llvm-6502/test/CodeGen/NVPTX/fp-literals.ll
Justin Holewinski 11ae250ec9 [NVPTX] Improve handling of FP fusion
We now consider the FPOpFusion flag when determining whether
to fuse ops.  We also explicitly emit add.rn when fusion is
disabled to prevent ptxas from fusing the operations on its
own.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213287 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-17 18:10:09 +00:00

22 lines
690 B
LLVM

; RUN: llc < %s -march=nvptx -mcpu=sm_20 -fp-contract=fast | FileCheck %s
target triple = "nvptx64-unknown-cuda"
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v16:16:16-v32:32:32-v64:64:64-v128:128:128-n16:32:64"
; Make sure we can properly differentiate between single-precision and
; double-precision FP literals.
; CHECK: myaddf
; CHECK: add.f32 %f{{[0-9]+}}, %f{{[0-9]+}}, 0f3F800000
define float @myaddf(float %a) {
%ret = fadd float %a, 1.0
ret float %ret
}
; CHECK: myaddd
; CHECK: add.f64 %fd{{[0-9]+}}, %fd{{[0-9]+}}, 0d3FF0000000000000
define double @myaddd(double %a) {
%ret = fadd double %a, 1.0
ret double %ret
}