Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ggml : add ggml_top_k (#17365)

Browse Source 
* ggml : add ggml_top_k

* cont : add ggml_argsort_top_k

* metal : add top_k support

* ggml : cleanup

* tests : add virtual err() function for test_case

* ggml : add comments
This commit is contained in:
Georgi Gerganov
2025-11-25 15:31:43 +02:00
committed by GitHub
parent 05872ac885
commit 583cb83416
15 changed files with 516 additions and 85 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-metal/ggml-metal.metal
+22 -15
View File
@@ -4670,11 +4670,12 @@ kernel void kernel_argsort_f32_i32(
ushort3 ntg[[threads_per_threadgroup]]) {
// bitonic sort
const int col = tpitg[0];
const int ib = tgpig[0] / args.ne01;
const int i00 = (tgpig[0]/args.ne01)*ntg.x;
const int i01 = tgpig[0]%args.ne01;
const int i02 = tgpig[1];
const int i03 = tgpig[2];
const int i00 = ib*ntg.x;
const int i01 = tgpig[0] % args.ne01;
const int i02 = tgpig[1];
const int i03 = tgpig[2];
device const float * src0_row = (device const float *) (src0 + args.nb01*i01 + args.nb02*i02 + args.nb03*i03);
@@ -4710,9 +4711,11 @@ kernel void kernel_argsort_f32_i32(
}
}
const int64_t i0 = ib*args.top_k;
// copy the result to dst without the padding
if (i00 + col < args.ne00) {
dst += i00 + args.ne00*i01 + args.ne00*args.ne01*i02 + args.ne00*args.ne01*args.ne02*i03;
if (i0 + col < args.ne0 && col < args.top_k) {
dst += i0 + args.ne0*i01 + args.ne0*args.ne1*i02 + args.ne0*args.ne1*args.ne2*i03;
dst[col] = shmem_i32[col];
}
@@ -4747,22 +4750,22 @@ kernel void kernel_argsort_merge_f32_i32(
const int start = im * (2 * args.len);
const int len0 = MIN(args.len, MAX(0, args.ne00 - (int)(start)));
const int len1 = MIN(args.len, MAX(0, args.ne00 - (int)(start + args.len)));
const int len0 = MIN(args.len, MAX(0, args.ne0 - (int)(start)));
const int len1 = MIN(args.len, MAX(0, args.ne0 - (int)(start + args.len)));
const int total = len0 + len1;
device const int32_t * tmp0 = tmp + start
+ i01*args.ne00
+ i02*args.ne00*args.ne01
+ i03*args.ne00*args.ne01*args.ne02;
+ i01*args.ne0
+ i02*args.ne0*args.ne01
+ i03*args.ne0*args.ne01*args.ne02;
device const int32_t * tmp1 = tmp0 + args.len;
dst += start
+ i01*args.ne00
+ i02*args.ne00*args.ne01
+ i03*args.ne00*args.ne01*args.ne02;
+ i01*args.top_k
+ i02*args.top_k*args.ne01
+ i03*args.top_k*args.ne01*args.ne02;
device const float * src0_row = (device const float *)(src0
+ args.nb01*i01
@@ -4776,7 +4779,11 @@ kernel void kernel_argsort_merge_f32_i32(
const int chunk = (total + ntg.x - 1) / ntg.x;
const int k0 = tpitg.x * chunk;
const int k1 = min(k0 + chunk, total);
const int k1 = MIN(MIN(k0 + chunk, total), args.top_k);
if (k0 >= args.top_k) {
return;
}
if (k0 >= total) {
return;