Commit Graph

109 Commits

Author SHA1 Message Date
Michel Danzer 0cc991e17b R600/SI: Add pattern for sign extension of i1 to i32.
16 more little piglits with radeonsi.

NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175887
2013-02-22 11:22:58 +00:00
Michel Danzer 00fb283560 R600/SI: Add pattern for logical or of i1 values.
24 more little piglits with radeonsi.

NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175886
2013-02-22 11:22:54 +00:00
Michel Danzer c3ea4041b9 R600/SI: Add pattern for fceil.
9 more little piglits with radeonsi.

NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175885
2013-02-22 11:22:49 +00:00
Christian Konig 71088e68e8 R600/SI: inline V_ADD|SUB_F32 patterns
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175758
2013-02-21 15:17:41 +00:00
Christian Konig 7c9de8e6e8 R600/SI: replace IMPLICIT_DEF with SIOperand.ZERO
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175757
2013-02-21 15:17:36 +00:00
Christian Konig 2aca043312 R600/SI: replace SI_V_CNDLT with a pattern
It actually fixes quite a bunch of piglit tests.

This is a candidate for the mesa-stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175756
2013-02-21 15:17:32 +00:00
Christian Konig 8dbe6f617c R600/SI: use patterns for clamp, fabs, fneg
Instead of using custom inserters, it's simpler and
should make DAG folding easier.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175755
2013-02-21 15:17:27 +00:00
Christian Konig bf114b42a8 R600/SI: add all the other missing asm operands v2
v2: put implicit parameters in []

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175754
2013-02-21 15:17:22 +00:00
Christian Konig 08e768b4cf R600/SI: add the missing M*BUF|IMG asm operands
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175753
2013-02-21 15:17:17 +00:00
Christian Konig e0130a2f25 R600/SI: add the missing S_* asm operands
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175752
2013-02-21 15:17:13 +00:00
Christian Konig f5754a011d R600/SI: rework VOP3 classes
Order the classes and add asm operands.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175751
2013-02-21 15:17:09 +00:00
Christian Konig b19849a682 R600/SI: simplify VOPC_* pattern v2
Fixing asm operation names.

v2: fix name of the e64 encoding, also add asm operands

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175750
2013-02-21 15:17:04 +00:00
Christian Konig ae034e63f1 R600/SI: rework VOP2_* pattern v2
Fixing asm operation names.

v2: use ZERO constant, also add asm operands

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175749
2013-02-21 15:16:58 +00:00
Christian Konig 3da7017e81 R600/SI: rework VOP1_* patterns v2
Fixing asm operation names.

v2: use ZERO constant, also add asm operands

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175748
2013-02-21 15:16:53 +00:00
Christian Konig eabf8333d6 R600/SI: add constant for inline zero operand
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175747
2013-02-21 15:16:49 +00:00
Christian Konig 72d5d5c754 R600/SI: cleanup SIInstrInfo.td and SIInstrFormat.td
Those two files got mixed up.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 175746
2013-02-21 15:16:44 +00:00
Tom Stellard 0d171c8877 R600: Fix for Unigine when MachineSched is enabled
Fixes for-loop.cl piglit test

Patch By: Vincent Lejeune

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175742
2013-02-21 15:06:59 +00:00
Michel Danzer 7f02a8c7a7 R600/SI: Make sure M0 is loaded for V_INTERP_MOV_F32
NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175733
2013-02-21 08:57:10 +00:00
Jim Grosbach d15cd2a11c R600: Update for name changes from r175667.
llvm-svn: 175668
2013-02-20 21:31:28 +00:00
Tom Stellard d4409e2cec R600: Add AR_X to the R600_TReg_X register class.
NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175519
2013-02-19 15:22:47 +00:00
Tom Stellard a24a516737 R600: Mark all members of the TRegMem register class as reserved
This stops the Machine Verifier from complaining about uses of undefined
physical registers.

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175518
2013-02-19 15:22:45 +00:00
Tom Stellard 8d469edbe3 R600: Fix scheduler crash caused by invalid MachinePointerInfo
Kernel function arguments are lowered to loads from the PARAM_I address
space.  When creating these load instructions, we were initializing
their MachinePointerInfo with an Arguement object that was not attached
to any function.  This was causing the MachineScheduler to crash when
it tried to access the parent of the Arguement.

This has been fixed by initializing the MachinePointerInfo with a
UndefValue instead.

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175517
2013-02-19 15:22:44 +00:00
Tom Stellard 0f965aaf9b R600: Fix tracking of implicit defs in the IndirectAddressing pass
In some cases, we were losing track of live implicit registers which
was creating dead defs and causing the scheduler to produce invalid
code.

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175516
2013-02-19 15:22:42 +00:00
David Blaikie 772d4f75f6 Use LLVM_DELETED_FUNCTION rather than '// do not implement' comments.
Also removes some redundant DNI comments on function declarations already
using the macro.

llvm-svn: 175466
2013-02-18 23:11:17 +00:00
Vincent Lejeune 1ce13f553e R600/SI: Use MULADD_IEEE/V_MAD_F32 instruction for mad pattern
llvm-svn: 175446
2013-02-18 14:11:28 +00:00
Vincent Lejeune 685018009b R600: Support for TBO
NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
llvm-svn: 175445
2013-02-18 14:11:19 +00:00
Vincent Lejeune 4c1602b5c9 R600: Increase number of ArrayBase Reg to 32
Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
llvm-svn: 175443
2013-02-18 13:48:09 +00:00
NAKAMURA Takumi eddbc713e1 Target/R600/CMakeLists.txt: Prune SILowerLiteralConstants.cpp corresponding to r175354.
llvm-svn: 175361
2013-02-16 15:30:28 +00:00
Christian Konig b559b079b4 R600/SI: Add pattern to simplify i64 loading
This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175356
2013-02-16 11:28:36 +00:00
Christian Konig a881179ffe R600/SI: nuke SReg_1 v3
It's completely unnecessary and can be replace with proper
SReg_64 handling instead.

