Lay out trellis-shaped CFGs optimally.
A trellis of the shape below:
A B
|\ /|
| \ / |
| X |
| / \ |
|/ \|
C D
would be laid out A; B->C ; D by the current layout algorithm. Now we identify
trellises and lay them out either A->C; B->D or A->D; B->C. This scales with an
increasing number of predecessors. A trellis is a a group of 2 or more
predecessor blocks that all have the same successors.
because of this we can tail duplicate to extend existing trellises.
As an example consider the following CFG:
B D F H
/ \ / \ / \ / \
A---C---E---G---Ret
Where A,C,E,G are all small (Currently 2 instructions).
The CFG preserving layout is then A,B,C,D,E,F,G,H,Ret.
The current code will copy C into B, E into D and G into F and yield the layout
A,C,B(C),E,D(E),F(G),G,H,ret
define void @straight_test(i32 %tag) {
entry:
br label %test1
test1: ; A
%tagbit1 = and i32 %tag, 1
%tagbit1eq0 = icmp eq i32 %tagbit1, 0
br i1 %tagbit1eq0, label %test2, label %optional1
optional1: ; B
call void @a()
br label %test2
test2: ; C
%tagbit2 = and i32 %tag, 2
%tagbit2eq0 = icmp eq i32 %tagbit2, 0
br i1 %tagbit2eq0, label %test3, label %optional2
optional2: ; D
call void @b()
br label %test3
test3: ; E
%tagbit3 = and i32 %tag, 4
%tagbit3eq0 = icmp eq i32 %tagbit3, 0
br i1 %tagbit3eq0, label %test4, label %optional3
optional3: ; F
call void @c()
br label %test4
test4: ; G
%tagbit4 = and i32 %tag, 8
%tagbit4eq0 = icmp eq i32 %tagbit4, 0
br i1 %tagbit4eq0, label %exit, label %optional4
optional4: ; H
call void @d()
br label %exit
exit:
ret void
}
here is the layout after D27742:
straight_test: # @straight_test
; ... Prologue elided
; BB#0: # %entry ; A (merged with test1)
; ... More prologue elided
mr 30, 3
andi. 3, 30, 1
bc 12, 1, .LBB0_2
; BB#1: # %test2 ; C
rlwinm. 3, 30, 0, 30, 30
beq 0, .LBB0_3
b .LBB0_4
.LBB0_2: # %optional1 ; B (copy of C)
bl a
nop
rlwinm. 3, 30, 0, 30, 30
bne 0, .LBB0_4
.LBB0_3: # %test3 ; E
rlwinm. 3, 30, 0, 29, 29
beq 0, .LBB0_5
b .LBB0_6
.LBB0_4: # %optional2 ; D (copy of E)
bl b
nop
rlwinm. 3, 30, 0, 29, 29
bne 0, .LBB0_6
.LBB0_5: # %test4 ; G
rlwinm. 3, 30, 0, 28, 28
beq 0, .LBB0_8
b .LBB0_7
.LBB0_6: # %optional3 ; F (copy of G)
bl c
nop
rlwinm. 3, 30, 0, 28, 28
beq 0, .LBB0_8
.LBB0_7: # %optional4 ; H
bl d
nop
.LBB0_8: # %exit ; Ret
ld 30, 96(1) # 8-byte Folded Reload
addi 1, 1, 112
ld 0, 16(1)
mtlr 0
blr
The tail-duplication has produced some benefit, but it has also produced a
trellis which is not laid out optimally. With this patch, we improve the layouts
of such trellises, and decrease the cost calculation for tail-duplication
accordingly.
This patch produces the layout A,C,E,G,B,D,F,H,Ret. This layout does have
back edges, which is a negative, but it has a bigger compensating
positive, which is that it handles the case where there are long strings
of skipped blocks much better than the original layout. Both layouts
handle runs of executed blocks equally well. Branch prediction also
improves if there is any correlation between subsequent optional blocks.
Here is the resulting concrete layout:
straight_test: # @straight_test
; BB#0: # %entry ; A (merged with test1)
mr 30, 3
andi. 3, 30, 1
bc 12, 1, .LBB0_4
; BB#1: # %test2 ; C
rlwinm. 3, 30, 0, 30, 30
bne 0, .LBB0_5
.LBB0_2: # %test3 ; E
rlwinm. 3, 30, 0, 29, 29
bne 0, .LBB0_6
.LBB0_3: # %test4 ; G
rlwinm. 3, 30, 0, 28, 28
bne 0, .LBB0_7
b .LBB0_8
.LBB0_4: # %optional1 ; B (Copy of C)
bl a
nop
rlwinm. 3, 30, 0, 30, 30
beq 0, .LBB0_2
.LBB0_5: # %optional2 ; D (Copy of E)
bl b
nop
rlwinm. 3, 30, 0, 29, 29
beq 0, .LBB0_3
.LBB0_6: # %optional3 ; F (Copy of G)
bl c
nop
rlwinm. 3, 30, 0, 28, 28
beq 0, .LBB0_8
.LBB0_7: # %optional4 ; H
bl d
nop
.LBB0_8: # %exit
Differential Revision: https://reviews.llvm.org/D28522
llvm-svn: 295223
We currently can't legalize those, but we should really not be creating
them in the first place, since legalization would probably look similar to the
way we legalize CONCAT_VECTORS - basically replace the INSERT with a BUILD.
This fixes PR311956.
