Skip to content

head, seq, batch grid order for triton flash attention bwd. #141

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

Merged
micmelesse merged 1 commit into from
May 16, 2025
Merged

head, seq, batch grid order for triton flash attention bwd. #141

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
16 changes: 8 additions & 8 deletions flash_attn/flash_attn_triton_amd/bwd_prefill_onekernel.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -449,7 +449,7 @@ def _bwd_dq_inner(
use_cuda_graph=True,
)
@triton.jit
def bwd_kernel_causal( # grid = (tl.cdiv(max_seqlen_q // BLOCK_M2), batch, nheads_q)
def bwd_kernel_causal( # grid = (nheads_k, tl.cdiv(max_seqlen_q // BLOCK_M2), batch)
Q, K, V, sm_scale, DO, DQ, DK, DV,
M, Delta,
stride_qb, stride_qh, stride_qm, stride_qd,
Expand Down Expand Up @@ -487,9 +487,9 @@ def bwd_kernel_causal( # grid = (tl.cdiv(max_seqlen_q // BLOCK_M2), batch, nhead
DEBUG_TRITON_DETAIL: tl.constexpr,
):
# program ids
pid = tl.program_id(0)
bid = tl.program_id(1)
hkid = tl.program_id(2)
hkid = tl.program_id(0)
pid = tl.program_id(1)
bid = tl.program_id(2)
if DEBUG_TRITON: print(f"\npid: {pid}, bid: {bid}, hkid: {hkid}") # noqa: E701
# figure out varlen start and end
q_start = 0
Expand Down Expand Up @@ -843,9 +843,9 @@ def bwd_kernel_noncausal(
DEBUG_TRITON_DETAIL: tl.constexpr,
):
# program ids
pid = tl.program_id(0)
bid = tl.program_id(1)
hkid = tl.program_id(2)
hkid = tl.program_id(0)
pid = tl.program_id(1)
bid = tl.program_id(2)
if DEBUG_TRITON: print(f"\npid: {pid}, bid: {bid}, hkid: {hkid}") # noqa: E701
# figure out varlen start and end
q_start = 0
Expand Down Expand Up @@ -1202,7 +1202,7 @@ def attention_prefill_backward_triton_split_oneKernel_impl(
dropout_mask.stride()

seqlen = max(max_seqlen_q_final, max_seqlen_k_final)
grid = lambda META: ((seqlen + META['BLOCK_N1'] - 1) // META['BLOCK_N1'], batch, nheads_k)
grid = lambda META: (nheads_k, (seqlen + META['BLOCK_N1'] - 1) // META['BLOCK_N1'], batch, )
if causal:
if DEBUG_TRITON: print(f"bwd_kernel: grid = {grid}" ) # noqa: E701
bwd_kernel_causal[grid](
Expand Down
Loading