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

Small changes from review

Browse source
This commit is contained in:
Phillip Kravtsov 2023-10-02 10:21:16 -07:00
parent 2b565916dd
commit e6bf87f785
2 changed files with 17 additions and 24 deletions
Download patch file Download diff file Expand all files Collapse all files

29
ggml-metal.m
View file

@ -750,19 +750,18 @@ void ggml_metal_graph_compute(
id<MTLBuffer> id_dst = dst ? ggml_metal_get_buffer(ctx, dst, &offs_dst) : nil;
//GGML_METAL_LOG_INFO("%s: op - %s\n", __func__, ggml_op_name(dst->op));
if (src0) {
GGML_METAL_LOG_INFO("%s: src0 - %4s [%5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(src0t), ne00, ne01, ne02,
ggml_is_contiguous(src0), src0->name);
}
if (src1) {
GGML_METAL_LOG_INFO("%s: src1 - %4s [%5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(src1t), ne10, ne11, ne12,
ggml_is_contiguous(src1), src1->name);
}
if (dst) {
GGML_METAL_LOG_INFO("%s: dst - %4s [%5lld, %5lld, %5lld], 1, %s\n", __func__, ggml_type_name(dstt), ne0, ne1, ne2,
dst->name);
}
*/
//if (src0) {
// GGML_METAL_LOG_INFO("%s: src0 - %4s [%5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(src0t), ne00, ne01, ne02,
// ggml_is_contiguous(src0), src0->name);
//}
//if (src1) {
// GGML_METAL_LOG_INFO("%s: src1 - %4s [%5lld, %5lld, %5lld], %d, %s\n", __func__, ggml_type_name(src1t), ne10, ne11, ne12,
// ggml_is_contiguous(src1), src1->name);
//}
//if (dst) {
// GGML_METAL_LOG_INFO("%s: dst - %4s [%5lld, %5lld, %5lld], 1, %s\n", __func__, ggml_type_name(dstt), ne0, ne1, ne2,
// dst->name);
//}
switch (dst->op) {
case GGML_OP_NONE:
@ -775,8 +774,6 @@ void ggml_metal_graph_compute(
} break;
case GGML_OP_CONCAT:
{
GGML_ASSERT(ggml_is_contiguous(src0));
GGML_ASSERT(ggml_is_contiguous(src1));
int64_t nb = ne00;
[encoder setComputePipelineState:ctx->pipeline_concat];
@ -951,6 +948,8 @@ void ggml_metal_graph_compute(
} break;
case GGML_OP_SQR:
{
GGML_ASSERT(ggml_is_contiguous(src0));
[encoder setComputePipelineState:ctx->pipeline_sqr];
[encoder setBuffer:id_src0 offset:offs_src0 atIndex:0];
[encoder setBuffer:id_dst offset:offs_dst atIndex:1];

12
llama.cpp
View file

@ -3037,11 +3037,9 @@ static struct ggml_cgraph * llm_build_baichaun(
struct ggml_init_params params = {
/*.mem_size =*/ buf_compute.size,
/*.mem_buffer =*/ buf_compute.data,
/*.no_alloc =*/ false,
/*.no_alloc =*/ true,
};
params.no_alloc = true;
struct ggml_context * ctx0 = ggml_init(params);
ggml_cgraph * gf = ggml_new_graph(ctx0);
@ -3445,11 +3443,9 @@ static struct ggml_cgraph * llm_build_falcon(
struct ggml_init_params params = {
/*.mem_size =*/ buf_compute.size,
/*.mem_buffer =*/ buf_compute.data,
/*.no_alloc =*/ false,
/*.no_alloc =*/ true,
};
params.no_alloc = true;
struct ggml_context * ctx0 = ggml_init(params);
ggml_cgraph * gf = ggml_new_graph(ctx0);
@ -3805,11 +3801,9 @@ static struct ggml_cgraph * llm_build_starcoder(
struct ggml_init_params params = {
/*.mem_size =*/ buf_compute.size,
/*.mem_buffer =*/ buf_compute.data,
/*.no_alloc =*/ false,
/*.no_alloc =*/ true,
};
params.no_alloc = true;
struct ggml_context * ctx0 = ggml_init(params);
ggml_cgraph * gf = ggml_new_graph(ctx0);