18.825686 0.000061129 0.001150795359494 0.69 1.13 0.78

SW info

SW Nightly commit Main commit
Pytorch a5a4048 6f44090
Torchbench / 7617d3f5
torchaudio 475b6ae ede4309
torchtext 142d029 45e4b8c
torchvision 8636bf3 4ac707a
torchdata eb9bf61 d76d92c
dynamo_benchmarks e6efc29 /

Repro:

inductor_single_run.sh

bash inductor_single_run.sh single inference performance torchbench pyhpc_equation_of_state float32 first static/dynamic default

Suspected guilty commit: 7fb09b8

cc @ezyang @msaroufim @wconstab @bdhirsh @anijain2305 @zou3519

Metadata

Metadata

Assignees

Labels

module: performanceIssues related to performance, either of kernel code or framework glueoncall: cpu inductorCPU Inductor issues for Intel team to triageoncall: pt2triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate module

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions

    0