For store->load dependencies that may alias, we should always use
TrueMemOrderLatency, which may eventually become a subtarget hook. In
effect, we should guarantee at least TrueMemOrderLatency on at least
one DAG path from a store to a may-alias load.
This should fix the standard mode as well as -enable-aa-sched-mi".
llvm-svn: 158380
We turned off the CMN instruction because it had semantics which we weren't
getting correct. If we are comparing with an immediate, then it's okay to use
the CMN instruction.
<rdar://problem/7569620>
llvm-svn: 158302
Over the entire test-suite, this has an insignificantly negative average
performance impact, but reduces some of the worst slowdowns from the
anti-dep. change (r158294).
Largest speedups:
SingleSource/Benchmarks/Stanford/Quicksort - 28%
SingleSource/Benchmarks/Stanford/Towers - 24%
SingleSource/Benchmarks/Shootout-C++/matrix - 23%
MultiSource/Benchmarks/SciMark2-C/scimark2 - 19%
MultiSource/Benchmarks/MiBench/automotive-bitcount/automotive-bitcount - 15%
(matrix and automotive-bitcount were both in the top-5 slowdown list from the
anti-dep. change)
Largest slowdowns:
MultiSource/Benchmarks/McCat/03-testtrie/testtrie - 28%
MultiSource/Benchmarks/mediabench/gsm/toast/toast - 26%
MultiSource/Benchmarks/MiBench/automotive-susan/automotive-susan - 21%
SingleSource/Benchmarks/CoyoteBench/lpbench - 20%
MultiSource/Applications/d/make_dparser - 16%
llvm-svn: 158296
The PPC64 backend had patterns for i32 <-> i64 extensions and truncations that
would leave self-moves in the final assembly. Replacing those patterns with ones
based on the SUBREG builtins yields better-looking code.
Thanks to Jakob and Owen for their suggestions in this matter.
llvm-svn: 158283
Tail merging had been disabled on PPC because it would disturb bundling decisions
made during pre-RA scheduling on the 970 cores. Now, however, all bundling decisions
are made during post-RA scheduling, and tail merging is generally beneficial (the
average test-suite speedup is insignificantly positive).
Largest test-suite speedups:
MultiSource/Benchmarks/mediabench/gsm/toast/toast - 30%
MultiSource/Benchmarks/BitBench/uuencode/uuencode - 23%
SingleSource/Benchmarks/Shootout-C++/ary - 21%
SingleSource/Benchmarks/Stanford/Queens - 17%
Largest slowdowns:
MultiSource/Benchmarks/MiBench/security-sha/security-sha - 24%
MultiSource/Benchmarks/McCat/03-testtrie/testtrie - 22%
MultiSource/Applications/JM/ldecod/ldecod - 14%
MultiSource/Benchmarks/mediabench/g721/g721encode/encode - 9%
This is improved by using full (instead of just critical) anti-dependency breaking,
but doing so still causes miscompiles and so cannot yet be enabled by default.
llvm-svn: 158259
The fast register allocator is not supposed to work in the optimizing
pipeline. It doesn't make sense to compute live intervals, run full copy
coalescing, and then run RAFast.
Fast register allocation in the optimizing pipeline is better done by
RABasic.
llvm-svn: 158242
Thanks to Jakob's help, this now causes no new test suite failures!
Over the entire test suite, this gives an average 1% speedup. The largest speedups are:
SingleSource/Benchmarks/Misc/pi - 108%
SingleSource/Benchmarks/CoyoteBench/lpbench - 54%
MultiSource/Benchmarks/Prolangs-C/unix-smail/unix-smail - 50%
SingleSource/Benchmarks/Shootout/ary3 - 32%
SingleSource/Benchmarks/Shootout-C++/matrix - 30%
The largest slowdowns are:
MultiSource/Benchmarks/mediabench/gsm/toast/toast - -30%
MultiSource/Benchmarks/Prolangs-C/bison/mybison - -25%
MultiSource/Benchmarks/BitBench/uuencode/uuencode - -22%
MultiSource/Applications/d/make_dparser - -14%
SingleSource/Benchmarks/Shootout-C++/ary - -13%
In light of these slowdowns, additional profiling work is obviously needed!
