fix msvc build

InternLM · Jan 1, 2025 · 5266f27 · 5266f27
1 parent 54df9f1
commit 5266f27
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 4 deletions.
diff --git a/src/turbomind/engine/request.h b/src/turbomind/engine/request.h
@@ -4,6 +4,8 @@
 
 #include <atomic>
 #include <cstdint>
+#include <functional>
+#include <memory>
 #include <ostream>
 
 #include "src/turbomind/utils/Tensor.h"
@@ -107,7 +109,8 @@ struct Request {
 
     int ec;  // set when disabling conflicting requests
 
-    enum {
+    enum
+    {
         kOk       = 0,
         kInvalid  = 1,  // Sequence not exist or both `start` & `stop` (instead of `end`) is set
         kConflict = 2,  // Concurrent requests to the same sequence

diff --git a/src/turbomind/kernels/sampling_topp_kernels.cu b/src/turbomind/kernels/sampling_topp_kernels.cu
@@ -22,6 +22,7 @@
 #include "3rdparty/cub/cub.cuh"
 #endif
 
+#include "src/turbomind/kernels/core/math.h"
 #include "src/turbomind/kernels/reduce_kernel_utils.cuh"
 #include "src/turbomind/kernels/sampling_topp_kernels.h"
 #include "src/turbomind/utils/constant.h"
@@ -216,9 +217,9 @@ void invokeTopPSort(TopPSortParams& params, cudaStream_t stream)
     size_t topp_id_val_buf_size  = sizeof(int) * params.batch_size * params.vocab_size_padded;
     size_t begin_offset_buf_size = sizeof(int) * params.batch_size;
     size_t end_offset_buf_size   = sizeof(int) * params.batch_size;
-    topp_id_val_buf_size         = div_up(topp_id_val_buf_size, 256UL) * 256;
-    begin_offset_buf_size        = div_up(begin_offset_buf_size, 256UL) * 256;
-    end_offset_buf_size          = div_up(end_offset_buf_size, 256UL) * 256;
+    topp_id_val_buf_size         = div_up<size_t>(topp_id_val_buf_size, 256) * 256;
+    begin_offset_buf_size        = div_up<size_t>(begin_offset_buf_size, 256) * 256;
+    end_offset_buf_size          = div_up<size_t>(end_offset_buf_size, 256) * 256;
 
     if (params.workspace == nullptr) {
         size_t cub_temp_storage_size;