0a6a1f1d05
This brings our tree to NetBSD 7.0, as found on -current on the 10-10-2015. This updates: - LLVM to 3.6.1 - GCC to GCC 5.1 - Replace minix/commands/zdump with usr.bin/zdump - external/bsd/libelf has moved to /external/bsd/elftoolchain/ - Import ctwm - Drop sprintf from libminc Change-Id: I149836ac18e9326be9353958bab9b266efb056f0
31 lines
769 B
Plaintext
31 lines
769 B
Plaintext
// RUN: %clang_cc1 %s -triple nvptx-unknown-unknown -fcuda-is-device -emit-llvm -o - | FileCheck %s
|
|
|
|
#include "Inputs/cuda.h"
|
|
|
|
#define MAX_THREADS_PER_BLOCK 256
|
|
#define MIN_BLOCKS_PER_MP 2
|
|
|
|
// Test both max threads per block and Min cta per sm.
|
|
extern "C" {
|
|
__global__ void
|
|
__launch_bounds__( MAX_THREADS_PER_BLOCK, MIN_BLOCKS_PER_MP )
|
|
Kernel1()
|
|
{
|
|
}
|
|
}
|
|
|
|
// CHECK: !{{[0-9]+}} = !{void ()* @Kernel1, !"maxntidx", i32 256}
|
|
// CHECK: !{{[0-9]+}} = !{void ()* @Kernel1, !"minctasm", i32 2}
|
|
|
|
// Test only max threads per block. Min cta per sm defaults to 0, and
|
|
// CodeGen doesn't output a zero value for minctasm.
|
|
extern "C" {
|
|
__global__ void
|
|
__launch_bounds__( MAX_THREADS_PER_BLOCK )
|
|
Kernel2()
|
|
{
|
|
}
|
|
}
|
|
|
|
// CHECK: !{{[0-9]+}} = !{void ()* @Kernel2, !"maxntidx", i32 256}
|