Creates a new scheduling strategy that attempts to maximize ILP for a single
wave.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D130869
In the 2e29b0138c we introduce a specific solving algorithm
that analyzes the VGPR to SGPR copies use chains and either lowers
the copy to v_readfirstlane_b32 or converts the whole chain to VALU forms.
Same time we still have the code that blindly converts to VALU REG_SEQUENCE and PHIs
in case they produce SGPR but have VGPRs input operands. In case the REG_SEQUENCE and PHIs
are in the VGPR to SGPR copy use chain, and this chain was considered long enough to convert
copy to v_readfistlane_b32, further lowering them to VALU leads to several kinds of issues.
At first, we have v_readfistlane_b32 which is completely useless because most parts of its use chain
were moved to VALU forms. Second, we may encounter subtle bugs related to the EXEC-dependent CF
because of the weird mixing of SALU and VALU instructions.
This change removes the code that moves REG_SEQUENCE and PHIs to VALU. Instead, we use the fact
that both REG_SEQUENCE and PHIs have copy semantics. That is, if they define SGPR but have VGPR inputs,
we insert VGPR to SGPR copies to make them pure SGPR. Then, the new copies are processed by the common
VGPR to SGPR lowering algorithm.
This is Part 2 in the series of commits aiming at the massive refactoring of the SIFixSGPRCopies pass.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D130367
This pass seems to have very little effect because all it does is hoist
some instructions, but it is followed later in the codegen pipeline by
the IR CodeSinking pass which does the opposite.
Differential Revision: https://reviews.llvm.org/D130258
This improves a corner case where v_fmac can be converted to v_fma on
GFX10+ even if it has a literal operand.
Differential Revision: https://reviews.llvm.org/D130992
For VALU write and memory (VM, L/DS, FLAT) instructions, SQ would insert
wait-states to avoid data hazard. However when there is a DGEMM instruction
in-between them, SQ incorrectly disables the wait-states thus the data hazard
needs to be handled with this workaround.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D130677
Extend hazard recognizer of ReadM0MovRelInterpHazard with
DS_READ_ADDTID and DS_WRITE_ADDTID, as they also
require a manually inserted S_NOP after SALU writing m0.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D130783
It is not necessary to wait for all outstanding memory operations before
barriers on hardware that can back off of the barrier in the event of an
exception when traps are enabled. Add a new subtarget feature which
tracks which HW has this ability.
Reviewed By: #amdgpu, rampitec
Differential Revision: https://reviews.llvm.org/D130722
Since 814a0abcce, this would break if we
had a function in the module that becomes dead in any codegen IR
pass. The function wasn't deleted since it was initially used in dead
code, but is detached from the call graph and doesn't appear in the PO
traversal. Do a second walk over the module to populate the resources
of any functions which weren't already processed.
Summary:
Flat scratch load of D16 type by default has tied vdst_in operand (with vdst). This should be taken
care of at the time of "removeOperand" in eliminateFrameIndex. Otherwise we will hit an assert saying
"Cannot move tied operands". This patch unties vdst_in before the move, and retie it with vdst afterwards.
Reviewers:
arsenm, foad
Differential Revision: https://reviews.llvm.org/D130537
When register pressure tracking is disabled, the scheduler attempts to load
pressures at SReg_32 and VGPR_32. This causes an index out of bounds error.
This patch fixes this issue by disabling the initialization of RPTracker
when not needed. NFC
Reviewed By: rampitec, kerbowa, arsenm
Differential Revision: https://reviews.llvm.org/D129322
This builtin allows the creation of custom scheduling pipelines on a per-region
basis. Like the sched_barrier builtin this is intended to be used either for
testing, in situations where the default scheduler heuristics cannot be
improved, or in critical kernels where users are trying to get performance that
is close to handwritten assembly. Obviously using these builtins will require
extra work from the kernel writer to maintain the desired behavior.
The builtin can be used to create groups of instructions called "scheduling
groups" where ordering between the groups is enforced by the scheduler.
__builtin_amdgcn_sched_group_barrier takes three parameters. The first parameter
is a mask that determines the types of instructions that you would like to
synchronize around and add to a scheduling group. These instructions will be
selected from the bottom up starting from the sched_group_barrier's location
during instruction scheduling. The second parameter is the number of matching
instructions that will be associated with this sched_group_barrier. The third
parameter is an identifier which is used to describe what other
sched_group_barriers should be synchronized with. Note that multiple
sched_group_barriers must be added in order for them to be useful since they
only synchronize with other sched_group_barriers. Only "scheduling groups" with
a matching third parameter will have any enforced ordering between them.