llvm-svn: 158223
The pass itself works well, but the something in the Machine* infrastructure
does not understand terminators which define registers. Without the ability
to use the block-placement pass, etc. this causes performance regressions (and
so is turned off by default). Turning off the analysis turns off the problems
with the Machine* infrastructure.
llvm-svn: 158206
The code which tests for an induction operation cannot assume that any
ADDI instruction will have a register operand because the operand could
also be a frame index; for example:
%vreg16<def> = ADDI8 <fi#0>, 0; G8RC:%vreg16
llvm-svn: 158205
This pass is derived from the Hexagon HardwareLoops pass. The only significant enhancement over the Hexagon
pass is that PPCCTRLoops will also attempt to delete the replaced add and compare operations if they are
no longer otherwise used. Also, invalid preheader DebugLoc is not used.
llvm-svn: 158204
This patch will generate the following for integer ABS:
movl %edi, %eax
negl %eax
cmovll %edi, %eax
INSTEAD OF
movl %edi, %ecx
sarl $31, %ecx
leal (%rdi,%rcx), %eax
xorl %ecx, %eax
There exists a target-independent DAG combine for integer ABS, which converts
integer ABS to sar+add+xor. For X86, we match this pattern back to neg+cmov.
This is implemented in PerformXorCombine.
rdar://10695237
llvm-svn: 158175
This patch will optimize the following
movq %rdi, %rax
subq %rsi, %rax
cmovsq %rsi, %rdi
movq %rdi, %rax
to
cmpq %rsi, %rdi
cmovsq %rsi, %rdi
movq %rdi, %rax
Perform this optimization if the actual result of SUB is not used.
rdar: 11540023
llvm-svn: 158126
The commit is intended to fix rdar://11540023.
It is implemented as part of peephole optimization. We can actually implement
this in the SelectionDAG lowering phase.
llvm-svn: 158122
when a compile time constant is known. This occurs when implicitly zero
extending function arguments from 16 bits to 32 bits.
<rdar://problem/11481151>
llvm-svn: 157966
It seems that this no longer causes test suite failures on PPC64 (after r157159),
and often gives a performance benefit, so it can be enabled by default.
llvm-svn: 157911
This patch will optimize the following:
sub r1, r3
cmp r3, r1 or cmp r1, r3
bge L1
TO
sub r1, r3
bge L1 or ble L1
If the branch instruction can use flag from "sub", then we can eliminate
the "cmp" instruction.
llvm-svn: 157831
This implements codegen support for accesses to thread-local variables
using the local-dynamic model, and adds a clean-up pass so that the base
address for the TLS block can be re-used between local-dynamic access on
an execution path.
llvm-svn: 157818
types, as well as int<->ptr casts. This allows us to tailcall functions
with some trivial casts between the call and return (i.e. because the
return types disagree).
llvm-svn: 157798
This patch will optimize the following
movq %rdi, %rax
subq %rsi, %rax
cmovsq %rsi, %rdi
movq %rdi, %rax
to
cmpq %rsi, %rdi
cmovsq %rsi, %rdi
movq %rdi, %rax
Perform this optimization if the actual result of SUB is not used.
rdar: 11540023
llvm-svn: 157755
I disabled FMA3 autodetection, since the result may differ from expected for some benchmarks.
I added tests for GodeGen and intrinsics.
I did not change llvm.fma.f32/64 - it may be done later.
llvm-svn: 157737
It helps compile exotic inline asm. In the test case, normal GR32
virtual registers use up eax-edx so the final GR32_ABCD live range has
no registers left. Since all the live ranges were tiny, we had no way of
prioritizing the smaller register class.
This patch allows tiny unspillable live ranges to be evicted by tiny
unspillable live ranges from a smaller register class.
<rdar://problem/11542429>
llvm-svn: 157715
integer registers. This is already supported by the fastcc convention, but it doesn't
hurt to support it in the standard conventions as well.
In cases where we can cheat at the calling convention, this allows us to avoid returning
things through memory in more cases.
llvm-svn: 157698
Besides adding the new insertPass function, this patch uses it to
enhance the existing -print-machineinstrs so that the MachineInstrs
after a specific pass can be printed.
