tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-18 22:20:16 -04:00
Code Issues Packages Projects Releases Wiki Activity

metal : pad n_ctx by 32 (#6177)

Browse Source 
* metal : require ne00 >= 128 for mat-mat kernels

ggml-ci

* llama : pad n_ctx by 32

ggml-ci
This commit is contained in:
Georgi Gerganov
2024-03-22 09:36:03 +02:00
committed by GitHub
parent 59c17f02de
commit 95d576b48e
4 changed files with 14 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
tests/test-backend-ops.cpp
View File

@@ -2091,6 +2091,13 @@ static bool test_backend(ggml_backend_t backend, test_mode mode, const char * op
}
}
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 64, 2, 128, { 8, 1}, {1, 1}));
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 83, 2, 128, { 8, 1}, {4, 1}));
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 64, 2, 64, { 8, 1}, {4, 1}));
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 83, 2, 64, { 8, 1}, {4, 1}));
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 64, 45, 128, { 8, 1}, {4, 1}));
test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 128, 45, 64, { 8, 1}, {4, 1}));
for (ggml_type type_a : all_types) {
for (ggml_type type_b : {GGML_TYPE_F32 /*, GGML_TYPE_F16 */}) {
for (int n_mats : {2, 4, 8}) {