biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[CPU] Add head sizes 80 and 112 with vec16 fallback (#31968)

Browse Source 
Signed-off-by: Rehan Khan <Rehan.Khan7@ibm.com>
This commit is contained in:
R3hankhan
2026-01-09 19:44:46 +05:30
committed by GitHub
parent 7cdf7e2fe0
commit 8e27663b6a
4 changed files with 12 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
csrc/cpu/cpu_attn_amx.hpp
View File

@@ -377,7 +377,7 @@ class AttentionImpl<ISA::AMX, scalar_t, head_dim> {
const int32_t q_heads_per_kv, const int64_t q_num_stride,
const int64_t q_head_stride, const float scale) {
constexpr int64_t bytes_per_head = head_dim * sizeof(scalar_t);
static_assert(bytes_per_head % AMX_TILE_ROW_BYTES == 0);
// static_assert(bytes_per_head % AMX_TILE_ROW_BYTES == 0);
constexpr int64_t head_size_block_num = bytes_per_head / AMX_TILE_ROW_BYTES;
constexpr int64_t head_elem_num_pre_block =
AMX_TILE_ROW_BYTES / sizeof(scalar_t);