-
Notifications
You must be signed in to change notification settings - Fork 75
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
Yugao/fix moreh microbenchmark #3214
Conversation
Please rebase to latest main. |
8427b57
to
3590105
Compare
pytest $TT_METAL_HOME/tests/scripts/test_moreh_microbenchmark.py::test_noc -k $ARCH_NAME | ||
pytest $TT_METAL_HOME/tests/scripts/test_moreh_microbenchmark.py::test_matmul_dram -k $ARCH_NAME | ||
pytest $TT_METAL_HOME/tests/scripts/test_moreh_microbenchmark.py::test_matmul_l1 -k $ARCH_NAME | ||
for ((i = 0; i < 1; i++)); do |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
1 for loop? Should we just get rid of this?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yeah I'll delete that
@@ -266,7 +266,10 @@ def test_pcie_h2d_l1(iteration, test_vector): | |||
header = ['Transfer Size', 'WriteToDeviceL1', 'WriteToBuffer', 'EnqueueWriteBuffer'] | |||
data = [] | |||
for test_point in test_vector: | |||
bw_wdd = test_write_device_l1(iteration, 1, test_point) | |||
if test_point < 1048576: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
is there a better way than using a hard-coded number?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yeah, the test vector should be smaller than L1 size, I think I should pass it in as pytest parameter?
49159df
to
f1bc5fd
Compare
@TT-billteng I told @yugaoTT he can merge in because he addressed my comment, but I didn't notice your comment. Normally, all conversations should be resolved before merging per our contributing standards. Yu showed me his passing u benchmarks pipeline, so I hope it's not a huge problem this time. sorry about that Bill |
fixed the device hang problem