As an example, the code below tries to create a pipeline of 1 VMEM_READ
instruction followed by 1 VALU instruction followed by 5 MFMA instructions...
// 1 VMEM_READ
__builtin_amdgcn_sched_group_barrier(32, 1, 0)
// 1 VALU
__builtin_amdgcn_sched_group_barrier(2, 1, 0)
// 5 MFMA
__builtin_amdgcn_sched_group_barrier(8, 5, 0)
// 1 VMEM_READ
__builtin_amdgcn_sched_group_barrier(32, 1, 0)
// 3 VALU
__builtin_amdgcn_sched_group_barrier(2, 3, 0)
// 2 VMEM_WRITE
__builtin_amdgcn_sched_group_barrier(64, 2, 0)
Reviewed By: jrbyrnes
Differential Revision: https://reviews.llvm.org/D128158
In the 2e29b0138c we introduce a specific solving algorithm
that analyzes the VGPR to SGPR copies use chains and either lowers
the copy to v_readfirstlane_b32 or converts the whole chain to VALU forms.
Same time we still have the code that blindly converts to VALU REG_SEQUENCE and PHIs
in case they produce SGPR but have VGPRs input operands. In case the REG_SEQUENCE and PHIs
are in the VGPR to SGPR copy use chain, and this chain was considered long enough to convert
copy to v_readfistlane_b32, further lowering them to VALU leads to several kinds of issues.
At first, we have v_readfistlane_b32 which is completely useless because most parts of its use chain
were moved to VALU forms. Second, we may encounter subtle bugs related to the EXEC-dependent CF
because of the weird mixing of SALU and VALU instructions.
This change removes the code that moves REG_SEQUENCE and PHIs to VALU. Instead, we use the fact
that both REG_SEQUENCE and PHIs have copy semantics. That is, if they define SGPR but have VGPR inputs,
we insert VGPR to SGPR copies to make them pure SGPR. Then, the new copies are processed by the common
VGPR to SGPR lowering algorithm.
This is Part 2 in the series of commits aiming at the massive refactoring of the SIFixSGPRCopies pass.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D130367
By not clustering loads and adjusting heuristics to more aggressively reduce
register pressure we may be able to increase occupancy for the function if it
was dropped in a first pass scheduling.
Similarly, try to reduce spilling if register usage exceeds lower bound
occupancy.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D130329
Clear all kill flags on source register when folding a COPY.
This is necessary because the kills may now be out of order with the uses.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D130622
The instruction is used to modify wave priority with the intent
to affect VALU execution and currently we can reschedule VALU
around it since that VALU does not have side effects.
Differential Revision: https://reviews.llvm.org/D130654
I don't have any evidence these particular uses are actually causing any
issues, but we should avoid accidentally truncating immediate values
depending on the host.
It errors out in the Bazel CI:
AMDGPULowerModuleLDSPass.cpp:384:12: error: chosen constructor is
explicit in copy-initialization
return {SGV, std::move(Map)};
Reviewed By: rupprecht
Differential Revision: https://reviews.llvm.org/D130623
Tries to make the different scheduling stages a bit more self contained and
modifiable. Intended to be NFC. Preface to other changes.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D130147
Set the priorities consistently to number of registers in the tuple -
1. Previously we started at 1, and also tried to give SGPR higher
values than VGPRs. There's no point in assigning SGPRs higher values
now that those are allocated in a separate regalloc run.
This avoids overflowing the 5 bits used for the class priority in the
allocation heuristic for 32 element tuples. This avoids some cases
where smaller registers unexpectedly get prioritized over larger.
This patch merges a consecutive sequence of
s_or_saveexec s_o, s_i
s_xor exec, exec, s_o
into a single
s_andn2_saveexec s_o, s_i instruction.
This patch also cleans up the SIOptimizeExecMasking pass a bit.
Reviewed By: nhaehnle
Differential Revision: https://reviews.llvm.org/D129073
VOPC DPP should not be formed when the row_mask and bank_mask are not
0xf (full) because the resulting VOP DPP would have different semantics
than the MOV DPP followed by VOP. Existing checks in GCNDPPCombine cover
this case but for different reasons, so assert the property for
future-proofing.
