Tom Stellard
ec5b9ab433
R600/SI: Fix asam errors in SIFoldOperands
...
We were trying to fold into implicit uses, which led to out of bounds
access of the MCInstrDesc::OpInfo arrray.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229533 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-17 20:11:54 +00:00
Matt Arsenault
e7ff4c1321
R600/SI: Fix phys reg copies in SIFoldOperands
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229227 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-14 02:55:57 +00:00
Matt Arsenault
1751616522
R600/SI: Allow f64 inline immediates in i64 operands
...
This requires considering the size of the operand when
checking immediate legality.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@229135 91177308-0d34-0410-b5e6-96231b3b80d8
2015-02-13 19:05:03 +00:00
Matt Arsenault
ea3e75f4d4
Fix typo
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227697 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-31 23:37:27 +00:00
Tom Stellard
d90e5063ca
R600/SI: Add pattern for bitcasting fp immediates to integers
...
The backend now assumes that all immediates are integers. This allows
us to simplify immediate handling code, becasue we no longer need to
handle fp and integer immediates differently.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225844 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-13 22:59:41 +00:00
Tom Stellard
a36b682c17
R600/SI: Commute instructions to enable more folding opportunities
...
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225410 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-07 22:44:19 +00:00
Tom Stellard
a3ee583339
R600/SI: Only fold immediates that have one use
...
Folding the same immediate into multiple instruction will increase
program size, which can hurt performance.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225405 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-07 22:18:27 +00:00
Tom Stellard
f7587043ef
R600/SI: Add a V_MOV_B64 pseudo instruction
...
This is used to simplify the SIFoldOperands pass and make it easier to
fold immediates.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225373 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-07 20:27:25 +00:00
Tom Stellard
546520a727
R600/SI: Teach SIFoldOperands to split 64-bit constants when folding
...
This allows folding of sequences like:
s[0:1] = s_mov_b64 4
v_add_i32 v0, s0, v0
v_addc_u32 v1, s1, v1
into
v_add_i32 v0, 4, v0
v_add_i32 v1, 0, v1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225369 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-07 19:56:17 +00:00
Tom Stellard
87ce6100a7
R600/SI: Refactor SIFoldOperands to simplify immediate folding
...
This will make a future patch much less intrusive.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@225358 91177308-0d34-0410-b5e6-96231b3b80d8
2015-01-07 17:42:16 +00:00
Tom Stellard
edcd88ce1a
R600/SI: Add SIFoldOperands pass
...
This pass attempts to fold the source operands of mov and copy
instructions into their uses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@222581 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-21 22:06:37 +00:00
