Remove trueFP16 gemms based on MSHADOW_USE_PASCAL=1. (dmlc#283)

crazy-cat · Aug 18, 2017 · e510277 · e510277
1 parent 1824582
commit e510277
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 18 deletions.
diff --git a/make/mshadow.mk b/make/mshadow.mk
@@ -125,6 +125,7 @@ else
 	MSHADOW_CFLAGS+= -DMSHADOW_DIST_PS=0
 endif
 
-# Set MSHADOW_USE_PASCAL to one to enable nvidia pascal gpu features.
-# Like cublasHgemm
+# MSHADOW_USE_PASCAL=1 used to enable true-fp16 gemms.  Now, mshadow
+# only uses pseudo-fp16 gemms, so this flag will be removed after
+# dependent projects no longer reference it.
 MSHADOW_CFLAGS += -DMSHADOW_USE_PASCAL=0
diff --git a/mshadow/dot_engine-inl.h b/mshadow/dot_engine-inl.h
@@ -430,14 +430,9 @@ struct BLASEngine<gpu, half::half_t> {
                           const half::half_t *B, int ldb, half::half_t beta,
                           half::half_t *C, int ldc) {
 #if defined(CUDA_VERSION) && CUDA_VERSION >= 7050
-  if (
-#if MSHADOW_USE_PASCAL == 1
-      false ||
-#endif
-      stream->dev_id == -1 || (stream->prop.major <= 5 && stream->prop.minor <= 2)) {
-    // Not PASCAL
-    float alpha_f = float(alpha);  // NOLINT(*)
-    float beta_f = float(beta);  // NOLINT(*)
+  // Always use pseudo-fp16: fp32 compute with fp16 I/O.
+  float alpha_f = float(alpha);  // NOLINT(*)
+  float beta_f = float(beta);  // NOLINT(*)
   #if CUDA_VERSION >= 8000
     cublasStatus_t err = cublasSgemmEx(Stream<gpu>::GetBlasHandle(stream),
                                        GetT(transa), GetT(transb), m, n, k, &alpha_f,
@@ -451,14 +446,6 @@ struct BLASEngine<gpu, half::half_t> {
                                        ldb, &beta_f, C, CUBLAS_DATA_HALF, ldc);
     CHECK_EQ(err, CUBLAS_STATUS_SUCCESS) << "Cublas SgemmEx fail";
   #endif  // CUDA_VERSION >= 8000
-  } else {
-    // PASCAL
-    cublasStatus_t err = cublasHgemm(Stream<gpu>::GetBlasHandle(stream),
-                                     GetT(transa), GetT(transb), m, n, k, &alpha.cuhalf_,
-                                     &A->cuhalf_, lda, &B->cuhalf_, ldb,
-                                     &beta.cuhalf_, &C->cuhalf_, ldc);
-    CHECK_EQ(err, CUBLAS_STATUS_SUCCESS) << "Cublas Hgemm fail";
-  }
 #else
     LOG(FATAL) << "Require CUDA version >= 7.5!";
 #endif  // defined(CUDA_VERSION) && CUDA_VERSION >= 7050