projects / gem5.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 57b9f53) | patch
x86: implement rcpps and rcpss SSE insts
authorSteve Reinhardt <steve.reinhardt@amd.com>
Wed, 7 Oct 2015 00:26:50 +0000 (17:26 -0700)
committerSteve Reinhardt <steve.reinhardt@amd.com>
Wed, 7 Oct 2015 00:26:50 +0000 (17:26 -0700)
commita2c875c746a7b9b5dcb94fd93d94ab70286dbbb4
tree03ac1c0befec0a164e233b655759efac0f3207c0tree
parent57b9f53afa5660152a77b7f3b7affb39f5b0e176commit | diff
x86: implement rcpps and rcpss SSE insts

These are packed single-precision approximate reciprocal operations,
vector and scalar versions, respectively.

This code was basically developed by copying the code for
sqrtps and sqrtss.  The mrcp micro-op was simplified relative to
msqrt since there are no double-precision versions of this operation.
src/arch/x86/isa/decoder/two_byte_opcodes.isa diff | blob | history
src/arch/x86/isa/insts/simd128/floating_point/arithmetic/reciprocal_estimation.py diff | blob | history
src/arch/x86/isa/microops/mediaop.isa diff | blob | history