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.

Sign up for GitHub

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

Jump to bottom

Align resulting computation of GBs, TFLOPs in tutorials #4752

Merged
merged 1 commit into from
Sep 18, 2024
Merged

Align resulting computation of GBs, TFLOPs in tutorials #4752

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion python/tutorials/01-vector-add.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -123,7 +123,7 @@ def benchmark(size, provider):
ms, min_ms, max_ms = triton.testing.do_bench(lambda: x + y, quantiles=quantiles)
if provider == 'triton':
ms, min_ms, max_ms = triton.testing.do_bench(lambda: add(x, y), quantiles=quantiles)
gbps = lambda ms: 3 * x.numel() * x.element_size() / ms * 1e-6
gbps = lambda ms: 3 * x.numel() * x.element_size() * 1e-9 / (ms * 1e-3)
return gbps(ms), gbps(max_ms), gbps(min_ms)


Expand Down
2 changes: 1 addition & 1 deletion python/tutorials/02-fused-softmax.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -231,7 +231,7 @@ def benchmark(M, N, provider):
ms = triton.testing.do_bench(lambda: torch.softmax(x, axis=-1))
if provider == 'triton':
ms = triton.testing.do_bench(lambda: softmax(x))
gbps = lambda ms: 2 * x.nelement() * x.element_size() * 1e-9 / (ms * 1e-3)
gbps = lambda ms: 2 * x.numel() * x.element_size() * 1e-9 / (ms * 1e-3)
Copy link
Contributor Author

Choose a reason for hiding this comment

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

numel is used across the codebase so I decided to replace .nelement() to .numel() to be consistent.

return gbps(ms)


Expand Down
4 changes: 2 additions & 2 deletions python/tutorials/05-layer-norm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -353,12 +353,12 @@ def y_fwd():

# forward pass
if mode == 'forward':
gbps = lambda ms: 2 * x.numel() * x.element_size() / ms * 1e-6
gbps = lambda ms: 2 * x.numel() * x.element_size() * 1e-9 / (ms * 1e-3)
ms, min_ms, max_ms = triton.testing.do_bench(y_fwd, quantiles=quantiles, rep=500)
# backward pass
if mode == 'backward':
y = y_fwd()
gbps = lambda ms: 3 * x.numel() * x.element_size() / ms * 1e-6 # noqa: F811, E704
gbps = lambda ms: 3 * x.numel() * x.element_size() * 1e-9 / (ms * 1e-3) # noqa: F811, E704
ms, min_ms, max_ms = triton.testing.do_bench(lambda: y.backward(dy, retain_graph=True), quantiles=quantiles,
grad_to_none=[x], rep=500)
return gbps(ms), gbps(max_ms), gbps(min_ms)
Expand Down
2 changes: 1 addition & 1 deletion python/tutorials/06-fused-attention.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -633,7 +633,7 @@ def bench_flash_attention(BATCH, H, N_CTX, HEAD_DIM, causal, mode, provider, dev
total_flops *= 0.5
if mode == "bwd":
total_flops *= 2.5 # 2.0(bwd) + 0.5(recompute)
return total_flops / ms * 1e-9
return total_flops * 1e-12 / (ms * 1e-3)


if __name__ == "__main__":
Expand Down
Loading