Skip to content

RMSNorm alignment calculation hotfix #436

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

Closed
Micky774 wants to merge 2 commits into from
Closed

RMSNorm alignment calculation hotfix #436

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8461f7e
Added hotfix
Micky774 Jan 28, 2026
d9aa309
Updated to account for element size
Micky774 Jan 28, 2026
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
14 changes: 9 additions & 5 deletions transformer_engine/pytorch/triton_kernels/rmsnorm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -365,8 +365,8 @@ def te_rmsnorm_bwd_triton(dz, x, rsigma, gamma, sm_margin, zero_centered_gamma):
dg_tmp = torch.empty(dg_tmp_rows(x_, sm_margin), N, device=x.device, dtype=torch.float32, requires_grad=False) if need_reduction else None

grid_bwd = lambda meta: (NUM_PRGMS, )
input_aligned_16 = (x_.data_ptr() % 16 == 0) and (x_.stride(-1) % 16 == 0)
grad_output_aligned_16 = (dz_.data_ptr() % 16 == 0) and (dz_.stride(-1) % 16 == 0)
input_aligned_16 = (x_.data_ptr() % 16 == 0) and (((x_.stride(0) * x_.dtype.itemsize) % 16) == 0)
grad_output_aligned_16 = (dz_.data_ptr() % 16 == 0) and (((dz_.stride(0) * dz_.dtype.itemsize) % 16) == 0)
_rmsnorm_bwd_triton[grid_bwd](dz_, x_, gamma_, rsigma_, dx, dg_tmp if need_reduction else dgamma,
x_.stride(0), dz_.stride(0), M, N, zero_centered_gamma, blk_size,
USE_BLOCKED, NUM_PRGMS, input_aligned_16, grad_output_aligned_16, num_warps=8)
Expand Down Expand Up @@ -453,10 +453,14 @@ def te_rmsnorm_fwd_triton(
grid_fwd = lambda meta: (NUM_PRGMS, )
# TODO(micky774) Implement fused MXFP8 quantization within the kernel
kernel = _rmsnorm_fwd_triton if autotune else _rmsnorm_fwd_triton_impl
input_aligned_16 = (input.data_ptr() % 16 == 0) and (input.stride(-1) % 16 == 0)
input_aligned_16 = (
(input.data_ptr() % 16 == 0) and
(((input.stride(0) * input.dtype.itemsize) % 16) == 0)
)
out_alignment_tensor = out._data if hasattr(out, "_data") else out
output_aligned_16 = (out_alignment_tensor.data_ptr() % 16 == 0) and (
out_alignment_tensor.stride(-1) % 16 == 0
output_aligned_16 = (
(out_alignment_tensor.data_ptr() % 16 == 0) and
(((out_alignment_tensor.stride(0) * out_alignment_tensor.dtype.itemsize) % 16) == 0)
)
kernel[grid_fwd](
out_ptr,
Expand Down