tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-26 03:03:25 -04:00
Code Issues Packages Projects Releases Wiki Activity

metal : support bcast add & dup & cont op (#2323)

Browse Source
This commit is contained in:
Jiahao Li
2023-07-23 19:00:37 +08:00
committed by GitHub
parent d2a43664f9
commit 83a00ce69b
2 changed files with 22 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
ggml-metal.metal
View File

@ -67,6 +67,17 @@ kernel void kernel_add(
dst[tpig] = src0[tpig] + src1[tpig];
}
// assumption: src1 is a row
// broadcast src1 into src0
kernel void kernel_add_row(
device const float * src0,
device const float * src1,
device float * dst,
constant int64_t & ne00,
uint tpig[[thread_position_in_grid]]) {
dst[tpig] = src0[tpig] + src1[tpig % ne00];
}
kernel void kernel_mul(
device const float * src0,
device const float * src1,