projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: d3b7528) | patch
[ARM][BFloat] Lowering of create/get/set/dup intrinsics
authorMikhail Maltsev <mikhail.maltsev@arm.com>
Fri, 19 Jun 2020 12:49:17 +0000 (12:49 +0000)
committerMikhail Maltsev <mikhail.maltsev@arm.com>
Fri, 19 Jun 2020 12:52:40 +0000 (12:52 +0000)
commit75268812464f86e5a6cf583ea5c04e71f7cbb680
tree08debaf7320849696641094704b20552fa2fc3c4tree | snapshot
parentd3b752845df0331348dad48000fc8b82afb3de5bcommit | diff
[ARM][BFloat] Lowering of create/get/set/dup intrinsics

This patch adds codegen for the following BFloat
operations to the ARM backend:
* concatenation of bf16 vectors
* bf16 vector element extraction
* bf16 vector element insertion
* duplication of a bf16 value into each lane of a vector
* duplication of a bf16 vector lane into each lane

Differential Revision: https://reviews.llvm.org/D81411
llvm/lib/Target/ARM/ARMInstrNEON.td diff | blob | history
llvm/test/CodeGen/ARM/bf16-create-get-set-dup.ll [new file with mode: 0644] blob
llvm/test/CodeGen/ARM/bf16-getlane-with-fp16.ll [new file with mode: 0644] blob
llvm/test/CodeGen/ARM/fp16-insert-extract.ll diff | blob | history
Domain: System / Toolchain;