This actually fixes a piglit test on SI.

v2: use correct register class in addRegisterClass,
    set special classes as not allocatable
v3: revert setting special classes as not allocateable

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175355
2013-02-16 11:28:30 +00:00
Christian Konig c756cb9901 R600/SI: cleanup literal handling v3
Seems to be allot simpler, and also paves the
way for further improvements.

v2: rebased on master, use 0 in BUFFER_LOAD_FORMAT_XYZW,
    use VGPR0 in dummy EXP, avoid compiler warning, break
    after encoding the first literal.
v3: correctly use V_ADD_F32_e64

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175354
2013-02-16 11:28:22 +00:00
Christian Konig b9e281a723 R600/SI: replace AllReg_* with [SV]Src_* v2
Mark all the operands that can also have an immediate.

v2: SOFFSET is also an SSrc_32 operand

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175353
2013-02-16 11:28:13 +00:00
Christian Konig 3c3a7bfb06 R600/SI: fix VOPC encoding v2
Previously it only worked because of coincident.

v2: fix 64bit versions, use 0x80 (inline 0) instead of SGPR0
    for the unused SRC2

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175352
2013-02-16 11:28:07 +00:00
Christian Konig e3cba88714 R600/SI: move *_Helper definitions to SIInstrFormat.td
This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175351
2013-02-16 11:28:02 +00:00
Christian Konig 8590c1e371 R600/SI: remove some more unused code
This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175350
2013-02-16 11:27:56 +00:00
Christian Konig d886099f13 R600/structurizer: improve inverting conditions
Stop adding more instructions than necessary.

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175349
2013-02-16 11:27:50 +00:00
Christian Konig fc6a985c12 R600/structurizer: improve loop handling
Generate more than one loop if it seems to make sense.

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175348
2013-02-16 11:27:45 +00:00
Christian Konig b5d8866b84 R600/structurizer: improve finding condition values
Using the new NearestCommonDominator class.

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175347
2013-02-16 11:27:40 +00:00
Christian Konig 0bccf9d60b R600/structurizer: improve PHI value finding
Using the new NearestCommonDominator class.

This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175346
2013-02-16 11:27:35 +00:00
Christian Konig d08e3d753e R600/structurizer: add class to find the Nearest Common Dominator
This is a candidate for the stable branch.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175345
2013-02-16 11:27:29 +00:00
Michel Danzer e9bb18b555 R600/SI: Fix int_SI_fs_interp_constant
The important fix is that the constant interpolation value is stored in the
parameter slot P0, which is encoded as 2.

In addition, drop the SI_INTERP_CONST pseudo instruction, pass the parameter
slot as an operand to V_INTERP_MOV_F32 instead of hardcoding it there, and
add a special operand class for the parameter slots for type checking and
pretty printing.

NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175193
2013-02-14 19:03:25 +00:00
Vincent Lejeune f940fd05bd R600: Do not fold single instruction with more that 3 kcache read
It fixes around 100 tfb piglit tests and 16 glean tests.

NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
llvm-svn: 175183
2013-02-14 16:57:19 +00:00
Vincent Lejeune ea710fe419 R600: Export instructions are no longer terminator
This allows MachineInstScheduler to reorder them, and thus make scheduling more
efficient.

Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
llvm-svn: 175182
2013-02-14 16:55:11 +00:00
Vincent Lejeune d80bc1561a R600: Fold zero/one in export instructions
Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
llvm-svn: 175181
2013-02-14 16:55:06 +00:00
Vincent Lejeune f694c10c8e R600: Do not fold modifier/litterals in vector inst
This fixes a couple of regressions on (probably not just) cayman

NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
llvm-svn: 175180
2013-02-14 16:55:01 +00:00
Michel Danzer ae0a403dab R600/SI: Check for empty stack in SIAnnotateControlFlow::isTopOfStack
Fixes assertion failure in newly added lit test. Might just be a bandaid that
needs to be revisited.

llvm-svn: 175139
2013-02-14 08:00:33 +00:00
Tom Stellard 91da4e9199 R600: Add support for 128-bit parameters
NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175096
2013-02-13 22:05:20 +00:00
Michel Danzer 3bb17ebd93 R600: Fix regression with shadow array sampler on pre-SI GPUs.
'R600/SI: Use proper instructions for array/shadow samplers.' removed two
cases from TEX_SHADOW. Vincent Lejeune reported on IRC that this broke some
shadow array piglit tests with the r600g driver. Reinstating the removed
cases should fix this, and still works with radeonsi as well.

I will follow up with some lit tests which would have caught the regression.

NOTE: This is a candidate for the Mesa stable branch.

Tested-by: Vincent Lejeune <vljn@ovi.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174963
2013-02-12 12:11:23 +00:00
Michel Danzer 10ed47f927 R600/SI: Use V_ADD_F32 instead of V_MOV_B32 for clamp/neg/abs modifiers.
The modifiers don't seem to have any effect with V_MOV_B32, supposedly it's
meant to just move bits untouched.

Fixes 46 piglit tests with radeonsi, though unfortunately 11 of those had
just regressed because they started using the clamp modifier.

NOTE: This is a candidate for the Mesa stable branch.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174890
2013-02-11 15:58:21 +00:00
Vincent Lejeune 44bf8158c5 Test Commit - Remove some trailing whitespace in R600Instructions.td
llvm-svn: 174839
2013-02-10 17:57:33 +00:00