This enables subreg liveness in the arm backend when MVE is present,
which allows the register allocator to detect when subregister are
alive/dead, compared to only acting on full registers. This can helps
produce better code on MVE with the way MQPR registers are made up of
SPR registers, but is especially helpful for MQQPR and MQQQQPR
registers, where there are very few "registers" available and being able
to split them up into subregs can help produce much better code.
Differential Revision: https://reviews.llvm.org/D107642
This adjusts the place that the t2DoLoopStart reg allocation hint is
inserted, adding it in the ARMTPAndVPTOptimizaionPass in a similar place
as other tail predicated loop optimizations. This removes the need for
doing so in a custom inserter, and should make the hint more accurate,
only adding it where we expect to create a DLS (not DLSTP or WLS).
Recently we improved the lowering of low overhead loops and tail
predicated loops, but concentrated first on the DLS do style loops. This
extends those improvements over to the WLS while loops, improving the
chance of lowering them successfully. To do this the lowering has to
change a little as the instructions are terminators that produce a value
- something that needs to be treated carefully.
Lowering starts at the Hardware Loop pass, inserting a new
llvm.test.start.loop.iterations that produces both an i1 to control the
loop entry and an i32 similar to the llvm.start.loop.iterations
intrinsic added for do loops. This feeds into the loop phi, properly
gluing the values together:
%wls = call { i32, i1 } @llvm.test.start.loop.iterations.i32(i32 %div)
%wls0 = extractvalue { i32, i1 } %wls, 0
%wls1 = extractvalue { i32, i1 } %wls, 1
br i1 %wls1, label %loop.ph, label %loop.exit
...
loop:
%lsr.iv = phi i32 [ %wls0, %loop.ph ], [ %iv.next, %loop ]
..
%iv.next = call i32 @llvm.loop.decrement.reg.i32(i32 %lsr.iv, i32 1)
%cmp = icmp ne i32 %iv.next, 0
br i1 %cmp, label %loop, label %loop.exit
The llvm.test.start.loop.iterations need to be lowered through ISel
lowering as a pair of WLS and WLSSETUP nodes, which each get converted
to t2WhileLoopSetup and t2WhileLoopStart Pseudos. This helps prevent
t2WhileLoopStart from being a terminator that produces a value,
something difficult to control at that stage in the pipeline. Instead
the t2WhileLoopSetup produces the value of LR (essentially acting as a
lr = subs rn, 0), t2WhileLoopStart consumes that lr value (the Bcc).
These are then converted into a single t2WhileLoopStartLR at the same
point as t2DoLoopStartTP and t2LoopEndDec. Otherwise we revert the loop
to prevent them from progressing further in the pipeline. The
t2WhileLoopStartLR is a single instruction that takes a GPR and produces
LR, similar to the WLS instruction.
%1:gprlr = t2WhileLoopStartLR %0:rgpr, %bb.3
t2B %bb.1
...
bb.2.loop:
%2:gprlr = PHI %1:gprlr, %bb.1, %3:gprlr, %bb.2
...
%3:gprlr = t2LoopEndDec %2:gprlr, %bb.2
t2B %bb.3
The t2WhileLoopStartLR can then be treated similar to the other low
overhead loop pseudos, eventually being lowered to a WLS providing the
branches are within range.
Differential Revision: https://reviews.llvm.org/D97729
This adds some simple known bits handling for the three CSINC/NEG/INV
instructions. From the operands known bits we can compute the common
bits of the first operand and incremented/negated/inverted second
operand. The first, especially CSINC ZR, ZR, comes up fair amount in the
tests. The others are more rare so a unit test for them is added.
Differential Revision: https://reviews.llvm.org/D97788
This removes the existing patterns for inserting two lanes into an
f16/i16 vector register using VINS, instead using a DAG combine to
pattern match the same code sequences. The tablegen patterns were
already on the large side (foreach LANE = [0, 2, 4, 6]) and were not
handling all the cases they could. Moving that to a DAG combine, whilst
not less code, allows us to better control and expand the selection of
VINSs. Additionally this allows us to remove the AddedComplexity on
VCVTT.
