vulkan: extend sparse FA support to coopmat2 (#30003)

This commit is contained in:
Jeff Bolz authored and GitHub committed 2026-10-08 11:36:20 +02:00
1 parent 03aa006acb
commit dac3087394
3 files changed
+77 -41

No files matched your search

+11
View File
@@ -11526,6 +11526,15 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_eval() {
// KV not a multiple of the compaction workgroup size.
test_cases.emplace_back(new test_flash_attn_ext(128, 128, 1, { 8, 1}, 5003, 4, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_F16, GGML_TYPE_F16, {0, 1, 2, 3}, true, false, 512));
// Sparse gather: native block sizes, padded slots, and head/batch strides.
for (ggml_type type : {GGML_TYPE_F32, GGML_TYPE_F16, GGML_TYPE_BF16, GGML_TYPE_Q4_0, GGML_TYPE_Q4_1, GGML_TYPE_Q5_0, GGML_TYPE_Q5_1, GGML_TYPE_Q8_0, GGML_TYPE_IQ4_NL}) {
test_cases.emplace_back(new test_flash_attn_ext(128, 96, 2, {8, 2}, 5003, 3, true, false, 0, 0, GGML_PREC_F32, type, type, {0, 1, 2, 3}, true, false, 257));
}
test_cases.emplace_back(new test_flash_attn_ext(128, 96, 2, {8, 2}, 5003, 3, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_Q8_0, GGML_TYPE_F16, {0, 2, 1, 3}, true, false, 257));
test_cases.emplace_back(new test_flash_attn_ext(128, 96, 2, {8, 2}, 5003, 3, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_F16, GGML_TYPE_Q4_0, {0, 2, 1, 3}, true, false, 257));
test_cases.emplace_back(new test_flash_attn_ext(128, 96, 2, {8, 2}, 5003, 3, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_Q8_0, GGML_TYPE_F32, {0, 1, 2, 3}, true, false, 257));
test_cases.emplace_back(new test_flash_attn_ext(128, 96, 2, {8, 2}, 5003, 3, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_F32, GGML_TYPE_Q8_0, {0, 1, 2, 3}, true, false, 257));
// more V-is-sub-view-of-K cases: other head shapes, and full views with equal head sizes
test_cases.emplace_back(new test_flash_attn_ext(320, 256, 1, {32, 1}, 512, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_F16, GGML_TYPE_F16, {0, 1, 2, 3}, true, true));
test_cases.emplace_back(new test_flash_attn_ext(192, 128, 4, {8, 1}, 512, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_F16, GGML_TYPE_F16, {0, 1, 2, 3}, true, true));
@@ -12017,6 +12026,8 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_perf() {
test_cases.emplace_back(new test_flash_attn_ext(576, 512, 1, {16, 1}, kv, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_F16, GGML_TYPE_F16, {0, 1, 2, 3}, true, true, 512));
test_cases.emplace_back(new test_flash_attn_ext(256, 256, 2, {12, 1}, kv, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_F16, GGML_TYPE_F16, {0, 1, 2, 3}, true, false, 2048));
test_cases.emplace_back(new test_flash_attn_ext(256, 256, 2, {12, 1}, kv, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0, {0, 1, 2, 3}, true, false, 2048));
test_cases.emplace_back(new test_flash_attn_ext(256, 256, 2, {12, 1}, kv, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_Q4_0, GGML_TYPE_Q4_0, {0, 1, 2, 3}, true, false, 2048));
test_cases.emplace_back(new test_flash_attn_ext(256, 256, 2, {12, 1}, kv, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_BF16, GGML_TYPE_BF16, {0, 1, 2, 3}, true, false, 2048));
test_cases.emplace_back(new test_flash_attn_ext(256, 256, 2, {12, 1}, kv, 1, true, false, 0, 0, GGML_PREC_F32, GGML_TYPE_Q8_0, GGML_TYPE_Q8_0, {0, 1, 2, 3}, true, false, 0));
}