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

[Operator] Add cat op #213

Merged
merged 1 commit into from
Sep 18, 2024
Merged

[Operator] Add cat op #213

merged 1 commit into from
Sep 18, 2024

Conversation

zfu82
Copy link
Collaborator

@zfu82 zfu82 commented Sep 14, 2024

Performance

Tested on NV-A100

benchmark/test_special_perf.py::test_perf_cat Operator cat Performance Test (torch.float16)
Size        Torch Latency (ms)   Gems Latency (ms)
--------------------------------------------------
1024                  0.018432            0.016384
6144                   0.05632            0.045056
11264                 0.095232            0.069632
16384                 0.134144            0.093184
21504                  0.17408             0.11776
26624                 0.212992            0.141312
31744                 0.253952            0.165888
36864                 0.293888            0.192512
41984                 0.333824            0.217088
47104                 0.374784             0.24064
52224                  0.41472            0.264192
57344                  0.45568            0.288768
62464                  0.49664             0.31232
67584                 0.538624            0.338944
72704                 0.579584             0.36352
77824                 0.620544            0.388096
Operator cat Performance Test (torch.float32)
Size        Torch Latency (ms)   Gems Latency (ms)
--------------------------------------------------
1024                  0.018432             0.02048
6144                   0.07168            0.072704
11264                 0.120832            0.120832
16384                 0.173056            0.167936
21504                 0.223232             0.21504
26624                 0.274432            0.262144
31744                 0.324608            0.310272
36864                 0.375808            0.357376
41984                 0.425984            0.406528
47104                 0.477184             0.45568
52224                  0.52736            0.504832
57344                  0.57856             0.55296
62464                  0.62976            0.601088
67584                 0.679936             0.65024
72704                 0.731136            0.700416
77824                 0.782336            0.746496
Operator cat Performance Test (torch.bfloat16)
Size        Torch Latency (ms)   Gems Latency (ms)
--------------------------------------------------
1024                  0.016384            0.016384
6144                   0.05632            0.045056
11264                 0.095232            0.069632
16384                 0.134144            0.093184
21504                  0.17408             0.11776
26624                 0.214016            0.141312
31744                 0.253952            0.165888
36864                 0.292864            0.192512
41984                 0.333824            0.217088
47104                 0.374784             0.24064
52224                  0.41472            0.264192
57344                  0.45568            0.288768
62464                  0.49664             0.31232
67584                 0.538624            0.338944
72704                 0.579584             0.36352
77824                 0.620544            0.388096
PASSED
benchmark/test_special_perf.py::test_perf_cat_int Operator cat Performance Test (torch.int16)
Size        Torch Latency (ms)   Gems Latency (ms)
--------------------------------------------------
1024                  0.017408            0.016384
6144                   0.05632            0.045056
11264                 0.094208            0.069632
16384                 0.135168            0.093184
21504                  0.17408             0.11776
26624                 0.212992            0.141312
31744                 0.253952            0.165888
36864                 0.292864            0.193536
41984                 0.333824            0.217088
47104                 0.374784             0.24064
52224                  0.41472            0.264192
57344                  0.45568            0.288768
62464                  0.49664             0.31232
67584                 0.538624            0.338944
72704                  0.57856             0.36352
77824                 0.620544            0.388096
Operator cat Performance Test (torch.int32)
Size        Torch Latency (ms)   Gems Latency (ms)
--------------------------------------------------
1024                  0.019456             0.02048
6144                   0.07168            0.073728
11264                 0.120832            0.120832
16384                 0.172032            0.167936
21504                 0.223232             0.21504
26624                 0.274432            0.262144
31744                 0.324608            0.310272
36864                 0.375808            0.357376
41984                 0.427008            0.406528
47104                 0.477184             0.45568
52224                 0.528384            0.504832
57344                 0.579584             0.55296
62464                  0.62976            0.601088
67584                 0.679936             0.65024
72704                 0.731136            0.700416
77824                 0.782336            0.746496
PASSED

Copy link
Collaborator

@iclementine iclementine left a comment

Choose a reason for hiding this comment

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

LGTM

@iclementine iclementine merged commit f721fd7 into master Sep 18, 2024
4 checks passed
@iclementine iclementine deleted the dev_cat2 branch September 18, 2024 01:55
@iclementine iclementine self-assigned this Sep 18, 2024
DuanYaQi pushed a commit that referenced this pull request Sep 20, 2024
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.

2 participants