The extra trick that this has learned in the process is to move two
adjacent lanes using a single f32 vmov, allowing some extra
inefficiencies to be removed.
Differenial Revision: https://reviews.llvm.org/D96876
Currently the findIncDecAfter will only look at the next instruction for
post-inc candidates in the load/store optimizer. This extends that to a
search through the current BB, until an instruction that modifies or
uses the increment reg is found. This allows more post-inc load/stores
and ldm/stm's to be created, especially in cases where a schedule might
move instructions further apart.
We make sure not to look any further for an SP, as that might invalidate
stack slots that are still in use.
Differential Revision: https://reviews.llvm.org/D95881
This patch adds tablegen patterns for pairs of i16/f16 insert/extracts.
If we are inserting into two adjacent vector lanes (0 and 1 for
example), we can use either a vmov;vins or vmovx;vins to insert the pair
together, avoiding a round-trip from GRP registers. This is quite a
large patterns with a number of EXTRACT_SUBREG/INSERT_SUBREG/
COPY_TO_REGCLASS nodes, but hopefully as most of those become copies all
that will be cleaned up by further optimizations.
The VINS pattern was also adjusted to allow it to represent that it is
inserting into the top half of an existing register.
Differential Revision: https://reviews.llvm.org/D95381
A DLS lr, lr instruction only moves lr to itself. It need not be emitted
on it's own to save a instruction in the loop preheader.
Differential Revision: https://reviews.llvm.org/D78916
When LSR converts a branch on the pre-inc IV into a branch on the
post-inc IV, the nowrap flags on the addition may no longer be valid.
Previously, a poison result of the addition might have been ignored,
in which case the program was well defined. After branching on the
post-inc IV, we might be branching on poison, which is undefined behavior.
Fix this by discarding nowrap flags which are not present on the SCEV
expression. Nowrap flags on the SCEV expression are proven by SCEV
to always hold, independently of how the expression will be used.
This is essentially the same fix we applied to IndVars LFTR, which
also performs this kind of pre-inc to post-inc conversion.
I believe a similar problem can also exist for getelementptr inbounds,
but I was not able to come up with a problematic test case. The
inbounds case would have to be addressed in a differently anyway
(as SCEV does not track this property).
Fixes https://bugs.llvm.org/show_bug.cgi?id=46943.
Differential Revision: https://reviews.llvm.org/D95286
It turns our that the BranchFolder and IfCvt does not like unanalyzable
branches that fall-through. This means that removing the unconditional
branches from the end of tail predicated instruction can run into
asserts and verifier issues.
This effectively reverts 372eb2bbb6, but
adds handling to t2DoLoopEndDec which are not branches, so can be safely
skipped.
This treats low overhead loop branches the same as jump tables and
indirect branches in analyzeBranch - they cannot be analyzed but the
direct branches on the end of the block may be removed. This helps
remove the unnecessary branches earlier, which can help produce better
codegen (and change block layout in a number of cases).
Differential Revision: https://reviews.llvm.org/D94392
We currently have problems with the way that low overhead loops are
specified, with LR being spilled between the t2LoopDec and the t2LoopEnd
forcing the entire loop to be reverted late in the backend. As they will
eventually become a single instruction, this patch introduces a
t2LoopEndDec which is the combination of the two, combined before
registry allocation to make sure this does not fail.
Unfortunately this instruction is a terminator that produces a value
(and also branches - it only produces the value around the branching
edge). So this needs some adjustment to phi elimination and the register
allocator to make sure that we do not spill this LR def around the loop
(needing to put a spill after the terminator). We treat the loop very
carefully, making sure that there is nothing else like calls that would
break it's ability to use LR. For that, this adds a
isUnspillableTerminator to opt in the new behaviour.
There is a chance that this could cause problems, and so I have added an
escape option incase. But I have not seen any problems in the testing
that I've tried, and not reverting Low overhead loops is important for
our performance. If this does work then we can hopefully do the same for
t2WhileLoopStart and t2DoLoopStart instructions.