Differential Revision: https://reviews.llvm.org/D29961
llvm-svn: 295213
This patch reverts region's scheduling to the original untouched state
in case if we have have decreased occupancy.
In addition it switches to use TargetRegisterInfo occupancy callback
for pressure limits instead of gradually increasing limits which were
just passed by. We are going to stay with the best schedule so we do
not need to tolerate worsened scheduling anymore.
Differential Revision: https://reviews.llvm.org/D29971
llvm-svn: 295206
Summary:
We don't seem to have great rules on what a valid VBROADCAST node looks like. And as a consequence we end up with a lot of patterns to try to catch everything. We have patterns with scalar inputs, 128-bit vector inputs, 256-bit vector inputs, and 512-bit vector inputs.
As you can see from the things improved here we are currently missing patterns for 128-bit loads being extended to 256-bit before the vbroadcast.
I'd like to propose that VBROADCAST should always take a 128-bit vector type as input. As a first step towards that this patch adds an EXTRACT_SUBVECTOR in front of VBROADCAST when the input is 256 or 512-bits. In the future I would like to add scalar_to_vector around all the scalar operations. And maybe we should consider adding a VBROADCAST+load node to avoid separating loads from the broadcasting operation when the load itself isn't foldable.
This requires an additional change in target shuffle combining to look for the extract subvector and look through it to find the original operand. I'm sure this change isn't perfect but was enough to fix a few test failures that were being caused.
Another interesting thing I noticed is that the changes in masked_gather_scatter.ll show cases were we don't remove a useless insert into element 1 before broadcasting element 0.
Reviewers: delena, RKSimon, zvi
Reviewed By: zvi
Subscribers: igorb, llvm-commits
Differential Revision: https://reviews.llvm.org/D28747
llvm-svn: 295155
This patch corrects the maximum workgroups per CU if we have big
workgroups (more than 128). This calculation contributes to the
occupancy calculation in respect to LDS size.
Differential Revision: https://reviews.llvm.org/D29974
llvm-svn: 295134
Summary:
Blocks ending in unreachable are typically cold because they end the
program or throw an exception, so merging them with other identical
blocks is usually profitable because it reduces the size of cold code.
MachineBlockPlacement generally does not arrange to fall through to such
blocks, so commoning these blocks will not introduce additional
unconditional branches.
Reviewers: hans, iteratee, haicheng
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29153
llvm-svn: 295105
This instruction clears the low bits of a pointer without requiring (possibly
dodgy if pointers aren't ints) conversions to and from an integer. Since (as
far as I'm aware) all masks are statically known, the instruction takes an
immediate operand rather than a register to specify the mask.
llvm-svn: 295103
This adds MXCSR to the set of recognized registers for X86 targets and updates the instructions that read or write it. I do not intend for all of the various floating point instructions that implicitly use the control bits or update the status bits of this register to ever have that usage modeled by default. However, when constrained floating point modes (such as strict FP exception status modeling or dynamic rounding modes) are enabled, implicit use/def information for MXCSR will be added to those instructions.
Until those additional updates are made this should cause (almost?) no functional changes. Theoretically, this will prevent instructions like LDMXCSR and STMXCSR from being moved past one another, but that should be prevented anyway and I haven't found a case where it is happening now.
Differential Revision: https://reviews.llvm.org/D29903
llvm-svn: 295004
Backends don't support this yet. They would have to move to the swifterror
register before the tail call to make sure it is live-in to the call.
rdar://30495920
llvm-svn: 294982
This is consistent with what we do for GlobalISel. That way, it is easy
to see whether or not FastISel is able to fully select a function.
At some point we may want to switch that to an optimization remark.
llvm-svn: 294970
Summary:
The attached test case fails with "fatal error: error in backend:
misaligned pc-relative fixup value" as the jump table is misaligned.
The EmitAlignment existed already for ARM and Thumb-1 code, but was
missing for Thumb-2.
The test checks that the fatal error disappears when generating an obj
file, as well as checking the align directive is there when producing an
asm file.
Reviewers: rengolin, grosbach, t.p.northover, jmolloy, SjoerdMeijer, samparker
Reviewed By: samparker
Subscribers: samparker, aemerson, llvm-commits
Differential Revision: https://reviews.llvm.org/D29650
llvm-svn: 294950
We match a sequence of 3-4 instructions into a tTBB pseudo. One of our checks is that
a particular register in that sequence is killed (so it can be clobbered by the pseudo).
We weren't noticing if an errant MOV or other instruction had infiltrated the
sequence we were walking. If it had, and it defined the register we've already
identified as killed, it makes it live across the tBR_JT and thus unclobberable.
Notice this case and bail out.
llvm-svn: 294949
When generating a floating point comparison we currently unconditionally
generate VCMPE. This has the sideeffect of setting the cumulative Invalid
bit in FPSCR if any of the operands are QNaN.
It is expected that use of a relational predicate on a QNaN value should
raise Invalid. Quoting from the C standard:
The relational and equality operators support the usual mathematical
relationships between numeric values. For any ordered pair of numeric
values exactly one of relationships the less, greater, equal and is true.
Relational operators may raise the floating-point exception when argument
values are NaNs.
The standard doesn't explicitly state the expectation for equality operators,
but the implication and obvious expectation is that equality operators
should not raise Invalid on a QNaN input, as those predicates are wholly
defined on unordered inputs (to return not equal).
