01afb3fb99
Also get rid of explicitly specified '-march' values for old architectures. This simplifies %ptxas-verify statements. After the change, we can potentially miss cases where a new functionality is added to the architecture without appropriate checks in the backend. On the other hand, this is mostly true for old architectures that have been thoroughly tested. Differential Revision: https://reviews.llvm.org/D141736
34 lines
1.2 KiB
LLVM
34 lines
1.2 KiB
LLVM
; RUN: llc < %s -march=nvptx64 -mcpu=sm_30 -mattr=+ptx60 | FileCheck %s
|
|
; RUN: %if ptxas %{ llc < %s -march=nvptx64 -mcpu=sm_30 -mattr=+ptx60 | %ptxas-verify %}
|
|
|
|
declare void @llvm.nvvm.bar.warp.sync(i32)
|
|
declare void @llvm.nvvm.barrier.sync(i32)
|
|
declare void @llvm.nvvm.barrier.sync.cnt(i32, i32)
|
|
|
|
; CHECK-LABEL: .func{{.*}}barrier_sync
|
|
define void @barrier_sync(i32 %id, i32 %cnt) {
|
|
; CHECK: ld.param.u32 [[ID:%r[0-9]+]], [barrier_sync_param_0];
|
|
; CHECK: ld.param.u32 [[CNT:%r[0-9]+]], [barrier_sync_param_1];
|
|
|
|
; CHECK: barrier.sync [[ID]], [[CNT]];
|
|
call void @llvm.nvvm.barrier.sync.cnt(i32 %id, i32 %cnt)
|
|
; CHECK: barrier.sync [[ID]], 32;
|
|
call void @llvm.nvvm.barrier.sync.cnt(i32 %id, i32 32)
|
|
; CHECK: barrier.sync 3, [[CNT]];
|
|
call void @llvm.nvvm.barrier.sync.cnt(i32 3, i32 %cnt)
|
|
; CHECK: barrier.sync 4, 64;
|
|
call void @llvm.nvvm.barrier.sync.cnt(i32 4, i32 64)
|
|
|
|
; CHECK: barrier.sync [[ID]];
|
|
call void @llvm.nvvm.barrier.sync(i32 %id)
|
|
; CHECK: barrier.sync 1;
|
|
call void @llvm.nvvm.barrier.sync(i32 1)
|
|
|
|
; CHECK: bar.warp.sync [[ID]];
|
|
call void @llvm.nvvm.bar.warp.sync(i32 %id)
|
|
; CHECK: bar.warp.sync 6;
|
|
call void @llvm.nvvm.bar.warp.sync(i32 6)
|
|
ret void;
|
|
}
|
|
|