This patch also contains the code needed to convert or revert the
t2LoopEndDec in the backend (which just needs a subs; bne) and the code
pre-ra to create them.
Differential Revision: https://reviews.llvm.org/D91358
This checks to see if the loop will likely become a tail predicated loop
and disables wls loop generation if so, as the likelihood for reverting
is currently too high. These should be fairly rare situations anyway due
to the way iterations and element counts are used during lowering. Just
not trying can alter how SCEV's are materialized however, leading to
different codegen.
It also adds a option to disable all while low overhead loops, for
debugging.
Differential Revision: https://reviews.llvm.org/D91663
This hints the operand of a t2DoLoopStart towards using LR, which can
help make it more likely to become t2DLS lr, lr. This makes it easier to
move if needed (as the input is the same as the output), or potentially
remove entirely.
The hint is added after others (from COPY's etc) which still take
precedence. It needed to find a place to add the hint, which currently
uses the post isel custom inserter.
Differential Revision: https://reviews.llvm.org/D89883
This changes the definition of t2DoLoopStart from
t2DoLoopStart rGPR
to
GPRlr = t2DoLoopStart rGPR
This will hopefully mean that low overhead loops are more tied together,
and we can more reliably generate loops without reverting or being at
the whims of the register allocator.
This is a fairly simple change in itself, but leads to a number of other
required alterations.
- The hardware loop pass, if UsePhi is set, now generates loops of the
form:
%start = llvm.start.loop.iterations(%N)
loop:
%p = phi [%start], [%dec]
%dec = llvm.loop.decrement.reg(%p, 1)
%c = icmp ne %dec, 0
br %c, loop, exit
- For this a new llvm.start.loop.iterations intrinsic was added, identical
to llvm.set.loop.iterations but produces a value as seen above, gluing
the loop together more through def-use chains.
- This new instrinsic conceptually produces the same output as input,
which is taught to SCEV so that the checks in MVETailPredication are not
affected.
- Some minor changes are needed to the ARMLowOverheadLoop pass, but it has
been left mostly as before. We should now more reliably be able to tell
that the t2DoLoopStart is correct without having to prove it, but
t2WhileLoopStart and tail-predicated loops will remain the same.
- And all the tests have been updated. There are a lot of them!
This patch on it's own might cause more trouble that it helps, with more
tail-predicated loops being reverted, but some additional patches can
hopefully improve upon that to get to something that is better overall.
Differential Revision: https://reviews.llvm.org/D89881
This adapts tail-predication to the new semantics of get.active.lane.mask as
defined in D86147. This means that:
- we can remove the BTC + 1 overflow checks because now the loop tripcount is
passed in to the intrinsic,
- we can immediately use that value to setup a counter for the number of
elements processed by the loop and don't need to materialize BTC + 1.
Differential Revision: https://reviews.llvm.org/D86303
This refactors option -disable-mve-tail-predication to take different arguments
so that we have 1 option to control tail-predication rather than several
different ones.
This is also a prep step for D82953, in which we want to reject reductions
unless that is requested with this option.
Differential Revision: https://reviews.llvm.org/D83133
We current extract and convert from a top lane of a f16 vector using a
VMOVX;VCVTB pair. We can simplify that to use a single VCVTT. The
pattern is mostly copied from a vector extract pattern, but produces a
VCVTTHS f32 directly.
This had to move some code around so that ARMInstrVFP had access to the
required pattern frags that were previously part of ARMInstrNEON.
Differential Revision: https://reviews.llvm.org/D81556
To set up a tail-predicated loop, we need to to calculate the number of
elements processed by the loop. We can now use intrinsic
@llvm.get.active.lane.mask() to do this, which is emitted by the vectoriser in
D79100. This intrinsic generates a predicate for the masked loads/stores, and
consumes the Backedge Taken Count (BTC) as its second argument. We can now use
that to reconstruct the loop tripcount, instead of the IR pattern match
approach we were using before.