Therefore, add a new operand to ARMISD::FPCMP and FPCMPZ indicating if
QNaN should raise Invalid, and pipe that through to TableGen.
llvm-svn: 294945
We now detect that both the extract and insert indices are non-zero and convert to a shuffle. This will be lowered as a blend for 256-bit vectors or as a vshuf operations for 512-bit vectors.
llvm-svn: 294931
The bug was introduced with:
https://reviews.llvm.org/rL294863
...and manifests as a selection failure in x86, but that's actually
another bug. This fix prevents wrong codegen with -0.0, but in the
more common case when we have NSZ and NNAN (-ffast-math), we should
still be able to fold this setcc/compare.
llvm-svn: 294924
Initial 256-bit vector support - 512-bit support requires extra checks for AVX512BW support (PMOVZXBW) that will be handled in a future patch.
llvm-svn: 294896
I don't know if anything other than x86 vectors is affected by this change, but this may allow
us to remove target-specific intrinsics for blendv* (vector selects). The simplification arises
from the fact that blendv* instructions only use the sign-bit when deciding which vector element
to choose for the destination vector. The mechanism to fold VSELECT into SHRUNKBLEND nodes already
exists in x86 lowering; this demanded bits change just enables the transform to fire more often.
The original motivation starts with a bug for DSE of masked stores that seems completely unrelated,
but I've explained the likely steps in this series here:
https://llvm.org/bugs/show_bug.cgi?id=11210
Differential Revision: https://reviews.llvm.org/D29687
llvm-svn: 294863
Seems the execution dependency pass likes to use FP instructions when most of the consuming code is integer if a vextractf128 instruction produced the register. Without AVX2 we don't have the corresponding integer instruction available.
This patch suppresses the domain on these instructions to GenericDomain if AVX2 is not supported so that they are ignored by domain fixing. If AVX2 is supported we'll report the correct domain and allow them to switch between integer and fp.
Overall I think this produces better results in the modified test cases.
llvm-svn: 294824
The recommit includes some changes of testcases. No functional change to the patch.
In RateRegister of existing LSR, if a formula contains a Reg which is a SCEVAddRecExpr,
and this SCEVAddRecExpr's loop is an outerloop, the formula will be marked as Loser
and dropped.
Suppose we have an IR that %for.body is outerloop and %for.body2 is innerloop. LSR only
handle inner loop now so only %for.body2 will be handled.
Using the logic above, formula like
reg(%array) + reg({1,+, %size}<%for.body>) + 1*reg({0,+,1}<%for.body2>) will be dropped
no matter what because reg({1,+, %size}<%for.body>) is a SCEVAddRecExpr type reg related
with outerloop. Only formula like
reg(%array) + 1*reg({{1,+, %size}<%for.body>,+,1}<nuw><nsw><%for.body2>) will be kept
because the SCEVAddRecExpr related with outerloop is folded into the initial value of the
SCEVAddRecExpr related with current loop.
But in some cases, we do need to share the basic induction variable
reg{0 ,+, 1}<%for.body2> among LSR Uses to reduce the final total number of induction
variables used by LSR, so we don't want to drop the formula like
reg(%array) + reg({1,+, %size}<%for.body>) + 1*reg({0,+,1}<%for.body2>) unconditionally.
From the existing comment, it tries to avoid considering multiple level loops at the same time.
However, existing LSR only handles innermost loop, so for any SCEVAddRecExpr with a loop other
than current loop, it is an invariant and will be simple to handle, and the formula doesn't have
to be dropped.
Differential Revision: https://reviews.llvm.org/D26429
llvm-svn: 294814
Since r274013, we've been looking through bitcasts on broadcast inputs.
In the scalar-folding case (from a load, build_vector, or sc2vec),
the input type didn't matter, as we'd simply bitcast the resulting
scalar back.
However, when broadcasting a 128-bit-lane-aligned element, we create an
EXTRACT_SUBVECTOR. Use proper types, by creating an extract_subvector
of the original input type.
llvm-svn: 294774
In the encoding of system registers in the M-class MSR instruction the mask bits
should be 2 for registers that don't take a _<bits> qualifier (the instruction
is unpredictable otherwise), and should also be 2 if the register takes a
_<bits> qualifier but it's not present as no _<bits> is an alias for _nzcvq.
Differential Revision: https://reviews.llvm.org/D29828
llvm-svn: 294762
This makes sure we get the same redefinition rules regardless of who
is printing (asm parser, codegen) and to what (asm, obj).
This fixes an unintentional regression in r293936.
llvm-svn: 294752
The patch comes in 2 parts:
1 - it makes use of the SelectionDAG::NewNodesMustHaveLegalTypes flag to tell when it can safely constant fold illegal types.
2 - it correctly resets SelectionDAG::NewNodesMustHaveLegalTypes at the start of each call to SelectionDAGISel::CodeGenAndEmitDAG so all the pre-legalization stages can make use of it - not just the first basic block that gets handled.
Fix for PR30760
Differential Revision: https://reviews.llvm.org/D29568
llvm-svn: 294749
In some cases we call getTargetConstantBitsFromNode for nodes that haven't been lowered from BUILD_VECTOR yet
Note: We're getting very close to being able to move most of the constant extraction code from getTargetShuffleMaskIndices into getTargetConstantBitsFromNode
llvm-svn: 294746
This change returns empty PSet list for M0 register. Otherwise its
PSet as defined by tablegen is SReg_32. This results in incorrect
register pressure calculation every time an instruction uses M0.
