From: Wallentri <redacted>
Date: Sat, 14 Mar 2026 07:43:13 +0000 (+0300)
Subject: Use fp32 in cuBLAS V100 to avoid overflows, env variables to override cuBLAS compute... 
X-Git-Tag: upstream/1.8.4~26
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=8ad5cb1e9d9c5d81c2d9ed64993625509b16a89f;p=pkg%2Fggml%2Fsources%2Fwhisper.cpp

Use fp32 in cuBLAS V100 to avoid overflows, env variables to override cuBLAS compute type (llama/19959)

* Update ggml-cuda.cu

* Update ggml-cuda.cu

* Update build.md

* Update build.md

* Update ggml/src/ggml-cuda/ggml-cuda.cu

Co-authored-by: Johannes Gäßler <redacted>

* Update ggml-cuda.cu

* Update build.md

* Update ggml/src/ggml-cuda/ggml-cuda.cu

Co-authored-by: Johannes Gäßler <redacted>

* Update build.md

* Update ggml-cuda.cu

* Update ggml-cuda.cu

---------

Co-authored-by: Johannes Gäßler <redacted>
---

diff --git a/ggml/src/ggml-cuda/ggml-cuda.cu b/ggml/src/ggml-cuda/ggml-cuda.cu
index 9d2aacf4..ce7a80ac 100644
--- a/ggml/src/ggml-cuda/ggml-cuda.cu
+++ b/ggml/src/ggml-cuda/ggml-cuda.cu
@@ -1242,6 +1242,34 @@ static cudaError_t ggml_cuda_cpy_tensor_2d(
     }
 }
 
+struct cublas_force_compute_type {
+    bool fp32 = false;
+    bool fp16 = false;
+};
+
+static const cublas_force_compute_type & ggml_cuda_cublas_get_force_compute_type() {
+    static const cublas_force_compute_type compute_type = [] {
+        cublas_force_compute_type result;
+
+        const bool ggml_cuda_force_cublas_compute_32f_env = getenv("GGML_CUDA_FORCE_CUBLAS_COMPUTE_32F") != nullptr;
+        const bool ggml_cuda_force_cublas_compute_16f_env = getenv("GGML_CUDA_FORCE_CUBLAS_COMPUTE_16F") != nullptr;
+
+        GGML_ASSERT(ggml_cuda_force_cublas_compute_16f_env == false || ggml_cuda_force_cublas_compute_32f_env == false);
+
+        if (ggml_cuda_force_cublas_compute_32f_env) {
+            GGML_LOG_INFO("Detected GGML_CUDA_FORCE_CUBLAS_COMPUTE_32F\n");
+            result.fp32 = true;
+        } else if (ggml_cuda_force_cublas_compute_16f_env) {
+            GGML_LOG_INFO("Detected GGML_CUDA_FORCE_CUBLAS_COMPUTE_16F\n");
+            result.fp16 = true;
+        }
+
+        return result;
+    }();
+
+    return compute_type;
+}
+
 static void ggml_cuda_op_mul_mat_cublas(
     ggml_backend_cuda_context & ctx,
     const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst, const char * src0_dd_i, const float * src1_ddf_i,
@@ -1324,7 +1352,13 @@ static void ggml_cuda_op_mul_mat_cublas(
 
         CUBLAS_CHECK(cublasSetStream(ctx.cublas_handle(id), stream));
 
-        if (GGML_CUDA_CC_IS_CDNA(cc) || GGML_CUDA_CC_IS_RDNA4(cc)) {
+        const auto & force_compute_type = ggml_cuda_cublas_get_force_compute_type();
+
+        if (!force_compute_type.fp16 && (GGML_CUDA_CC_IS_CDNA(cc)
+                                        || GGML_CUDA_CC_IS_RDNA4(cc)
+                                        || cc == GGML_CUDA_CC_VOLTA
+                                        || force_compute_type.fp32))
+        {
             const float alpha = 1.0f;
             const float beta = 0.0f;
             CUBLAS_CHECK(
@@ -1923,10 +1957,23 @@ static void ggml_cuda_mul_mat_batched_cublas_impl(ggml_backend_cuda_context & ct
     cudaDataType_t cu_data_type_b = traits::data_type;
     const void * alpha = traits::get_alpha();
     const void * beta = traits::get_beta();
-    const float alpha_f32 = 1.0f;
-    const float beta_f32 = 0.0f;
 
-    if (dst->op_params[0] == GGML_PREC_DEFAULT) {
+    const auto & force_compute_type = ggml_cuda_cublas_get_force_compute_type();
+
+    int id = ggml_cuda_get_device();
+    const int cc = ggml_cuda_info().devices[id].cc;
+    static constexpr bool is_src0_type_f16 = src0_type == GGML_TYPE_F16;
+
+    // bf16 and fp32 are already being computed in fp32 (ensure it using static_assert),
+    // so checking necessity of forced fp32 only for fp16 src0_type
+    static_assert(is_src0_type_f16 || traits::compute_type == CUBLAS_COMPUTE_32F);
+
+    const bool need_compute_32f = is_src0_type_f16 && !force_compute_type.fp16 && (GGML_CUDA_CC_IS_CDNA(cc)
+                                                                                  || GGML_CUDA_CC_IS_RDNA4(cc)
+                                                                                  || cc == GGML_CUDA_CC_VOLTA
+                                                                                  || force_compute_type.fp32);
+
+    if (dst->op_params[0] == GGML_PREC_DEFAULT && !need_compute_32f) {
         if constexpr (src0_type == GGML_TYPE_F32) {
             dst_t = (char *) dst_ddf;  // Direct F32 output
         } else {
@@ -1936,18 +1983,10 @@ static void ggml_cuda_mul_mat_batched_cublas_impl(ggml_backend_cuda_context & ct
         }
     } else {
         dst_t = (char *) dst_ddf;
-        cu_compute_type = CUBLAS_COMPUTE_32F;
-        cu_data_type = CUDA_R_32F;
-        alpha = &alpha_f32;
-        beta = &beta_f32;
-    }
-
-    int id = ggml_cuda_get_device();
-    const int cc = ggml_cuda_info().devices[id].cc;
-    if (GGML_CUDA_CC_IS_CDNA(cc) || GGML_CUDA_CC_IS_RDNA4(cc)) {
-        cu_compute_type = CUBLAS_COMPUTE_32F;
-        alpha = &alpha_f32;
-        beta = &beta_f32;
+        cu_compute_type = batched_mul_mat_traits<GGML_TYPE_F32>::compute_type;
+        cu_data_type = batched_mul_mat_traits<GGML_TYPE_F32>::data_type;
+        alpha = batched_mul_mat_traits<GGML_TYPE_F32>::get_alpha();
+        beta = batched_mul_mat_traits<GGML_TYPE_F32>::get_beta();
     }
 
     GGML_ASSERT(ne12 % ne02 == 0);