Patch by Bin Zeng!
llvm-svn: 157655
This required light surgery on the assembler and disassembler
because the instructions use an uncommon encoding. They are
the only two instructions in x86 that use register operands
and two immediates.
llvm-svn: 157634
SimplifyCFG tends to form a lot of 2-3 case switches when merging branches. Move
the most likely condition to the front so it is checked first and the others can
be skipped. This is currently not as effective as it could be because SimplifyCFG
destroys profiling metadata when merging branches and switches. Merging branch
weight metadata is tricky though.
This code touches at most 3 cases so I didn't use a proper sorting algorithm.
llvm-svn: 157521
Live ranges with a constrained register class may benefit from splitting
around individual uses. It allows the remaining live range to use a
larger register class where it may allocate. This is like spilling to a
different register class.
This is only attempted on constrained register classes.
<rdar://problem/11438902>
llvm-svn: 157354
Now that the coalescer keeps live intervals and machine code in sync at
all times, it needs to deal with identity copies differently.
When merging two virtual registers, all identity copies are removed
right away. This means that other identity copies must come from
somewhere else, and they are going to have a value number.
Deal with such copies by merging the value numbers before erasing the
copy instruction. Otherwise, we leave dangling value numbers in the live
interval.
This fixes PR12927.
llvm-svn: 157340
may be RAUW'd by the recursive call to LegalizeOps; instead, retrieve
the other operands when calling UpdateNodeOperands. Fixes PR12889.
llvm-svn: 157162
Dead code elimination during coalescing could cause a virtual register
to be split into connected components. The following rewriting would be
confused about the already joined copies present in the code, but
without a corresponding value number in the live range.
Erase all joined copies instantly when joining intervals such that the
MI and LiveInterval representations are always in sync.
llvm-svn: 157135
The late dead code elimination is no longer necessary.
The test changes are cause by a register hint that can be either %rdi or
%rax. The choice depends on the use list order, which this patch changes.
llvm-svn: 157131
Use a dedicated MachO load command to annotate data-in-code regions.
This is the same format the linker produces for final executable images,
allowing consistency of representation and use of introspection tools
for both object and executable files.
Data-in-code regions are annotated via ".data_region"/".end_data_region"
directive pairs, with an optional region type.
data_region_directive := ".data_region" { region_type }
region_type := "jt8" | "jt16" | "jt32" | "jta32"
end_data_region_directive := ".end_data_region"
The previous handling of ARM-style "$d.*" labels was broken and has
been removed. Specifically, it didn't handle ARM vs. Thumb mode when
marking the end of the section.
rdar://11459456
llvm-svn: 157062
non-profitable commute using outdated info. The test case would still fail
because of poor pre-RA schedule. That will be fixed by MI scheduler.
rdar://11472010
llvm-svn: 157038
This is the same as the other tests: Clever tricks are required to make
the arguments and return value line up in a single-instruction function.
It rarely happens in real life.
We have plenty other examples of this behavior.
llvm-svn: 157030
This option has been disabled for a while, and it is going away so I can
clean up the coalescer code.
The tests that required physreg joining to be enabled were almost all of
the form "tiny function with interference between arguments and return
value". Such functions are usually inlined in the real world.
The problem exposed by phys_subreg_coalesce-3.ll is real, but fairly
rare.
llvm-svn: 157027
It is now possible to coalesce weird skewed sub-register copies by
picking a super-register class larger than both original registers. The
included test case produces code like this:
vld2.32 {d16, d17, d18, d19}, [r0]!
vst2.32 {d18, d19, d20, d21}, [r0]
We still perform interference checking as if it were a normal full copy
join, so this is still quite conservative. In particular, the f1 and f2
functions in the included test case still have remaining copies because
of false interference.
llvm-svn: 156878
- Remove code which lowers pseudo SETGP01.
- Fix LowerSETGP01. The first two of the three instructions that are emitted to
initialize the global pointer register now use register $2.
- Stop emitting .cpload directive.
llvm-svn: 156689
pointer register.
This is the first of the series of patches which clean up the way global pointer
register is used. The patches will make the following improvements:
- Make $gp an allocatable temporary register rather than reserving it.
- Use a virtual register as the global pointer register and let the register
allocator decide which register to assign to it or whether spill/reloads are
needed.
- Make sure $gp is valid at the entry of a called function, which is necessary
for functions using lazy binding.