Many thanks to Eli Friedman and Sam Parker for all their help with this work.
This also adds overflow checks for the different, new expressions that we
create: the loop tripcount, and the sub expression that calculates the
remaining elements to be processed. For the latter, SCEV is not able to
calculate precise enough bounds, so we work around that at the moment, but is
not entirely correct yet, it's conservative. The overflow checks can be
overruled with a force flag, which is thus potentially unsafe (but not really
because the vectoriser is the only place where this intrinsic is emitted at the
moment). It's also good to mention that the tail-predication pass is not yet
enabled by default. We will follow up to see if we can implement these
overflow checks better, either by a change in SCEV or we may want revise the
definition of llvm.get.active.lane.mask.
Differential Revision: https://reviews.llvm.org/D79175
Unlike Neon, MVE does not have a way of duplicating from a vector lane,
so a VDUPLANE currently selects to a VDUP(move_from_lane(..)). This
forces that to be done earlier as a dag combine to allow other folds to
happen.
It converts to a VDUP(EXTRACT). On FP16 this is then folded to a
VGETLANEu to prevent it from creating a vmovx;vmovhr pair, using a
single move_from_reg instead.
Differential Revision: https://reviews.llvm.org/D79606
Printing floating point number in decimal is inconvenient for humans.
Verbose asm output will print out floating point values in comments, it
helps.
But in lots of cases, users still need additional work to covert the
decimal back to hex or binary to check the bit patterns,
especially when there are small precision difference.
Hexadecimal form is one of the supported form in LLVM IR, and easier for
debugging.
This patch try to print all FP constant in hex form instead.
Reviewed By: RKSimon
Differential Revision: https://reviews.llvm.org/D73566
This patch uses helper function rewriteLoopExitValues that is refactored in
D72602 to rematerialise the iteration count in exit blocks, so that we can
clean-up loop update expressions inside the hardware-loops later in
ARMLowOverheadLoops, which is necessary to get actual performance gains for
tail-predicated loops.
Differential Revision: https://reviews.llvm.org/D72714
We were previously not necessarily favouring postinc for the MVE loads
and stores, leading to extra code prior to the loop to set up the
preinc. MVE in general can benefit from postinc (as we don't have
unrolled loops), and certain instructions like the VLD2's only post-inc
versions are available.
Differential Revision: https://reviews.llvm.org/D70790
Introduce a method to walk through use-def chains to decide whether
it's possible to remove a given instruction and its users. These
instructions are then stored in a set until the end of the transform
when they're erased. This is now used to perform checks on the
iteration count (LoopDec chain), element count (VCTP chain) and the
possibly redundant iteration count.
As well as being able to remove chains of instructions, we know also
check that the sub feeding the vctp is producing the expected value.
Differential Revision: https://reviews.llvm.org/D71837
Architecturally, it's allowed to have MVE-I without an FPU, thus
-mfpu=none should not disable MVE-I, or moves to/from FP-registers.
This patch removes `+/-fpregs` from features unconditionally added to
target feature list, depending on FPU and moves the logic to Clang
driver, where the negative form (`-fpregs`) is conditionally added to
the target features list for the cases of `-mfloat-abi=soft`, or
`-mfpu=none` without either `+mve` or `+mve.fp`. Only the negative
form is added by the driver, the positive one is derived from other
features in the backend.
Differential Revision: https://reviews.llvm.org/D71843
While we do manage to fold integer-typed IR in middle-end,
we can't do that for the main motivational case of pointers.
There is @llvm.ptrmask() intrinsic which may or may not be helpful,
but i'm not sure it is fully considered canonical yet,
not everything is fully aware of it likely.
Name: PR44448 ptr - (ptr & C) -> ptr & (~C)
%bias = and i32 %ptr, C
%r = sub i32 %ptr, %bias
=>
%r = and i32 %ptr, ~C
See
https://bugs.llvm.org/show_bug.cgi?id=44448https://reviews.llvm.org/D71499
This adds ICmp to the list of instructions that we sink a splat to in a
loop, allowing the register forms of instructions to be selected more
often. It does not add FCmp yet as the results look a little odd, trying
to keep the register in an float reg and having to move it back to a GPR.
