ViperEkura/AstrAI
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fix: improve attention kernel numerical stability and test precision checks

Browse Source 
- use fmaf() for V-accumulation in scalar decode paths to reduce rounding
- delay scale multiplication to after dot-product in scalar prefill
- unify __expf/expf across MMA and scalar paths for consistent numerics
- harmonize divide-by-zero guards to 1e-20f
- add both absolute and relative error checks in standalone tests (atol=0.01, rtol=0.01)
This commit is contained in:
ViperEkura
2026-07-21 23:05:17 +08:00
parent f7a16efc9d
commit 60d7ee614a
6 changed files with 57 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files
csrc/tests/attn_decode_test.cu
+14 -5
View File
@@ -135,18 +135,27 @@ static int run_test(int B, int Hq, int Hk, int sl, int D, int causal) {
float* ref=new float[nQ];
cpu_attention_ref(hQ, hK, hV, hMask, ref, B, Hq, Hk, 1, sl, D, causal ? 0 : -1);
float max_err=0;
float max_abs_err=0, max_rel_err=0;
for (size_t i=0;i<nQ;i++){
float d=fabsf(bf2f(hOut[i])-ref[i]);
if(d>max_err) max_err=d;
float err=fabsf(bf2f(hOut[i])-ref[i]);
if(err>max_abs_err) max_abs_err=err;
float rel=err/fmaxf(fabsf(ref[i]), 1e-8f);
if(rel>max_rel_err) max_rel_err=rel;
}
printf("kernel: %.3f ms max_err: %.6e\n\n",kms,max_err);
const float atol=0.01f, rtol=0.01f;
bool pass=true;
for (size_t i=0;i<nQ;i++){
float err=fabsf(bf2f(hOut[i])-ref[i]);
if (err > atol + rtol * fabsf(ref[i])) { pass=false; break; }
}
printf("kernel: %.3f ms max_abs_err: %.6e max_rel_err: %.6e %s\n\n",
kms, max_abs_err, max_rel_err, pass?"PASS":"FAIL");
cudaFree(dQ);cudaFree(dK);cudaFree(dV);cudaFree(dO);cudaFree(dMask);
free_scratch(sc);
delete[]hQ;delete[]hK;delete[]hV;delete[]hMask;delete[]hOut;delete[]ref;delete[]tmp;
return (max_err < 0.05f) ? 0 : 1;
return pass ? 0 : 1;
}
int main() {
csrc/tests/attn_paged_decode_test.cu
+13 -6
View File
@@ -135,23 +135,30 @@ static int run_test(int B, int Hq, int Hkv, int kv_len, int page_size, int causa
for (int i = 0; i < B * Hq * HEAD_DIM; i++)
h_o_paged[i] = __bfloat162float(h_o_bf16[i]);
float max_err = 0.0f;
float max_abs_err = 0.0f, max_rel_err = 0.0f;
int bad_idx = -1;
for (int i = 0; i < B * Hq * HEAD_DIM; i++) {
float e = fabsf(h_o_paged[i] - h_o_ref[i]);
if (e > max_err) { max_err = e; bad_idx = i; }
if (e > max_abs_err) { max_abs_err = e; bad_idx = i; }
float rel = e / fmaxf(fabsf(h_o_ref[i]), 1e-8f);
if (rel > max_rel_err) max_rel_err = rel;
}
bool pass = max_err < 0.02f;
const float atol = 0.01f, rtol = 0.01f;
bool pass = true;
for (int i = 0; i < B * Hq * HEAD_DIM; i++) {
float e = fabsf(h_o_paged[i] - h_o_ref[i]);
if (e > atol + rtol * fabsf(h_o_ref[i])) { pass = false; break; }
}
if (pass) {
printf("PASS (max_abs_err=%.4e)\n", max_err);
printf("PASS (max_abs_err=%.4e max_rel_err=%.4e)\n", max_abs_err, max_rel_err);
} else {
int b = bad_idx / (Hq * HEAD_DIM);
int h = (bad_idx / HEAD_DIM) % Hq;
int d = bad_idx % HEAD_DIM;
printf("FAIL (max_abs_err=%.4e at [%d,%d,%d]: ref=%.4f got=%.4f)\n",
max_err, b, h, d, h_o_ref[bad_idx], h_o_paged[bad_idx]);
printf("FAIL (max_abs_err=%.4e max_rel_err=%.4e at [%d,%d,%d]: ref=%.4f got=%.4f)\n",
max_abs_err, max_rel_err, b, h, d, h_o_ref[bad_idx], h_o_paged[bad_idx]);
printf(" ref[0..7]:");
for (int i = 0; i < 8 && i < HEAD_DIM; i++)
printf(" %.4f", h_o_ref[i]);
csrc/tests/attn_prefill_test.cu
+14 -5
View File
@@ -119,17 +119,26 @@ static int run_test(int B, int Hq, int Hk, int ql, int kl, int D, int causal) {
float* ref=new float[nQ];
cpu_attention_ref(hQ, hK, hV, nullptr, ref, B, Hq, Hk, ql, kl, D, causal ? 0 : -1);
float max_err=0;
float max_abs_err=0, max_rel_err=0;
for (size_t i=0;i<nQ;i++) {
float d=fabsf(bf2f(hOut[i])-ref[i]);
if(d>max_err) max_err=d;
float err=fabsf(bf2f(hOut[i])-ref[i]);
if(err>max_abs_err) max_abs_err=err;
float rel=err/fmaxf(fabsf(ref[i]), 1e-8f);
if(rel>max_rel_err) max_rel_err=rel;
}
printf("kernel: %.3f ms max_err: %.6e\n\n",kms,max_err);
const float atol=0.01f, rtol=0.01f;
bool pass=true;
for (size_t i=0;i<nQ;i++) {
float err=fabsf(bf2f(hOut[i])-ref[i]);
if (err > atol + rtol * fabsf(ref[i])) { pass=false; break; }
}
printf("kernel: %.3f ms max_abs_err: %.6e max_rel_err: %.6e %s\n\n",
kms, max_abs_err, max_rel_err, pass?"PASS":"FAIL");
cudaFree(dQ);cudaFree(dK);cudaFree(dV);cudaFree(dO);
delete[]hQ;delete[]hK;delete[]hV;delete[]hOut;delete[]ref;delete[]tmp;
return (max_err < 0.05f) ? 0 : 1;
return pass ? 0 : 1;
}
int main() {