cuda : fix compile warning (ggml-org#7454)

ggerganov · teleprint-me · commit 24bdeaa1a0b6 · 2024-05-23T13:00:47.000-04:00
diff --git a/ggml-cuda/fattn-tile-f32.cu b/ggml-cuda/fattn-tile-f32.cu
@@ -283,8 +283,7 @@ void launch_fattn_tile_f32_64_128(ggml_backend_cuda_context & ctx, ggml_tensor *
 }
 
 void ggml_cuda_flash_attn_ext_tile_f32(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {
-    const ggml_tensor * KQV = dst;
-    const ggml_tensor * Q   = dst->src[0];
+    const ggml_tensor * Q = dst->src[0];
 
     if (Q->ne[1] <= 16) {
         constexpr int cols_per_block = 16;

Original file line number	Diff line number	Diff line change
`@@ -283,8 +283,7 @@ void launch_fattn_tile_f32_64_128(ggml_backend_cuda_context & ctx, ggml_tensor *`
`283`	`283`	`}`
`284`	`284`
`285`	`285`	`void ggml_cuda_flash_attn_ext_tile_f32(ggml_backend_cuda_context & ctx, ggml_tensor * dst) {`
`286`		`- const ggml_tensor * KQV = dst;`
`287`		`- const ggml_tensor * Q = dst->src[0];`
	`286`	`+ const ggml_tensor * Q = dst->src[0];`
`288`	`287`
`289`	`288`	`if (Q->ne[1] <= 16) {`
`290`	`289`	`constexpr int cols_per_block = 16;`