]> Git Repo - qemu.git/commit
target/arm: Vectorize USHL and SSHL
authorRichard Henderson <[email protected]>
Sun, 16 Feb 2020 21:42:29 +0000 (13:42 -0800)
committerPeter Maydell <[email protected]>
Fri, 21 Feb 2020 16:07:02 +0000 (16:07 +0000)
commit87b74e8b6edd287ea2160caa0ebea725fa8f1ca1
treeacbe40c42da070d88cc77e7372cc43b2a55824a3
parent7abc8cabad977aeccbbb6e6b2026e68ab8e32c65
target/arm: Vectorize USHL and SSHL

These instructions shift left or right depending on the sign
of the input, and 7 bits are significant to the shift.  This
requires several masks and selects in addition to the actual
shifts to form the complete answer.

That said, the operation is still a small improvement even for
two 64-bit elements -- 13 vector operations instead of 2 * 7
integer operations.

Reviewed-by: Alex BennĂ©e <[email protected]>
Signed-off-by: Richard Henderson <[email protected]>
Message-id: 20200216214232[email protected]
Signed-off-by: Peter Maydell <[email protected]>
target/arm/helper.h
target/arm/neon_helper.c
target/arm/translate-a64.c
target/arm/translate.c
target/arm/translate.h
target/arm/vec_helper.c
This page took 0.025317 seconds and 4 git commands to generate.