diff options
| author | Vasileios Kalintiris <Vasileios.Kalintiris@imgtec.com> | 2015-08-11 08:56:25 +0000 |
|---|---|---|
| committer | Vasileios Kalintiris <Vasileios.Kalintiris@imgtec.com> | 2015-08-11 08:56:25 +0000 |
| commit | 1c78ca6a09cec862d6aad04e820dae98b71be9fc (patch) | |
| tree | ef6def2a0475fbea2672cf161260dc2830857755 /llvm/test/MC/Mips/mips32/valid.s | |
| parent | 7337ee23d8ee9864d6b5f4140aaa56cf68388361 (diff) | |
| download | bcm5719-llvm-1c78ca6a09cec862d6aad04e820dae98b71be9fc.tar.gz bcm5719-llvm-1c78ca6a09cec862d6aad04e820dae98b71be9fc.zip | |
[mips] Remap move as or.
Summary:
This patch remaps the assembly idiom 'move' to 'or' instead of 'daddu' or
'addu'. The use of addu/daddu instead of or as move was highlighted as a
performance issue during the analysis of a recent 64bit design. Originally
move was encoded as 'or' by binutils but was changed for the r10k cpu family
due to their pipeline which had 2 arithmetic units and a single logical unit,
and so could issue multiple (d)addu based moves at the same time but only 1
logical move.
This patch preserves the disassembly behaviour so that disassembling a old style
(d)addu move still appears as move, but assembling move always gives an or
Patch by Simon Dardis.
Reviewers: vkalintiris
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11796
llvm-svn: 244579
Diffstat (limited to 'llvm/test/MC/Mips/mips32/valid.s')
| -rw-r--r-- | llvm/test/MC/Mips/mips32/valid.s | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/llvm/test/MC/Mips/mips32/valid.s b/llvm/test/MC/Mips/mips32/valid.s index fbe1551f6c2..6958583030a 100644 --- a/llvm/test/MC/Mips/mips32/valid.s +++ b/llvm/test/MC/Mips/mips32/valid.s @@ -96,8 +96,8 @@ a: mflo $s1 mov.d $f20,$f14 mov.s $f2,$f27 - move $s8,$a0 - move $25,$a2 + move $s8,$a0 # CHECK: move $fp, $4 # encoding: [0x00,0x80,0xf0,0x25] + move $25,$a2 # CHECK: move $25, $6 # encoding: [0x00,0xc0,0xc8,0x25] movf $gp,$8,$fcc7 movf.d $f6,$f11,$fcc5 movf.s $f23,$f5,$fcc6 |
