biondizzle/nvfp4-megamoe-kernel
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

auto: pre-test commit

Browse Source
This commit is contained in:
biondizzle
2026-05-28 15:55:59 +00:00
parent 1ec3e1ed2c
commit 43516ed4ec
1 changed files with 4 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/unit/test_fmha_hd64_n16_v2.cu
View File

@@ -41,7 +41,7 @@ test_fmha_hd64_n16(const bf16_t* q, const bf16_t* k, const bf16_t* v,
bf16_t* sK0 = sQ0 + NKT_QK * TILE_SZ;
bf16_t* sPk = (bf16_t*)(((uintptr_t)(sK0 + NKT_QK * TILE_SZ) + 127) & ~(uintptr_t)127);
bf16_t* sV = (bf16_t*)(((uintptr_t)(sPk + TILE_SZ) + 127) & ~(uintptr_t)127);
float* s_p_vals = (float*)(sV + NKT_PV * V_SUB_TILE_SZ); // Only 1 V sub-tile at a time
float* s_p_vals = (float*)(sV + V_SUB_TILE_SZ); // Only 1 V sub-tile at a time
// Load Q K-tiles (4 × (128,16) canonical)
for (int kt = 0; kt < NKT_QK; kt++) {
@@ -234,6 +234,9 @@ int main() {
cudaFuncSetAttribute(test_fmha_hd64_n16, cudaFuncAttributeMaxDynamicSharedMemorySize, smem);
test_fmha_hd64_n16<<<1, 128, smem>>>(d_q, d_k, d_v, d_o, d_o_scalar, SCALE);
cudaError_t launch_err = cudaGetLastError();
if (launch_err != cudaSuccess) { printf("LAUNCH ERROR: %s\n", cudaGetErrorString(launch_err)); return 1; }
cudaError_t err = cudaDeviceSynchronize();
if (err != cudaSuccess) { printf("CUDA ERROR: %s\n", cudaGetErrorString(err)); return 1; }