Such uses count as SReg_32 PSet and inadequately increase pressure
on SGPRs.
Differential Revision: https://reviews.llvm.org/D29798
llvm-svn: 294691
Gcc supports target armv7ve which is armv7-a with virtualization
extensions. This change adds support for this in llvm for gcc
compatibility.
Also remove redundant FeatureHWDiv, FeatureHWDivARM for a few models as
this is specified automatically by FeatureVirtualization.
Patch by Manoj Gupta.
Differential Revision: https://reviews.llvm.org/D29472
llvm-svn: 294661
This requires that we communicate to X86InstrInfo::optimizeCompareInstr
that the second operand is neither a register nor an immediate. The way we
do that is by setting CmpMask to zero.
Note that there were already instructions where the second operand was not a
register nor an immediate, namely X86::SUB*rm, so also set CmpMask to zero
for those instructions. This seems like a latent bug, but I was unable to
trigger it.
Differential Revision: https://reviews.llvm.org/D28621
llvm-svn: 294634
Summary:
Fix two bugs in SelectionDAGBuilder::FindMergedConditions reported by
Mikael Holmen. Handle non-canonicalized xor not operation
correctly (was assuming operand 0 was always the non-constant operand)
and check that the negated condition is also in the same block as the
original and/or instruction (as is done for and/or operands already)
before proceeding with optimization.
Reviewers: bogner, MatzeB, qcolombet
Subscribers: mcrosier, uabelho, llvm-commits
Differential Revision: https://reviews.llvm.org/D29680
llvm-svn: 294605
If some of the trailing or leading bytes of a load combine pattern are zeroes we can combine the pattern to a load + zext and shift. Currently we don't support it, so the tests check the current codegen without load combine. This change will make the patch to support this kind of combine a bit more clear.
llvm-svn: 294591
Stack Smash Protection is not completely free, so in hot code, the overhead it causes can cause performance issues. By adding diagnostic information for which function have SSP and why, a user can quickly determine what they can do to stop SSP being applied to a specific hot function.
This change adds an SSP-specific DiagnosticInfo class and uses of it to the Stack Protection code. A subsequent change to clang will cause the remarks to be emitted when enabled.
Patch by: James Henderson
Differential Revision: https://reviews.llvm.org/D29023
llvm-svn: 294590
In combineOrCmpEqZeroToCtlzSrl, replace "getConstantOperand == 0" by "isNullConstant" to account for floating point constants.
Differential Revision: https://reviews.llvm.org/D29756
llvm-svn: 294588
LowerBuildVectorv16i8/LowerBuildVectorv8i16 insert values into a UNDEF vector if the build vector doesn't contain any zero elements, resulting in register dependencies with a previous use of the register.
This patch attempts to break the register dependency by either always zeroing the vector before hand or (if we're inserting to the 0'th element) by using VZEXT_MOVL(SCALAR_TO_VECTOR(i32 AEXT(Elt))) which lowers to (V)MOVD and performs a similar function. Additionally (V)MOVD is a shorter instruction than PINSRB/PINSRW. We already do something similar for SSE41 PINSRD.
On pre-SSE41 LowerBuildVectorv16i8 we go a little further and use VZEXT_MOVL(SCALAR_TO_VECTOR(i32 ZEXT(Elt))) if the build vector contains zeros to avoid the vector zeroing at the cost of a scalar zero extension, which can probably be brought over to the other cases in a future patch in some cases (load folding etc.)
Differential Revision: https://reviews.llvm.org/D29720
llvm-svn: 294581
This patch does the following.
1. Adds an Intrinsic int_x86_clzero which works with __builtin_ia32_clzero
2. Identifies clzero feature using cpuid info. (Function:8000_0008, Checks if EBX[0]=1)
3. Adds the clzero feature under znver1 architecture.
4. The custom inserter is added in Lowering.
5. A testcase is added to check the intrinsic.
6. The clzero instruction is added to assembler test.
Patch by Ganesh Gopalasubramanian with a couple formatting tweaks, a disassembler test, and using update_llc_test.py from me.
Differential revision: https://reviews.llvm.org/D29385
llvm-svn: 294558
Functions that have a dynamic alloca require a base register which is defined to
be X19 on AArch64 and r6 on ARM. We have defined the swifterror register to be
the same register. Use a different callee save register for swifterror instead:
X21 on AArch64
R8 on ARM
rdar://30433803
llvm-svn: 294551
It'll usually be immediately legalized back to a libcall, but occasionally
something can be done with it so we'd just as well enable that flexibility from
the start.
llvm-svn: 294530
We mark X0 as preserved by a call that passes the returned parameter.
x0 = ...
fun(x0) // no implicit def of x0
This no longer is valid if we pass the parameter in a different register then
the returned value as is the case with a swiftself parameter (passed in x20).
x20 = ...
fun(x20) // there should be an implict def of x8
rdar://30425845
llvm-svn: 294527
AArch64 has specific instructions to multiply two numbers at double the width
and produce the high part of the result. These can be used to implement LLVM's
mul.with.overflow instructions fairly simply. Helps with C++ operator new[].
llvm-svn: 294519
The AAPCS ABI is substantially more complicated so that's coming in a separate
patch. For now we can generate correct code for iOS though.
llvm-svn: 294493
I forgot to remove the neonfp target feature from the test, which means we'd
have trouble selecting VADDS on targets that have neonfp enabled by default.
llvm-svn: 294451
Add a register bank for floating point values and select simple instructions
using them (add, copies from GPR).
