From 97878e53fd30580bbc3e96479691de5b78527bb0 Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Sat, 6 Apr 2024 17:50:21 +0300
Subject: [PATCH] sync : llama.cpp (skip)

ggml-ci
---
 ggml-cuda/common.cuh |  6 ------
 ggml-cuda/dmmv.cu    |  4 ----
 ggml-cuda/dmmv.cuh   | 11 +++++++++++
 3 files changed, 11 insertions(+), 10 deletions(-)

diff --git a/ggml-cuda/common.cuh b/ggml-cuda/common.cuh
index 79e1d54..b98d7cb 100644
--- a/ggml-cuda/common.cuh
+++ b/ggml-cuda/common.cuh
@@ -231,12 +231,6 @@ typedef float dfloat; // dequantize float
 typedef float2 dfloat2;
 #endif //GGML_CUDA_F16
 
-// dmmv = dequantize_mul_mat_vec
-// TODO: remove this?
-#ifndef GGML_CUDA_DMMV_X
-#define GGML_CUDA_DMMV_X 32
-#endif
-
 [[noreturn]]
 static __device__ void no_device_code(
     const char * file_name, const int line, const char * function_name, const int arch, const char * arch_list) {
diff --git a/ggml-cuda/dmmv.cu b/ggml-cuda/dmmv.cu
index ea2950f..0b17e3c 100644
--- a/ggml-cuda/dmmv.cu
+++ b/ggml-cuda/dmmv.cu
@@ -2,10 +2,6 @@
 #include "dequantize.cuh"
 #include "convert.cuh"
 
-#ifndef GGML_CUDA_MMV_Y
-#define GGML_CUDA_MMV_Y 1
-#endif
-
 #ifndef K_QUANTS_PER_ITERATION
 #define K_QUANTS_PER_ITERATION 2
 #else
diff --git a/ggml-cuda/dmmv.cuh b/ggml-cuda/dmmv.cuh
index 3802678..4c5ebd4 100644
--- a/ggml-cuda/dmmv.cuh
+++ b/ggml-cuda/dmmv.cuh
@@ -1,5 +1,16 @@
 #include "common.cuh"
 
+// dmmv = dequantize_mul_mat_vec
+
+// TODO: remove this?
+#ifndef GGML_CUDA_DMMV_X
+#define GGML_CUDA_DMMV_X 32
+#endif
+
+#ifndef GGML_CUDA_MMV_Y
+#define GGML_CUDA_MMV_Y 1
+#endif
+
 void ggml_cuda_op_dequantize_mul_mat_vec(
     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,