vulkan: fix mul_mat_vec failure in backend tests (#12529)

The OOB calculation could be wrong if the last iteration was during one of
the unrolled loops. Adjust the unrolling counts to avoid this. Add a couple
new backend tests that hit this failure on NVIDIA GPUs.

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp
index 31ecd9f81a85ffbeec0ea0681e13d8eaeb2a38c9..775b48cd05e1af31c004d73c5381f9d99175f8ac 100644 (file)
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp
@@ -105,6 +105,16 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
     int unroll_count = 4;
     uint unrolled_iters = num_iters & ~(unroll_count - 1);
 
+#if K_PER_ITER == 2
+    // If the K dimension is odd, we need lastiter==true on the last iteration
+    // so OOB is computed correctly. Skip some unrolling to make that happen.
+    if ((p.ncols & 1) != 0 &&
+        unrolled_iters == num_iters &&
+        unrolled_iters > 0) {
+        unrolled_iters -= unroll_count;
+    }
+#endif
+
     uint i = 0;
     while (i < unrolled_iters) {
         // Manually partially unroll the loop
@@ -113,8 +123,18 @@ void compute_outputs(const uint32_t first_row, const uint32_t num_rows) {
             i++;
         }
     }
+
     unroll_count = 2;
     unrolled_iters = num_iters & ~(unroll_count - 1);
+
+#if K_PER_ITER == 2
+    if ((p.ncols & 1) != 0 &&
+        unrolled_iters == num_iters &&
+        unrolled_iters > 0) {
+        unrolled_iters -= unroll_count;
+    }
+#endif
+
     while (i < unrolled_iters) {
         // Manually partially unroll the loop
         [[unroll]] for (uint k = 0; k < unroll_count; ++k) {

diff --git a/tests/test-backend-ops.cpp b/tests/test-backend-ops.cpp
index ebc32d7919fe48ce936d77a96c8046610d7b40fd..28f860a7f2969bc0b01ea7e2ae0901acd57b771a 100644 (file)
--- a/tests/test-backend-ops.cpp
+++ b/tests/test-backend-ops.cpp
@@ -4204,6 +4204,8 @@ static std::vector<std::unique_ptr<test_case>> make_test_cases_eval() {
     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}));
+    test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 1056, 1, 193, {1,  1}, {4, 1}, {0, 2, 1, 3}));
+    test_cases.emplace_back(new test_mul_mat(GGML_TYPE_F16, GGML_TYPE_F32, 1056, 1, 67,  {1,  1}, {4, 1}, {0, 2, 1, 3}));
 
     for (auto bs : {1,2,4,8}) {
         for (auto nr : {1,4}) {