This assumes that the hardware can cope with a single precision add (VADDS)
instruction, so the legalizer will treat G_FADD as legal and the instruction
selector will refuse to select if the hardware doesn't support it. In the future
we'll want to be more careful about this, and legalize to libcalls if we have to
use soft float.
llvm-svn: 294442
This patch checks the number of operands in the resulting
instruction instead of just the alias, then skips over
tied operands when generating the printing method.
This allows us to generate the preferred assembly syntax
for the AArch64 'ins' instruction, which should always be
displayed as 'mov' according to the ARMARM.
Several unit tests have changed as a result, but only to
reflect the preferred disassembly.
Some other InstAlias patterns (movk/bic/orr) needed a
slight adjustment to stop them becoming the default
and breaking other unit tests.
Patch by Graham Hunter.
Differential Revision: https://reviews.llvm.org/D29219
llvm-svn: 294437
There are about 3 underlying bugs causing the tests to fail.
On top of that, some tests just we're 'generic' enough. i.e. 32-bit
registers.
llvm-svn: 294434
They are currently modelled incorrectly (as calls, which clobber
registers, confusing e.g. Machine Copy Propagation).
Reverting until we figure out the proper solution.
llvm-svn: 294348
Summary:
This change allows usage of store instruction for implicit null check.
Memory Aliasing Analisys is not used and change conservatively supposes
that any store and load may access the same memory. As a result
re-ordering of store-store, store-load and load-store is prohibited.
Patch by Serguei Katkov!
Reviewers: reames, sanjoy
Reviewed By: sanjoy
Subscribers: atrick, llvm-commits
Differential Revision: https://reviews.llvm.org/D29400
llvm-svn: 294338
Adds the vnot extended mnemonic for the vnor instruction.
Committing on behalf of brunoalr (Bruno Rosa).
Differential Revision: https://reviews.llvm.org/D29225
llvm-svn: 294330
Currently we don't support these nodes, so the tests check the current codegen without load combine. This change makes the review of the change to support these nodes more clear.
Separated from https://reviews.llvm.org/D29591 review.
llvm-svn: 294305
When constructing global address literals while targeting the RWPI
relocation model. LLVM currently only uses literal pools. If MOVW/MOVT
instructions are available we can use these instead. Beside being more
efficient it allows -arm-execute-only to work with
-relocation-model=RWPI as well.
When we generate MOVW/MOVT for global addresses when targeting the RWPI
relocation model, we need to use base relative relocations. This patch
does the needed plumbing in MC to generate these for MOVW/MOVT.
Differential Revision: https://reviews.llvm.org/D29487
Change-Id: I446786e43a6f5aa9b6a5bb2cd216d60d41c7755d
llvm-svn: 294298
This reverts commit r294186.
On an internal test, this triggers an out-of-memory error on PPC,
presumably because there is another dagcombine that does the exact
opposite triggering and endless loop consuming more and more memory.
Chandler has started at creating a reduced test case and we'll attach it
as soon as possible.
llvm-svn: 294288
This includes unmasked forms of variable shift and shifting by the lower element of a register.
Still need to do shift by immediate which was not foldable prior to avx512 and all the masked forms.
llvm-svn: 294285
joinReservedPhysReg() can only deal with a liverange in a single basic
block when copying from a vreg into a physreg.
See also rdar://30306405
Differential Revision: https://reviews.llvm.org/D29436
llvm-svn: 294268
For amdgcn target Clang generates addrspacecast to represent null pointers in private and local address spaces.
In LLVM codegen, the static variable initializer is lowered by virtual function AsmPrinter::lowerConstant which is target generic. Since addrspacecast is target specific, AsmPrinter::lowerConst
This patch overrides AsmPrinter::lowerConstant with AMDGPUAsmPrinter::lowerConstant, which is able to lower the target-specific addrspacecast in the null pointer representation so that -1 is co
Differential Revision: https://reviews.llvm.org/D29284
llvm-svn: 294265
AArch64 was asserting when it was asked to provide a register-bank of a size it
couldn't deal with (in this case an s128 IMPLICIT_DEF). But we want a robust
fallback path so this isn't allowed.
llvm-svn: 294248
We don't handle all cases yet (see arm64-fallback.ll for an example), but this
is enough to cover most common C++ code so it's a good place to start.
llvm-svn: 294247
When splitting up one store into several in splitStoreSplat we have to
make sure we get the MachinePointerInfo right, otherwise alias
analysis thinks they all store to the same location. This can then
cause invalid scheduling later on.
Differential Revision: https://reviews.llvm.org/D29446
llvm-svn: 294203
The full codegen is committed for larger multiply, so that won't make the test suite more fragile. However, it'll allow to expose the effects fo various DAG combine.
llvm-svn: 294196
Currently we only combine shuffle nodes if they have a single user to prevent us from causing code bloat by splitting the shuffles into several different combines.
We don't take into account that in some cases we will already have combined all the users during recursively calling up the shuffle tree.
This patch keeps a list of all the shuffle nodes that have been combined so far and permits combining of further shuffle nodes if all its users are in that list.
Differential Revision: https://reviews.llvm.org/D29399
llvm-svn: 294183
Similar to what we already do for zero elt insertion, we can quickly rematerialize 'allbits' vectors so to avoid a unnecessary gpr value and insertion into a vector
llvm-svn: 294162
Similar was already done for several other shuffles in this function.