- Remove the need for emitting .cprestore and .cpload directives.
llvm-svn: 156671
This patch will optimize the following cases:
sub r1, r3 | sub r1, imm
cmp r3, r1 or cmp r1, r3 | cmp r1, imm
bge L1
TO
subs r1, r3
bge L1 or ble L1
If the branch instruction can use flag from "sub", then we can replace
"sub" with "subs" and eliminate the "cmp" instruction.
rdar: 10734411
llvm-svn: 156599
This patch will optimize the following cases:
sub r1, r3 | sub r1, imm
cmp r3, r1 or cmp r1, r3 | cmp r1, imm
bge L1
TO
subs r1, r3
bge L1 or ble L1
If the branch instruction can use flag from "sub", then we can replace
"sub" with "subs" and eliminate the "cmp" instruction.
rdar: 10734411
llvm-svn: 156550
This patch will optimize -(x != 0) on X86
FROM
cmpl $0x01,%edi
sbbl %eax,%eax
notl %eax
TO
negl %edi
sbbl %eax %eax
In order to generate negl, I added patterns in Target/X86/X86InstrCompiler.td:
def : Pat<(X86sub_flag 0, GR32:$src), (NEG32r GR32:$src)>;
rdar: 10961709
llvm-svn: 156312
The primitive conservative heuristic seems to give a slight overall
improvement while not regressing stuff. Make it available to wider
testing. If you notice any speed regressions (or significant code
size regressions) let me know!
llvm-svn: 156258
This came up when a change in block placement formed a cmov and slowed down a
hot loop by 50%:
ucomisd (%rdi), %xmm0
cmovbel %edx, %esi
cmov is a really bad choice in this context because it doesn't get branch
prediction. If we emit it as a branch, an out-of-order CPU can do a better job
(if the branch is predicted right) and avoid waiting for the slow load+compare
instruction to finish. Of course it won't help if the branch is unpredictable,
but those are really rare in practice.
This patch uses a dumb conservative heuristic, it turns all cmovs that have one
use and a direct memory operand into branches. cmovs usually save some code
size, so we disable the transform in -Os mode. In-Order architectures are
unlikely to benefit as well, those are included in the
"predictableSelectIsExpensive" flag.
It would be better to reuse branch probability info here, but BPI doesn't
support select instructions currently. It would make sense to use the same
heuristics as the if-converter pass, which does the opposite direction of this
transform.
Test suite shows a small improvement here and there on corei7-level machines,
but the actual results depend a lot on the used microarchitecture. The
transformation is currently disabled by default and available by passing the
-enable-cgp-select2branch flag to the code generator.
Thanks to Chandler for the initial test case to him and Evan Cheng for providing
me with comments and test-suite numbers that were more stable than mine :)
llvm-svn: 156234
The new target machines are:
nvptx (old ptx32) => 32-bit PTX
nvptx64 (old ptx64) => 64-bit PTX
The sources are based on the internal NVIDIA NVPTX back-end, and
contain more functionality than the current PTX back-end currently
provides.
NV_CONTRIB
llvm-svn: 156196
to catch cases like:
%reg1024<def> = MOV r1
%reg1025<def> = MOV r0
%reg1026<def> = ADD %reg1024, %reg1025
r0 = MOV %reg1026
By commuting ADD, it let coalescer eliminate all of the copies. However, there
was a bug in the heuristics where it ended up commuting the ADD in:
%reg1024<def> = MOV r0
%reg1025<def> = MOV 0
%reg1026<def> = ADD %reg1024, %reg1025
r0 = MOV %reg1026
That did no benefit but rather ensure the last MOV would not be coalesced.
rdar://11355268
llvm-svn: 156048
This patch will optimize the following cases on X86
(a > b) ? (a-b) : 0
(a >= b) ? (a-b) : 0
(b < a) ? (a-b) : 0
(b <= a) ? (a-b) : 0
FROM
movl %edi, %ecx
subl %esi, %ecx
cmpl %edi, %esi
movl $0, %eax
cmovll %ecx, %eax
TO
xorl %eax, %eax
subl %esi, %edi
cmovll %eax, %edi
movl %edi, %eax
rdar: 10734411
llvm-svn: 155919
On x86-32, structure return via sret lets the callee pop the hidden
pointer argument off the stack, which the caller then re-pushes.
However if the calling convention is fastcc, then a register is used
instead, and the caller should not adjust the stack. This is
implemented with a check of IsTailCallConvention
X86TargetLowering::LowerCall but is now checked properly in
X86FastISel::DoSelectCall.
