[llvm] [LLVM][NVPTX] Add support for div.full instruction (PR #116482)

Pradeep Kumar via llvm-commits llvm-commits at lists.llvm.org
Thu Nov 21 03:50:43 PST 2024


================
@@ -1096,6 +1096,18 @@ def INT_NVVM_DIV_RM_D : F_MATH_2<"div.rm.f64 \t$dst, $src0, $src1;",
 def INT_NVVM_DIV_RP_D : F_MATH_2<"div.rp.f64 \t$dst, $src0, $src1;",
   Float64Regs, Float64Regs, Float64Regs, int_nvvm_div_rp_d>;
 
+def : Pat<(int_nvvm_div_full Float32Regs:$a, Float32Regs:$b),
----------------
schwarzschild-radius wrote:

@Artem-B I tried setting `--nvptx-prec-divf32=`  to different values and it does not fail. It only failed after adding the same `Requires` clause to the pattern that I notice the Instruction Selection to fail. Seems like Patterns don't inherit the `Requires` clause. What do you think?

https://github.com/llvm/llvm-project/pull/116482


More information about the llvm-commits mailing list