The test changes are because the old code used explicity zeroing for elements that could have been undef.
While I was here I also changed other shuffle vectors in the same function to use the same input twice instead of creating UNDEF nodes. getVectorShuffle can create the UNDEF for us.
llvm-svn: 294130
Summary:
Without this change, the getVR() call would hit an assert since it was
being passed a physical register.
Update the AArch64/ldst-opt.ll test with a case that triggers this
behavior by adding a run with strict-align, which causes an unaligned
STR XZR instruction to be split into byte stores, creating an
EXTRACT_SUBREG of XZR that triggers the original problem.
Reviewers: bogner, qcolombet, MatzeB, atrick
Subscribers: aemerson, mcrosier, llvm-commits
Differential Revision: https://reviews.llvm.org/D29495
llvm-svn: 294129
Summary: This avoid the need to duplicate all pattern and actually end up exposing some opportunity to optimize existing pattern that did not exists in both directions on an existing test case.
Reviewers: mkuper, spatel, bkramer, RKSimon, zvi
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29541
llvm-svn: 294125
This was originally introduced in r278321 to work around correctness
problems in the ExecutionDepsFix pass; Probably also to keep the
performance benefits of breaking the false dependencies which of course
also affect undef operands.
ExecutionDepsFix has been improved here recently (see for example
r278321) so we should not need this exception any longer.
Differential Revision: https://reviews.llvm.org/D29525
llvm-svn: 294087
An assert occurs when calling SlotIndexes::getInstructionIndex with
a DBG_VALUE instruction because the function expects an instruction
with a slot index. However, there is no slot index for a DBG_VALUE
instruction.
Differential Revision: https://reviews.llvm.org/D29048
llvm-svn: 294070
This re-applies commit r292189, reverted in r292191.
SelectionDAGBuilder recognizes libfuncs using some homegrown
parameter type-checking.
Use TLI instead, removing another heap of redundant code.
This isn't strictly NFC, as the SDAG code was too lax.
Concretely, this means changes are required to a few tests:
- calling a non-variadic function via a variadic prototype isn't OK;
it just happens to work on x86_64 (but not on, e.g., aarch64).
- mempcpy has a size_t parameter; the SDAG code accepts any integer
type, which meant using i32 on x86_64 worked.
- a handful of SystemZ tests check the SDAG support for lax prototype
checking: Ulrich agrees on removing them.
I don't think it's worth supporting any of these (IMO) invalid
testcases. Instead, fix them to be more meaningful.
llvm-svn: 294028
ISD::DELETED_NODE && "NodeToMatch was removed partway through
selection"' failed.
NodeToMatch can be modified during matching, but code does not handle
this situation.
Differential Revision: https://reviews.llvm.org/D29292
llvm-svn: 294003
Summary:
The tail call optimisation is performed before register allocation, so
at that point we don't know if LR is being spilt or not. If LR was spilt
to the stack, then we cannot do a tail call optimisation. That would
involve popping back into LR which is not possible in Thumb1 code.
Reviewers: rengolin, jmolloy, rovka, olista01
Reviewed By: olista01
Subscribers: llvm-commits, aemerson
Differential Revision: https://reviews.llvm.org/D29020
llvm-svn: 294000
Summary:
llc would hit a fatal error for errors in inline assembly. The
diagnostics message is now printed.
Reviewers: rengolin, MatzeB, javed.absar, anemet
Reviewed By: anemet
Subscribers: jyknight, nemanjai, llvm-commits
Differential Revision: https://reviews.llvm.org/D29408
llvm-svn: 293999
On one test this seems to have given more chance for DAG combine to do other INSERT_SUBVECTOR/EXTRACT_SUBVECTOR combines before the BLENDI was created. Looks like we can still improve more by teaching DAG combine to optimize INSERT_SUBVECTOR/EXTRACT_SUBVECTOR with BLENDI.
llvm-svn: 293944
This is the second in the series of patches to enable adding
of machine sched-models for ARM processors easier and compact.
This patch focuses on integer instructions and adds missing
sched definitions.
Reviewers: rovka, rengolin
Differential Revision: https://reviews.llvm.org/D29127
llvm-svn: 293935
Recommiting after fixing X86 inc/dec chain bug.
* Simplify Consecutive Merge Store Candidate Search
Now that address aliasing is much less conservative, push through
simplified store merging search and chain alias analysis which only
checks for parallel stores through the chain subgraph. This is cleaner
as the separation of non-interfering loads/stores from the
store-merging logic.
When merging stores search up the chain through a single load, and
finds all possible stores by looking down from through a load and a
TokenFactor to all stores visited.
This improves the quality of the output SelectionDAG and the output
Codegen (save perhaps for some ARM cases where we correctly constructs
wider loads, but then promotes them to float operations which appear
but requires more expensive constant generation).
Some minor peephole optimizations to deal with improved SubDAG shapes (listed below)
Additional Minor Changes:
1. Finishes removing unused AliasLoad code
2. Unifies the chain aggregation in the merged stores across code
paths
3. Re-add the Store node to the worklist after calling
SimplifyDemandedBits.
4. Increase GatherAllAliasesMaxDepth from 6 to 18. That number is
arbitrary, but seems sufficient to not cause regressions in
tests.
