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

Adding cpu inference with VXE ISA for s390x architecture (#12613)

Browse Source 
Signed-off-by: Dilip Gowda Bhagavan <dilip.bhagavan@ibm.com>
Signed-off-by: Rishika Kedia <rishika.kedia@in.ibm.com>
Co-authored-by: Rishika Kedia <rishika.kedia@in.ibm.com>
This commit is contained in:
Dilip Gowda Bhagavan
2025-03-06 22:10:53 +05:30
committed by GitHub
parent bf0560bda9
commit ada19210a3
7 changed files with 653 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
csrc/cpu/attention.cpp
View File

@@ -24,8 +24,8 @@ struct KernelVecType<float> {
template <>
struct KernelVecType<c10::Half> {
#ifdef __powerpc64__
// Power architecture-specific vector types
#if defined(__powerpc64__) || defined(__s390x__)
// Power and s390x architecture-specific vector types
using q_load_vec_type = vec_op::FP32Vec8;
using k_load_vec_type = vec_op::FP32Vec16;
using v_load_vec_type = vec_op::FP32Vec16;