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.
15 lines
754 B
LLVM
15 lines
754 B
LLVM
; RUN: llc < %s -mtriple=nvptx -mcpu=sm_20 | FileCheck %s --check-prefix=PTX32
|
|
; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_20 | FileCheck %s --check-prefix=PTX64
|
|
; RUN: %if ptxas && !ptxas-12.0 %{ llc < %s -mtriple=nvptx -mcpu=sm_20 | %ptxas-verify %}
|
|
; RUN: %if ptxas %{ llc < %s -mtriple=nvptx64 -mcpu=sm_20 | %ptxas-verify %}
|
|
|
|
; PTX32: .visible .global .align 4 .u32 i;
|
|
; PTX32: .visible .const .align 4 .u32 j;
|
|
; PTX32: .visible .shared .align 4 .u32 k;
|
|
; PTX64: .visible .global .align 4 .u32 i;
|
|
; PTX64: .visible .const .align 4 .u32 j;
|
|
; PTX64: .visible .shared .align 4 .u32 k;
|
|
@i = addrspace(1) externally_initialized global i32 0, align 4
|
|
@j = addrspace(4) externally_initialized global i32 0, align 4
|
|
@k = addrspace(3) global i32 undef, align 4
|