[PATCH] D73636: [AArch64][SVE] SVE2 intrinsics for complex integer arithmetic
Sander de Smalen via Phabricator via cfe-commits
cfe-commits at lists.llvm.org
Wed Feb 5 05:41:29 PST 2020
sdesmalen added inline comments.
================
Comment at: llvm/include/llvm/IR/IntrinsicsAArch64.td:1758
+def int_aarch64_sve_sqcadd_x : AdvSIMD_2VectorArgIndexed_Intrinsic;
+def int_aarch64_sve_cmla_x : AdvSIMD_3VectorArgIndexed_Intrinsic;
+def int_aarch64_sve_cmla_lane_x : AdvSIMD_SVE_CMLA_LANE_Intrinsic;
----------------
While the shape of AdvSIMD_3VectorArgIndexed_Intrinsic may be the same, the immediate of int_aarch64_sve_cmla_x is not an index. Instead, it is an immediate that expresses the complex rotation. I would just create a separate Intrinsic class for that.
Same for `cadd`, `sqcadd` and `sqrdcmlah`
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D73636/new/
https://reviews.llvm.org/D73636
More information about the cfe-commits
mailing list