(this time, actually commit what was reviewed!)
llvm-svn: 155825
ARM BUILD_VECTORs created after type legalization cannot use i8 or i16
operands, since those types are not legal. Instead use i32 operands, which
will be implicitly truncated by the BUILD_VECTOR to match the element type.
llvm-svn: 155824
This time, also fix the caller of AddGlue to properly handle
incomplete chains. AddGlue had failure modes, but shamefully hid them
from its caller. It's luck ran out.
Fixes rdar://11314175: BuildSchedUnits assert.
llvm-svn: 155749
On x86-32, structure return via sret lets the callee pop the hidden
pointer argument off the stack, which the caller then re-pushes.
However if the calling convention is fastcc, then a register is used
instead, and the caller should not adjust the stack. This is
implemented with a check of IsTailCallConvention
X86TargetLowering::LowerCall but is now checked properly in
X86FastISel::DoSelectCall.
llvm-svn: 155745
x == -y --> x+y == 0
x != -y --> x+y != 0
On x86, the generated code goes from
negl %esi
cmpl %esi, %edi
je .LBB0_2
to
addl %esi, %edi
je .L4
This case is correctly handled for ARM with "cmn".
Patch by Manman Ren.
rdar://11245199
PR12545
llvm-svn: 155739
* Model FPSW (the FPU status word) as a register.
* Add ISel patterns for the FUCOM*, FNSTSW and SAHF instructions.
* During Legalize/Lowering, build a node sequence to transfer the comparison
result from FPSW into EFLAGS. If you're wondering about the right-shift: That's
an implicit sub-register extraction (%ax -> %ah) which is handled later on by
the instruction selector.
Fixes PR6679. Patch by Christoph Erhardt!
llvm-svn: 155704
DAGCombine strangeness may result in multiple loads from the same
offset. They both may try to glue themselves to another load. We could
insist that the redundant loads glue themselves to each other, but the
beter fix is to bail out from bad gluing at the time we detect it.
Fixes rdar://11314175: BuildSchedUnits assert.
llvm-svn: 155668
On some cores it's a bad idea for performance to mix VFP and NEON instructions
and since these patterns are NEON anyway, the NEON load should be used.
llvm-svn: 155630
the feature set of v7a. This comes about if the user specifies something like
-arch armv7 -mcpu=cortex-m3. We shouldn't be generating instructions such as
uxtab in this case.
rdar://11318438
llvm-svn: 155601
using the pattern (vbroadcast (i32load src)). In some cases, after we generate
this pattern new users are added to the load node, which prevent the selection
of the blend pattern. This commit provides fallback patterns which perform
in-vector broadcast (using in-vector vbroadcast in AVX2 and pshufd on AVX1).
llvm-svn: 155437
on X86 Atom. Some of our tests failed because the tail merging part of
the BranchFolding pass was creating new basic blocks which did not
contain live-in information. When the anti-dependency code in the Post-RA
scheduler ran, it would sometimes rename the register containing
the function return value because the fact that the return value was
live-in to the subsequent block had been lost. To fix this, it is necessary
to run the RegisterScavenging code in the BranchFolding pass.
This patch makes sure that the register scavenging code is invoked
in the X86 subtarget only when post-RA scheduling is being done.
Post RA scheduling in the X86 subtarget is only done for Atom.
This patch adds a new function to the TargetRegisterClass to control
whether or not live-ins should be preserved during branch folding.
This is necessary in order for the anti-dependency optimizations done
during the PostRASchedulerList pass to work properly when doing
Post-RA scheduling for the X86 in general and for the Intel Atom in particular.
The patch adds and invokes the new function trackLivenessAfterRegAlloc()
instead of using the existing requiresRegisterScavenging().
It changes BranchFolding.cpp to call trackLivenessAfterRegAlloc() instead of
requiresRegisterScavenging(). It changes the all the targets that
implemented requiresRegisterScavenging() to also implement
trackLivenessAfterRegAlloc().
It adds an assertion in the Post RA scheduler to make sure that post RA
liveness information is available when it is needed.
It changes the X86 break-anti-dependencies test to use –mcpu=atom, in order
to avoid running into the added assertion.
