thek0tyara/llama-cpp-turboquant
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 6

metal : update sum_rows kernel to support float4 (#19524)

Browse source
This commit is contained in:
Georgi Gerganov 2026-02-12 11:35:28 +02:00 • committed by GitHub
parent 6845f7f87f
commit 3b3a948134
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 106 additions and 51 deletions
Download patch file Download diff file Expand all files Collapse all files

24
tests/test-backend-ops.cpp
View file

@ -8132,24 +8132,30 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_eval() {
}
test_cases.emplace_back(new test_sum());
test_cases.emplace_back(new test_sum_rows());
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, {11, 5, 6, 3}, {0, 2, 1, 3})); // row-contiguous but non-contiguous
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, {11, 5, 6, 3}, {0, 3, 2, 1}));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, {11, 5, 6, 3}, {0, 1, 3, 2}));
test_cases.emplace_back(new test_mean());
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 33, 1, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 33, 256, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 32769, 1, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 32, 1, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 32, 256, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 32768, 1, 1, 1 }));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 1, 1, 1 }));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 1024, 1, 1 }));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 256, 1, 1 }));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 256, 1, 1 }, { 1, 0, 2, 3 })); // sum dst not-contiguous
test_cases.emplace_back(new test_sum_rows());
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 11, 5, 6, 3 }, true, false));
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 11, 5, 6, 3 }, false, true));
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 11, 5, 6, 3 }, true, true));
test_cases.emplace_back(new test_mean());
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 1, 1, 1 }));
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 16, 5, 6, 3 }, true, false));
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 16, 5, 6, 3 }, false, true));
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 16, 5, 6, 3 }, true, true));
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 33, 1, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 33, 1, 1, 1 }));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 1024, 1, 1 }));
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 33, 1024, 1, 1 }));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 256, 1, 1 }));
test_cases.emplace_back(new test_sum(GGML_TYPE_F32, { 33, 256, 1, 1 }, { 1, 0, 2, 3 })); // sum dst not-contiguous
test_cases.emplace_back(new test_sum_rows(GGML_TYPE_F32, { 33, 256, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 33, 256, 1, 1 }));
test_cases.emplace_back(new test_mean(GGML_TYPE_F32, { 32769, 1, 1, 1 }));
test_cases.emplace_back(new test_group_norm(GGML_TYPE_F32, {64, 64, 320, 1}));
test_cases.emplace_back(new test_group_norm(GGML_TYPE_F32, {9, 9, 1280, 1}));
test_cases.emplace_back(new test_group_norm_mul_add(GGML_TYPE_F32, {64, 64, 320, 1}));