(Bugfix, ggml-cuda) Pool alloc count fix + small size computation type adjustment (#18559)

* CUDA: Fixed obj byte size instead of obj count being passed to pool alloc (fattn-common, dst_tmp_meta)

* CUDA: Explicitly casted some of the int alloc counts before multiplication in argsort

---------

Co-authored-by: pl752 <maximpl752@gmail.com>

ggml/src/ggml-cuda/argsort.cu

@@ -29,8 +29,8 @@ static void argsort_f32_i32_cuda_cub(ggml_cuda_pool & pool,
                                      const int        nrows,
                                      ggml_sort_order  order,
                                      cudaStream_t     stream) {
-    ggml_cuda_pool_alloc<int>   temp_indices_alloc(pool, ncols * nrows);
-    ggml_cuda_pool_alloc<float> temp_keys_alloc(pool, ncols * nrows);
+    ggml_cuda_pool_alloc<int>   temp_indices_alloc(pool, ((size_t) ncols) * nrows);
+    ggml_cuda_pool_alloc<float> temp_keys_alloc(pool, ((size_t) ncols) * nrows);
     ggml_cuda_pool_alloc<int>   offsets_alloc(pool, nrows + 1);
 
     int *   temp_indices = temp_indices_alloc.get();

ggml/src/ggml-cuda/fattn-common.cuh

@@ -918,7 +918,7 @@ void launch_fattn(
         blocks_num.y = 1;
         blocks_num.z = 1;
 
-        dst_tmp_meta.alloc(blocks_num.x*ncols * (2*2 + DV) * sizeof(float));
+        dst_tmp_meta.alloc(((size_t) blocks_num.x) * ncols * (2 + DV/2));
     } else {
         const int ntiles_KQ = (K->ne[1] + nbatch_fa - 1) / nbatch_fa; // Max. number of parallel blocks limited by tensor size.