Finally, this patch restores the use of anti-dependency checking
(which was turned off temporarily for the 3.1 release) for
Intel Atom in the Post RA scheduler.
Patch by Andy Zhang!
Thanks to Jakob and Anton for their reviews.
llvm-svn: 155395
test suite failures. The failures occur at each stage, and only get
worse, so I'm reverting all of them.
Please resubmit these patches, one at a time, after verifying that the
regression test suite passes. Never submit a patch without running the
regression test suite.
llvm-svn: 155372
The X86 target is editing the selection DAG while isel is selecting
nodes following a topological ordering. When the DAG hacking triggers
CSE, nodes can be deleted and bad things happen.
llvm-svn: 155257
when the set bits aren't the same for both args of the xor.
This transformation is in the function TargetLowering::SimplifyDemandedBits
in the file lib/CodeGen/SelectionDAG/TargetLowering.cpp.
I have tested this test using a previous version of llc which the defect and
the a version of llc which does not. I got the expected fail and pass,
respectively.
This test goes with rdar://11195364 and the check in with the fix: svn r154955
llvm-svn: 155156
also fix SimplifyLibCalls to use TLI rather than compile-time conditionals to enable optimizations on floor, ceil, round, rint, and nearbyint
llvm-svn: 154960
both fallthrough and a conditional branch target the same successor.
Gracefully delete the conditional branch and introduce any unconditional
branch needed to reach the actual successor. This fixes memory
corruption in 2009-06-15-RegScavengerAssert.ll and possibly other tests.
Also, while I'm here fix a latent bug I spotted by inspection. I never
applied the same fundamental fix to this fallthrough successor finding
logic that I did to the logic used when there are no conditional
branches. As a consequence it would have selected landing pads had they
be aligned in just the right way here. I don't have a test case as
I spotted this by inspection, and the previous time I found this
required have of TableGen's source code to produce it. =/ I hate backend
bugs. ;]
Thanks to Jim Grosbach for helping me reason through this and reviewing
the fix.
llvm-svn: 154867
This is mostly to test the waters. I'd like to get results from FNT
build bots and other bots running on non-x86 platforms.
This feature has been pretty heavily tested over the last few months by
me, and it fixes several of the execution time regressions caused by the
inlining work by preventing inlining decisions from radically impacting
block layout.
I've seen very large improvements in yacr2 and ackermann benchmarks,
along with the expected noise across all of the benchmark suite whenever
code layout changes. I've analyzed all of the regressions and fixed
them, or found them to be impossible to fix. See my email to llvmdev for
more details.
I'd like for this to be in 3.1 as it complements the inliner changes,
but if any failures are showing up or anyone has concerns, it is just
a flag flip and so can be easily turned off.
I'm switching it on tonight to try and get at least one run through
various folks' performance suites in case SPEC or something else has
serious issues with it. I'll watch bots and revert if anything shows up.
llvm-svn: 154816
once we start changing the block layout, so just nuke it. If anyone has
ideas about how to craft a code layout agnostic form of the test please
let me know.
llvm-svn: 154815
rotation. When there is a loop backedge which is an unconditional
branch, we will end up with a branch somewhere no matter what. Try
placing this backedge in a fallthrough position above the loop header as
that will definitely remove at least one branch from the loop iteration,
where whole loop rotation may not.
I haven't seen any benchmarks where this is important but loop-blocks.ll
tests for it, and so this will be covered when I flip the default.
llvm-svn: 154812
laid out in a form with a fallthrough into the header and a fallthrough
out of the bottom. In that case, leave the loop alone because any
rotation will introduce unnecessary branches. If either side looks like
it will require an explicit branch, then the rotation won't add any, do
it to ensure the branch occurs outside of the loop (if possible) and
maximize the benefit of the fallthrough in the bottom.
llvm-svn: 154806
This is a complex change that resulted from a great deal of
experimentation with several different benchmarks. The one which proved
the most useful is included as a test case, but I don't know that it
captures all of the relevant changes, as I didn't have specific
regression tests for each, they were more the result of reasoning about
what the old algorithm would possibly do wrong. I'm also failing at the
moment to craft more targeted regression tests for these changes, if
anyone has ideas, it would be welcome.
The first big thing broken with the old algorithm is the idea that we
can take a basic block which has a loop-exiting successor and a looping
successor and use the looping successor as the layout top in order to
get that particular block to be the bottom of the loop after layout.
