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

fix: bugfix to pr 135 #136

Merged
merged 5 commits into from
Feb 25, 2024
Merged

fix: bugfix to pr 135 #136

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
2e5221d
upd
yzh119 Feb 24, 2024
a9586b6
bugfix
yzh119 Feb 25, 2024
c69d352
bugfix
yzh119 Feb 25, 2024
002f393
upd
yzh119 Feb 25, 2024
a869311
bugfix
yzh119 Feb 25, 2024
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
20 changes: 15 additions & 5 deletions include/flashinfer/prefill.cuh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1530,9 +1530,14 @@ cudaError_t SinglePrefillWithKVCacheWorkEstimation(
(num_blocks_per_sm * num_sm) /
(num_kv_heads *
ceil_div(qo_len * group_size, num_rows_per_cta));
uint32_t chunk_size =
max(ceil_div(kv_len, max_num_kv_chunks), 256);
uint32_t num_chunks = ceil_div(kv_len, chunk_size);
uint32_t num_chunks;
if (max_num_kv_chunks > 0) {
uint32_t chunk_size =
max(ceil_div(kv_len, max_num_kv_chunks), 256);
num_chunks = ceil_div(kv_len, chunk_size);
} else {
num_chunks = 0;
}

max_grid_size = num_blocks_per_sm * num_sm;
if (num_chunks > 1) {
Expand Down Expand Up @@ -1626,8 +1631,13 @@ cudaError_t SinglePrefillWithKVCacheDispatched(DTypeIn* q, DTypeIn* k, DTypeIn*
uint32_t max_num_kv_chunks =
(num_blocks_per_sm * num_sm) /
(num_kv_heads * ceil_div(qo_len * GROUP_SIZE, num_rows_per_cta));
uint32_t chunk_size = max(ceil_div(kv_len, max_num_kv_chunks), 256);
uint32_t num_chunks = ceil_div(kv_len, chunk_size);
uint32_t num_chunks;
if (max_num_kv_chunks > 0) {
uint32_t chunk_size = max(ceil_div(kv_len, max_num_kv_chunks), 256);
num_chunks = ceil_div(kv_len, chunk_size);
} else {
num_chunks = 0;
}

if (num_chunks <= 1 || tmp == nullptr) {
// Enough parallelism, do not split-kv
Expand Down