vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

metal : add cpy f16 -> f32 kernel

Browse source
This commit is contained in:
Georgi Gerganov 2023-12-12 14:14:15 +02:00
parent a742d9f9b7
commit 08eb99179a
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
4 changed files with 84 additions and 15 deletions
Download patch file Download diff file Expand all files Collapse all files

45
ggml-metal.metal
View file

@ -1698,8 +1698,8 @@ template [[host_name("kernel_argsort_f32_i32_asc")]] kernel argsort_t kernel_ar
template [[host_name("kernel_argsort_f32_i32_desc")]] kernel argsort_t kernel_argsort_f32_i32<GGML_SORT_DESC>;
kernel void kernel_cpy_f16_f16(
device const half * src0,
device half * dst,
device const half * src0,
device half * dst,
constant int64_t & ne00,
constant int64_t & ne01,
constant int64_t & ne02,
@ -1738,6 +1738,47 @@ kernel void kernel_cpy_f16_f16(
}
}
kernel void kernel_cpy_f16_f32(
device const half * src0,
device float * dst,
constant int64_t & ne00,
constant int64_t & ne01,
constant int64_t & ne02,
constant int64_t & ne03,
constant uint64_t & nb00,
constant uint64_t & nb01,
constant uint64_t & nb02,
constant uint64_t & nb03,
constant int64_t & ne0,
constant int64_t & ne1,
constant int64_t & ne2,
constant int64_t & ne3,
constant uint64_t & nb0,
constant uint64_t & nb1,
constant uint64_t & nb2,
constant uint64_t & nb3,
uint3 tgpig[[threadgroup_position_in_grid]],
uint3 tpitg[[thread_position_in_threadgroup]],
uint3 ntg[[threads_per_threadgroup]]) {
const int64_t i03 = tgpig[2];
const int64_t i02 = tgpig[1];
const int64_t i01 = tgpig[0];
const int64_t n = i03*ne02*ne01*ne00 + i02*ne01*ne00 + i01*ne00;
const int64_t i3 = n / (ne2*ne1*ne0);
const int64_t i2 = (n - i3*ne2*ne1*ne0) / (ne1*ne0);
const int64_t i1 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0) / ne0;
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0);
device float * dst_data = (device float *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
for (int64_t i00 = tpitg.x; i00 < ne00; i00 += ntg.x) {
device const half * src = (device half *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
dst_data[i00] = src[0];
}
}
kernel void kernel_cpy_f32_f16(
device const float * src0,
device half * dst,