b279f6b098
Similar to 806761a7629df268c8aed49657aeccffa6bca449
-mtriple= specifies the full target triple while -march= merely sets the
architecture part of the default target triple (e.g. Windows, macOS),
leaving a target triple which may not make sense.
Therefore, -march= is error-prone and not recommended for tests without
a target triple. The issue has been benign as we recognize
nvptx{,64}-apple-darwin as ELF instead of rejecting it outrightly.
10 lines
283 B
LLVM
10 lines
283 B
LLVM
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_20 -nvptx-prec-divf32=0 | FileCheck %s
|
|
; RUN: %if ptxas %{ llc < %s -mtriple=nvptx64 -mcpu=sm_20 -nvptx-prec-divf32=0 | %ptxas-verify %}
|
|
|
|
define float @foo(float %a) {
|
|
; CHECK: div.approx.f32
|
|
%div = fdiv float %a, 13.0
|
|
ret float %div
|
|
}
|
|
|