[PATCH] D124357: [AArch64] Cost modelling for fptoi_sat

[Eli Friedman via Phabricator via llvm-commits]

efriedma added inline comments.


================
Comment at: llvm/test/Analysis/CostModel/AArch64/fptoi_sat.ll:325
 ; CHECK-FP16-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v16f16u32 = call <16 x i32> @llvm.fptoui.sat.v16i32.v16f16(<16 x half> undef)
 ; CHECK-FP16-NEXT:  Cost Model: Found an estimated cost of 400 for instruction: %v16f16s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f16(<16 x half> undef)
 ; CHECK-FP16-NEXT:  Cost Model: Found an estimated cost of 78 for instruction: %v16f16u64 = call <16 x i64> @llvm.fptoui.sat.v16i64.v16f16(<16 x half> undef)
----------------
Any idea what's going on with this "400"?  Even with our current terrible unrolled codegen, it's still only 50 instructions.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D124357/new/

https://reviews.llvm.org/D124357