5. Remove Chain dependencies of Memory operations on CopyfromReg
nodes as these are captured by data dependence
6. Forward loads-store values through tokenfactors containing
{CopyToReg,CopyFromReg} Values.
7. Peephole to convert buildvector of extract_vector_elt to
extract_subvector if possible (see
CodeGen/AArch64/store-merge.ll)
8. Store merging for the ARM target is restricted to 32-bit as
some in some contexts invalid 64-bit operations are being
generated. This can be removed once appropriate checks are
added.
This finishes the change Matt Arsenault started in r246307 and
jyknight's original patch.
Many tests required some changes as memory operations are now
reorderable, improving load-store forwarding. One test in
particular is worth noting:
CodeGen/PowerPC/ppc64-align-long-double.ll - Improved load-store
forwarding converts a load-store pair into a parallel store and
a memory-realized bitcast of the same value. However, because we
lose the sharing of the explicit and implicit store values we
must create another local store. A similar transformation
happens before SelectionDAG as well.
Reviewers: arsenm, hfinkel, tstellarAMD, jyknight, nhaehnle
llvm-svn: 293893
It is important to change the ArgInfo's type from pointer to integer, otherwise
the CC assign function won't know what to do. Instead of hacking it up, we use
ComputeValueVTs and introduce some of the helpers that we will need later on for
lowering more complex types.
llvm-svn: 293889
This is a first attempt at using the MOVMSK instructions to replace all_of/any_of reduction patterns (i.e. an and/or + shuffle chain).
So far this only matches patterns where we are reducing an all/none bits source vector (i.e. a comparison result) but we should be able to expand on this in conjunction with improvements to 'bool vector' handling both in the x86 backend as well as the vectorizers etc.
Differential Revision: https://reviews.llvm.org/D28810
llvm-svn: 293880
The operand types were defined to fit the fp16_to_fp node, which
has the half as an integer type. v_cvt_f32_f16 does support
source modifiers, so change this to have an FP type and modifiers.
For targets without legal f16, this requires recognizing the
bit operations and trying to produce them.
llvm-svn: 293857
Functions matching LDS use to occupancy return results for a workgroup
of 64 workitems. The numbers has to be adjusted for bigger workgroups.
For example a workgroup of size 256 already occupies 4 waves just by
itself. Given that all numbers of LDS use in the compiler are per
workgroup, occupancy shall be multiplied by 4 in this case. Each 64
workitems still limited by the same number, but 4 subrgoups 64 workitems
each can afford 4 times more LDS to get the same occupancy.
In addition change initializes LDS size in the subtarget to a real value
for SI+ targets. This is required since LDS size is a variable in these
calculations.
Differential Revision: https://reviews.llvm.org/D29423
llvm-svn: 293837
Summary:
Currently the test implicit-null-checks.mir crashes if we run llc with
-enable-implicit-null-checks -start-before implicit-null-checks
options. Change fixes the RET instruction causing the crash.
Patch by Serguei Katkov!
Reviewers: sanjoy, reames
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29390
llvm-svn: 293789
These were simply preserving the flags of the original operation,
which was too conservative in most cases and incorrect for mul.
nsw/nuw may be needed for some combines to cleanup messes when
intermediate sext_inregs are introduced later.
Tested valid combinations with alive.
llvm-svn: 293776
Summary:
This change allows a re-order of two intructions if their uses
are overlapped.
Patch by Serguei Katkov!
Reviewers: reames, sanjoy
Reviewed By: sanjoy
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29120
llvm-svn: 293775
The the following instructions:
- LD/LWZ (expanded from sjLj pseudo-instructions)
- LXVL/LXVLL vector loads
- STXVL/STXVLL vector stores
all require G8RC_NO0X class registers for RA.
Differential Revision: https://reviews.llvm.org/D29289
Committed for Lei Huang
llvm-svn: 293769
Add both cores to the target parser and TableGen. Test that eabi
attributes are set correctly for both cores. Additionally, test the
absence and presence of MOVT in Cortex-M23 and Cortex-M33, respectively.
Committed on behalf of Sanne Wouda.
Reviewers : rengolin, olista01.
Differential Revision: https://reviews.llvm.org/D29073
llvm-svn: 293761
This patch moves the class for scheduling adjacent instructions,
MacroFusion, to the target.
In AArch64, it also expands the fusion to all instructions pairs in a
scheduling block, beyond just among the predecessors of the branch at the
end.
Differential revision: https://reviews.llvm.org/D28489
llvm-svn: 293737
When choosing the best successor for a block, ordinarily we would have preferred
a block that preserves the CFG unless there is a strong probability the other
direction. For small blocks that can be duplicated we now skip that requirement
as well, subject to some simple frequency calculations.
Differential Revision: https://reviews.llvm.org/D28583
llvm-svn: 293716
x*rsqrt(x) returns NaN for x == 0, whereas 1/rsqrt(x) returns 0, as
desired.
Verified that the particular nvptx approximate instructions here do in
fact return 0 for x = 0.
llvm-svn: 293713
Well, sort of. But the lower-level code that invoke used to be using completely
botched the handling of varargs functions, which hopefully won't be possible if
they're using the same code.
llvm-svn: 293670
Summary:
For some reason instructions are being inserted in the wrong order with some
builds. I'm not sure why this is happening.
Reviewers: arsenm
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, tony-tye, tpr, llvm-commits
Differential Revision: https://reviews.llvm.org/D29325
llvm-svn: 293639
Summary:
The affected transforms all implicitly use associativity of addition,
for which we usually require unsafe math to be enabled.
