llvm/test/CodeGen/NVPTX/ld-generic.ll
Justin Holewinski 49683f3c96 This patch adds a new NVPTX back-end to LLVM which supports code generation for NVIDIA PTX 3.0. This back-end will (eventually) replace the current PTX back-end, while maintaining compatibility with it.
The new target machines are:

nvptx (old ptx32) => 32-bit PTX
nvptx64 (old ptx64) => 64-bit PTX

The sources are based on the internal NVIDIA NVPTX back-end, and
contain more functionality than the current PTX back-end currently
provides.

NV_CONTRIB

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156196 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04 20:18:50 +00:00

64 lines
1.5 KiB
LLVM

; RUN: llc < %s -march=nvptx -mcpu=sm_20 | FileCheck %s --check-prefix=PTX32
; RUN: llc < %s -march=nvptx64 -mcpu=sm_20 | FileCheck %s --check-prefix=PTX64
;; i8
define i8 @ld_global_i8(i8 addrspace(0)* %ptr) {
; PTX32: ld.u8 %rc{{[0-9]+}}, [%r{{[0-9]+}}]
; PTX32: ret
; PTX64: ld.u8 %rc{{[0-9]+}}, [%rl{{[0-9]+}}]
; PTX64: ret
%a = load i8 addrspace(0)* %ptr
ret i8 %a
}
;; i16
define i16 @ld_global_i16(i16 addrspace(0)* %ptr) {
; PTX32: ld.u16 %rs{{[0-9]+}}, [%r{{[0-9]+}}]
; PTX32: ret
; PTX64: ld.u16 %rs{{[0-9]+}}, [%rl{{[0-9]+}}]
; PTX64: ret
%a = load i16 addrspace(0)* %ptr
ret i16 %a
}
;; i32
define i32 @ld_global_i32(i32 addrspace(0)* %ptr) {
; PTX32: ld.u32 %r{{[0-9]+}}, [%r{{[0-9]+}}]
; PTX32: ret
; PTX64: ld.u32 %r{{[0-9]+}}, [%rl{{[0-9]+}}]
; PTX64: ret
%a = load i32 addrspace(0)* %ptr
ret i32 %a
}
;; i64
define i64 @ld_global_i64(i64 addrspace(0)* %ptr) {
; PTX32: ld.u64 %rl{{[0-9]+}}, [%r{{[0-9]+}}]
; PTX32: ret
; PTX64: ld.u64 %rl{{[0-9]+}}, [%rl{{[0-9]+}}]
; PTX64: ret
%a = load i64 addrspace(0)* %ptr
ret i64 %a
}
;; f32
define float @ld_global_f32(float addrspace(0)* %ptr) {
; PTX32: ld.f32 %f{{[0-9]+}}, [%r{{[0-9]+}}]
; PTX32: ret
; PTX64: ld.f32 %f{{[0-9]+}}, [%rl{{[0-9]+}}]
; PTX64: ret
%a = load float addrspace(0)* %ptr
ret float %a
}
;; f64
define double @ld_global_f64(double addrspace(0)* %ptr) {
; PTX32: ld.f64 %fl{{[0-9]+}}, [%r{{[0-9]+}}]
; PTX32: ret
; PTX64: ld.f64 %fl{{[0-9]+}}, [%rl{{[0-9]+}}]
; PTX64: ret
%a = load double addrspace(0)* %ptr
ret double %a
}