@@ -7819,6 +7819,7 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_perf() {
78197819 for (int bs : {1 , 4 , 8 , 32 , 64 , 128 , 256 , 512 }) {
78207820 for (ggml_type type_a : {GGML_TYPE_F32, GGML_TYPE_F16, GGML_TYPE_Q4_0, GGML_TYPE_Q8_0, GGML_TYPE_Q4_K, GGML_TYPE_Q6_K, GGML_TYPE_IQ2_XS}) {
78217821 for (ggml_type type_b : {GGML_TYPE_F32}) {
7822+ test_cases.emplace_back (new test_mul_mat_id (type_a, type_b, 128 , 8 , false , 768 , bs, 2048 ));
78227823 test_cases.emplace_back (new test_mul_mat_id_fusion (type_a, type_b, 128 , 8 , false , 768 , bs, 2048 , 1 ));
78237824 }
78247825 }
@@ -7827,6 +7828,7 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_perf() {
78277828 for (int bs : {1 , 4 , 8 , 32 , 64 , 128 , 256 , 512 }) {
78287829 for (ggml_type type_a : {GGML_TYPE_F32, GGML_TYPE_F16, GGML_TYPE_Q4_0, GGML_TYPE_Q8_0, GGML_TYPE_Q4_K, GGML_TYPE_Q6_K, GGML_TYPE_IQ2_XS}) {
78297830 for (ggml_type type_b : {GGML_TYPE_F32}) {
7831+ test_cases.emplace_back (new test_mul_mat_id (type_a, type_b, 32 , 4 , false , 1792 , bs, 2048 ));
78307832 test_cases.emplace_back (new test_mul_mat_id_fusion (type_a, type_b, 32 , 4 , false , 1792 , bs, 2048 , 1 ));
78317833 }
78327834 }
@@ -7837,6 +7839,7 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_perf() {
78377839 for (int bs : {1 , 4 , 8 , 512 }) {
78387840 for (ggml_type type_a : {GGML_TYPE_MXFP4}) {
78397841 for (ggml_type type_b : {GGML_TYPE_F32}) {
7842+ test_cases.emplace_back (new test_mul_mat_id (type_a, type_b, 32 , 4 , false , 2880 , bs, 2880 ));
78407843 test_cases.emplace_back (new test_mul_mat_id_fusion (type_a, type_b, 32 , 4 , false , 2880 , bs, 2880 , 1 ));
78417844 }
78427845 }
0 commit comments