This happens to work in many cases, but not in all.
The second big thing broken was that we didn't try to select the exit
which fell into the nearest enclosing loop (to which we exit at all). As
a consequence, even if the rotation worked perfectly, it would result in
one of two bad layouts. Either the bottom of the loop would get
fallthrough, skipping across a nearer enclosing loop and thereby making
it discontiguous, or it would be forced to take an explicit jump over
the nearest enclosing loop to earch its successor. The point of the
rotation is to get fallthrough, so we need it to fallthrough to the
nearest loop it can.
The fix to the first issue is to actually layout the loop from the loop
header, and then rotate the loop such that the correct exiting edge can
be a fallthrough edge. This is actually much easier than I anticipated
because we can handle all the hard parts of finding a viable rotation
before we do the layout. We just store that, and then rotate after
layout is finished. No inner loops get split across the post-rotation
backedge because we check for them when selecting the rotation.
That fix exposed a latent problem with our exitting block selection --
we should allow the backedge to point into the middle of some inner-loop
chain as there is no real penalty to it, the whole point is that it
*won't* be a fallthrough edge. This may have blocked the rotation at all
in some cases, I have no idea and no test case as I've never seen it in
practice, it was just noticed by inspection.
Finally, all of these fixes, and studying the loops they produce,
highlighted another problem: in rotating loops like this, we sometimes
fail to align the destination of these backwards jumping edges. Fix this
by actually walking the backwards edges rather than relying on loopinfo.
This fixes regressions on heapsort if block placement is enabled as well
as lots of other cases where the previous logic would introduce an
abundance of unnecessary branches into the execution.
llvm-svn: 154783
There is an assert at line 558 in ScheduleDAGInstrs::buildSchedGraph(AliasAnalysis *AA).
This assert needs to addressed for post RA scheduler. Until that assert is addressed,
any passes that uses post ra scheduler will fail. So, I am temporarily disabling the
hexagon tests until that fix is in.
The assert is as follows:
assert(!MI->isTerminator() && !MI->isLabel() &&
"Cannot schedule terminators or labels!");
llvm-svn: 154617
- FCOPYSIGN nodes that have operands of different types were not handled.
- Different code was generated depending on the endianness of the target.
Additionally, code is added that emits INS and EXT instructions, if they are
supported by target (they are R2 instructions).
llvm-svn: 154540
Original message:
Modify the code that lowers shuffles to blends from using blendvXX to vblendXX.
blendV uses a register for the selection while Vblend uses an immediate.
On sandybridge they still have the same latency and execute on the same execution ports.
llvm-svn: 154483
blendv uses a register for the selection while vblend uses an immediate.
On sandybridge they still have the same latency and execute on the same execution ports.
llvm-svn: 154396
legalizer always use the DAG entry node. This is wrong when the libcall is
emitted as a tail call since it effectively folds the return node. If
the return node's input chain is not the entry (i.e. call, load, or store)
use that as the tail call input chain.
PR12419
rdar://9770785
rdar://11195178
llvm-svn: 154370
in-register, such that we can use a single vector store rather then a
series of scalar stores.
For func_4_8 the generated code
vldr d16, LCPI0_0
vmov d17, r0, r1
vadd.i16 d16, d17, d16
vmov.u16 r0, d16[3]
strb r0, [r2, #3]
vmov.u16 r0, d16[2]
strb r0, [r2, #2]
vmov.u16 r0, d16[1]
strb r0, [r2, #1]
vmov.u16 r0, d16[0]
strb r0, [r2]
bx lr
becomes
vldr d16, LCPI0_0
vmov d17, r0, r1
vadd.i16 d16, d17, d16
vuzp.8 d16, d17
vst1.32 {d16[0]}, [r2, :32]
bx lr
I'm not fond of how this combine pessimizes 2012-03-13-DAGCombineBug.ll,
but I couldn't think of a way to judiciously apply this combine.
This
ldrh r0, [r0, #4]
strh r0, [r1]
becomes
vldr d16, [r0]
vmov.u16 r0, d16[2]
vmov.32 d16[0], r0
vuzp.16 d16, d17
vst1.32 {d16[0]}, [r1, :32]
PR11158
rdar://10703339
llvm-svn: 154340
Akira Hatanaka