[Mesa-dev] [PATCH 2/2] radeon/llvm: Add a v4f32 to v4i32 BitConvert pattern

Vincent Lejeune vljn at ovi.com
Thu Oct 25 14:18:00 PDT 2012


---
 lib/Target/AMDGPU/R600Instructions.td | 1 +
 1 file changed, 1 insertion(+)

diff --git a/lib/Target/AMDGPU/R600Instructions.td b/lib/Target/AMDGPU/R600Instructions.td
index 5b97e5e..2f9ac0c 100644
--- a/lib/Target/AMDGPU/R600Instructions.td
+++ b/lib/Target/AMDGPU/R600Instructions.td
@@ -1487,5 +1487,6 @@ def : Vector_Build <v4i32, R600_Reg128, i32, R600_Reg32>;
 def : BitConvert <i32, f32, R600_Reg32>;
 def : BitConvert <f32, i32, R600_Reg32>;
 def : BitConvert <v4f32, v4i32, R600_Reg128>;
+def : BitConvert <v4i32, v4f32, R600_Reg128>;
 
 } // End isR600toCayman Predicate
-- 
1.7.11.7



More information about the mesa-dev mailing list