Fix #431.

Summary: Added missing include for cstdint for Windows and removed problematic inline assembly.

Reviewed By: bottler

Differential Revision: D24838053

fbshipit-source-id: 95496be841c2c22a82068073d4740e98ee8a02ac

Author: classner

pytorch3d/csrc/pulsar/cuda/commands.h

@@ -209,10 +209,6 @@ __device__ static float atomicMin(float* address, float val) {
 #define RNORM3DF(x, y, z) rnorm3df(x, y, z)
 
 // High level.
-INLINE DEVICE void prefetch_l1(unsigned long addr) {
-  asm(" prefetch.global.L1 [ %1 ];" : "=l"(addr) : "l"(addr));
-}
-#define PREFETCH(PTR) prefetch_l1((unsigned long)(PTR))
 #define GET_SORT_WS_SIZE(RES_PTR, KEY_TYPE, VAL_TYPE, NUM_OBJECTS) \
   cub::DeviceRadixSort::SortPairsDescending(                       \
       (void*)NULL,                                                 \

pytorch3d/csrc/pulsar/include/camera.h

@@ -2,6 +2,7 @@
 #ifndef PULSAR_NATIVE_INCLUDE_CAMERA_H_
 #define PULSAR_NATIVE_INCLUDE_CAMERA_H_
 
+#include <stdint.h>
 #include "../global.h"
 
 namespace pulsar {