diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2016-02-13 01:01:59 +0000 |
|---|---|---|
| committer | Matt Arsenault <Matthew.Arsenault@amd.com> | 2016-02-13 01:01:59 +0000 |
| commit | ba011288a9659d57736956c47b8522eace520807 (patch) | |
| tree | 9b09f077808a5517692f892bf2847663c3881ca4 /amdgcn | |
| parent | b518692b52a0bbdf9cf0e2167b9629dd9501abcd (diff) | |
Split sources for amdgcn and r600
Most files remain in a common amdgpu directory.
Also switches barriers to to use convergent,
and use llvm.amdgcn.s.barrier.
This now requires 3.9/trunk to build amdgcn.
git-svn-id: https://llvm.org/svn/llvm-project/libclc/trunk@260777 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'amdgcn')
| -rw-r--r-- | amdgcn/lib/OVERRIDES | 0 | ||||
| -rw-r--r-- | amdgcn/lib/SOURCES | 1 | ||||
| -rw-r--r-- | amdgcn/lib/synchronization/barrier_impl.ll | 32 |
3 files changed, 33 insertions, 0 deletions
diff --git a/amdgcn/lib/OVERRIDES b/amdgcn/lib/OVERRIDES new file mode 100644 index 0000000..e69de29 --- /dev/null +++ b/amdgcn/lib/OVERRIDES diff --git a/amdgcn/lib/SOURCES b/amdgcn/lib/SOURCES new file mode 100644 index 0000000..c99f3fc --- /dev/null +++ b/amdgcn/lib/SOURCES @@ -0,0 +1 @@ +synchronization/barrier_impl.ll diff --git a/amdgcn/lib/synchronization/barrier_impl.ll b/amdgcn/lib/synchronization/barrier_impl.ll new file mode 100644 index 0000000..1809edd --- /dev/null +++ b/amdgcn/lib/synchronization/barrier_impl.ll @@ -0,0 +1,32 @@ +declare i32 @__clc_clk_local_mem_fence() #1 +declare i32 @__clc_clk_global_mem_fence() #1 +declare void @llvm.amdgcn.s.barrier() #0 + +define void @barrier(i32 %flags) #2 { +barrier_local_test: + %CLK_LOCAL_MEM_FENCE = call i32 @__clc_clk_local_mem_fence() + %0 = and i32 %flags, %CLK_LOCAL_MEM_FENCE + %1 = icmp ne i32 %0, 0 + br i1 %1, label %barrier_local, label %barrier_global_test + +barrier_local: + call void @llvm.amdgcn.s.barrier() + br label %barrier_global_test + +barrier_global_test: + %CLK_GLOBAL_MEM_FENCE = call i32 @__clc_clk_global_mem_fence() + %2 = and i32 %flags, %CLK_GLOBAL_MEM_FENCE + %3 = icmp ne i32 %2, 0 + br i1 %3, label %barrier_global, label %done + +barrier_global: + call void @llvm.amdgcn.s.barrier() + br label %done + +done: + ret void +} + +attributes #0 = { nounwind convergent } +attributes #1 = { nounwind alwaysinline } +attributes #2 = { nounwind convergent alwaysinline } |