Differential Revision: https://reviews.llvm.org/D70997
After creating a low-overhead loop, the loop update instruction was still
lingering around hurting performance. This removes dead loop update
instructions, which in our case are mostly SUBS instructions.
To support this, some helper functions were added to MachineLoopUtils and
ReachingDefAnalysis to analyse live-ins of loop exit blocks and find uses
before a particular loop instruction, respectively.
This is a first version that removes a SUBS instruction when there are no other
uses inside and outside the loop block, but there are some more interesting
cases in test/CodeGen/Thumb2/LowOverheadLoops/mve-tail-data-types.ll which
shows that there is room for improvement. For example, we can't handle this
case yet:
..
dlstp.32 lr, r2
.LBB0_1:
mov r3, r2
subs r2, #4
vldrh.u32 q2, [r1], #8
vmov q1, q0
vmla.u32 q0, q2, r0
letp lr, .LBB0_1
@ %bb.2:
vctp.32 r3
..
which is a lot more tricky because r2 is not only used by the subs, but also by
the mov to r3, which is used outside the low-overhead loop by the vctp
instruction, and that requires a bit of a different approach, and I will follow
up on this.
Differential Revision: https://reviews.llvm.org/D71007
With the extra optimisations we have done, these should now be fine to
enable by default. Which is what this patch does.
Differential Revision: https://reviews.llvm.org/D70968
MVE has a basic symmetry between it's normal loads/store operations and
the masked variants. This means that masked loads and stores can use
pre-inc and post-inc addressing modes, just like the standard loads and
stores already do.
To enable that, this patch adds all the relevant infrastructure for
treating masked loads/stores addressing modes in the same way as normal
loads/stores.
This involves:
- Adding an AddressingMode to MaskedLoadStoreSDNode, along with an extra
Offset operand that is added after the PtrBase.
- Extending the IndexedModeActions from 8bits to 16bits to store the
legality of masked operations as well as normal ones. This array is
fairly small, so doubling the size still won't make it very large.
Offset masked loads can then be controlled with
setIndexedMaskedLoadAction, similar to standard loads.
- The same methods that combine to indexed loads, such as
CombineToPostIndexedLoadStore, are adjusted to handle masked loads in
the same way.
- The ARM backend is then adjusted to make use of these indexed masked
loads/stores.
- The X86 backend is adjusted to hopefully be no functional changes.
Differential Revision: https://reviews.llvm.org/D70176
Add some more helper functions to ReachingDefs to query the uses of
a given MachineInstr and also to query whether two MachineInstrs use
the same def of a register.
For Arm, while tail-predicating, these helpers are used in the
low-overhead loops to remove the dead code that calculates the number
of loop iterations.
Differential Revision: https://reviews.llvm.org/D70240
This patch modifies ARMLowOverheadLoops to convert a predicated
vector low-overhead loop into a tail-predicatd one. This is currently
a very basic conversion, with the following restrictions:
- Operates only on single block loops.
- The loop can only contain a single vctp instruction.
- No other instructions can write to the vpr.
- We only allow a subset of the mve instructions in the loop.
TODO: Pass the number of elements, not the number of iterations to
dlstp/wlstp.
Differential Revision: https://reviews.llvm.org/D69945
The Arm backend will usually return false for isFMAFasterThanFMulAndFAdd,
where both the fused VFMA.f32 and a non-fused VMLA.f32 are usually
available for scalar code. For MVE we don't have the non-fused version
though. It makes more sense for isFMAFasterThanFMulAndFAdd to return
true, allowing us to simplify some of the existing ISel patterns.
The tests here are that non of the existing tests failed, and so we are
still selecting VFMA and VFMS. The one test that changed shows we can
now select from fast math flags, as opposed to just relying on the
isFMADLegalForFAddFSub option.
Differential Revision: https://reviews.llvm.org/D69115
David Green