The "Aggressive" flag is only meant to convey information about the
performance of the fused ops relative to a fmul+fadd sequence.
Fixes Bug 31626.
Reviewers: spatel, hfinkel, mehdi_amini, arsenm, tstellarAMD
Subscribers: jholewinski, nemanjai, wdng, llvm-commits
Differential Revision: https://reviews.llvm.org/D28675
llvm-svn: 293635
The Requires class overrides the target requirements of an instruction,
rather than adding to them, so all ARM instructions need to include the
IsARM predicate when they have overwitten requirements.
This caused the swp and swpb instructions to be allowed in thumb mode
assembly, and the ARM encoding of CDP to be selected in codegen (which
is different for conditional instructions).
Differential Revision: https://reviews.llvm.org/D29283
llvm-svn: 293634
Also add the ability to recognise PINSR(Vex, 0, Idx).
Targets shuffle combines won't replace multiple insertions with a bit mask until a depth of 3 or more, so we avoid codesize bloat.
The unnecessary vpblendw in clearupper8xi16a will be fixed in an upcoming patch.
llvm-svn: 293627
Just adds the vmr (Vector Move Register) mnemonic for the VOR instruction in
the PPC back end.
Committing on behalf of brunoalr (Bruno Rosa).
Differential Revision: https://reviews.llvm.org/D29133
llvm-svn: 293626
Summary:
This lets us lower to sqrt.approx and rsqrt.approx under more
circumstances.
* Now we emit sqrt.approx and rsqrt.approx for calls to @llvm.sqrt.f32,
when fast-math is enabled. Previously, we only would emit it for
calls to @llvm.nvvm.sqrt.f. (With this patch we no longer emit
sqrt.approx for calls to @llvm.nvvm.sqrt.f; we rely on intcombine to
simplify llvm.nvvm.sqrt.f into llvm.sqrt.f32.)
* Now we emit the ftz version of rsqrt.approx when ftz is enabled.
Previously, we only emitted rsqrt.approx when ftz was disabled.
Reviewers: hfinkel
Subscribers: llvm-commits, tra, jholewinski
Differential Revision: https://reviews.llvm.org/D28508
llvm-svn: 293605
I think this is safe as long as no inputs are known to ever
be nans.
Also add an intrinsic for fmed3 to be able to handle all safe
math cases.
llvm-svn: 293598
Summary:
In revision rL278321, ExecutionDepsFix learned how to pick a better
register for undef register reads, e.g. for instructions such as
`vcvtsi2sdq`. While this revision improved performance on a good number
of our benchmarks, it unfortunately also caused significant regressions
(up to 3x) on others. This regression turned out to be caused by loops
such as:
PH -> A -> B (xmm<Undef> -> xmm<Def>) -> C -> D -> EXIT
^ |
+----------------------------------+
In the previous version of the clearance calculation, we would visit
the blocks in order, remembering for each whether there were any
incoming backedges from blocks that we hadn't processed yet and if
so queuing up the block to be re-processed. However, for loop structures
such as the above, this is clearly insufficient, since the block B
does not have any unknown backedges, so we do not see the false
dependency from the previous interation's Def of xmm registers in B.
To fix this, we need to consider all blocks that are part of the loop
and reprocess them one the correct clearance values are known. As
an optimization, we also want to avoid reprocessing any later blocks
that are not part of the loop.
In summary, the iteration order is as follows:
Before: PH A B C D A'
Corrected (Naive): PH A B C D A' B' C' D'
Corrected (w/ optimization): PH A B C A' B' C' D
To facilitate this optimization we introduce two new counters for each
basic block. The first counts how many of it's predecssors have
completed primary processing. The second counts how many of its
predecessors have completed all processing (we will call such a block
*done*. Now, the criteria to reprocess a block is as follows:
- All Predecessors have completed primary processing
- For x the number of predecessors that have completed primary
processing *at the time of primary processing of this block*,
the number of predecessors that are done has reached x.
The intuition behind this criterion is as follows:
We need to perform primary processing on all predecessors in order to
find out any direct defs in those predecessors. When predecessors are
done, we also know that we have information about indirect defs (e.g.
in block B though that were inherited through B->C->A->B). However,
we can't wait for all predecessors to be done, since that would
cause cyclic dependencies. However, it is guaranteed that all those
predecessors that are prior to us in reverse postorder will be done
before us. Since we iterate of the basic blocks in reverse postorder,
the number x above, is precisely the count of the number of predecessors
prior to us in reverse postorder.
Reviewers: myatsina
Differential Revision: https://reviews.llvm.org/D28759
llvm-svn: 293571
Since we have no call support and late linking we can produce code
only for used symbols. This saves compilation time, size of the final
executable, and size of any intermediate dumps.
Run Internalize pass early in the opt pipeline followed by global
DCE pass. To enable it RT can pass -amdgpu-internalize-symbols option.
Differential Revision: https://reviews.llvm.org/D29214
llvm-svn: 293549
For some reason the exception selector register must be a pointer (that's
assumed by SDag); on the other hand, it gets moved into an IR-level type which
might be entirely different (i32 on AArch64). IRTranslator needs to be aware of
this.
llvm-svn: 293546
This is worse if the original constant is an inline immediate.
This should also be done for 64-bit adds, but requires fixing
operand folding bugs first.
llvm-svn: 293540
Matt Arsenault