Reviewed By: nhaehnle
Differential Revision: https://reviews.llvm.org/D130101
For the longest time we used `AAValueSimplify` and
`genericValueTraversal` to determine "potential values". This was
problematic for many reasons:
- We recomputed the result a lot as there was no caching for the 9
locations calling `genericValueTraversal`.
- We added the idea of "intra" vs. "inter" procedural simplification
only as an afterthought. `genericValueTraversal` did offer an option
but `AAValueSimplify` did not. Thus, we might end up with "too much"
simplification in certain situations and then gave up on it.
- Because `genericValueTraversal` was not a real `AA` we ended up with
problems like the infinite recursion bug (#54981) as well as code
duplication.
This patch introduces `AAPotentialValues` and replaces the
`AAValueSimplify` uses with it. `genericValueTraversal` is folded into
`AAPotentialValues` as are the instruction simplifications performed in
`AAValueSimplify` before. We further distinguish "intra" and "inter"
procedural simplification now.
`AAValueSimplify` was not deleted as we haven't ported the
re-materialization of instructions yet. There are other differences over
the former handling, e.g., we may not fold trivially foldable
instructions right now, e.g., `add i32 1, 1` is not folded to `i32 2`
but if an operand would be simplified to `i32 1` we would fold it still.
We are also even more aware of function/SCC boundaries in CGSCC passes,
which is good even if some tests look like they regress.
Fixes: https://github.com/llvm/llvm-project/issues/54981
Note: A previous version was flawed and consequently reverted in
6555558a80.
Implement an intrinsic for use lowering LDS variables to different
addresses from different kernels. This will allow kernels that cannot
reach an LDS variable to avoid wasting space for it.
There are a number of implicit arguments accessed by intrinsic already
so this implementation closely follows the existing handling. It is slightly
novel in that this SGPR is written by the kernel prologue.
It is necessary in the general case to put variables at different addresses
such that they can be compactly allocated and thus necessary for an
indirect function call to have some means of determining where a
given variable was allocated. Claiming an arbitrary SGPR into which
an integer can be written by the kernel, in this implementation based
on metadata associated with that kernel, which is then passed on to
indirect call sites is sufficient to determine the variable address.
The intent is to emit a __const array of LDS addresses and index into it.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D125060
For most DPP instructions, the old operand stores the value that was in
the current lane before the DPP operation, and is tied to the
destination. For VOPC DPP, this is unnecessary and incorrect.
There appears to have been a latent bug related to D122737 with
SIInstrInfo::isOperandLegal. If you checked if a register operand was legal
when the InstructionDesc expected an immediate, it reported that is valid.
Its fix is necessary for and tested in this patch.
Reviewed By: foad, rampitec
Differential Revision: https://reviews.llvm.org/D130040
AMDGPUPerfHintAnalysis doesn't set the memory bound attribute if
FuncInfo::InstCost outweighs MemInstCost even if we have a basic block
with relatively high global memory access. GCNSchedStrategy could revert
optimal scheduling in favour of occupancy which seems to degrade
performance for some kernels. This change introduces the
HasDenseGlobalMemAcc metric in the heuristic that makes the analysis
more conservative in these cases.
This fixes SWDEV-334259/SWDEV-343932
Differential Revision: https://reviews.llvm.org/D129759
This was stored in LiveIntervals, but not actually used for anything
related to LiveIntervals. It was only used in one check for if a load
instruction is rematerializable. I also don't think this was entirely
correct, since it was implicitly assuming constant loads are also
dereferenceable.
Remove this and rely only on the invariant+dereferenceable flags in
the memory operand. Set the flag based on the AA query upfront. This
should have the same net benefit, but has the possible disadvantage of
making this AA query nonlazy.
Preserve the behavior of assuming pointsToConstantMemory implying
dereferenceable for now, but maybe this should be changed.
Parse op_sel for *_e64_dpp VOP3 opcodes.
Depends on D129637 and setting of VOP3_OPSEL in dpp pseudos.
Differential Revision: https://reviews.llvm.org/D129767
Saves some add instructions on a couple Rage 2 shaders and is also a
prerequisite for a coming-soon change matching (register + immediate)
offsets.
Reviewed By: foad, arsenm
Differential Revision: https://reviews.llvm.org/D129095
This change introduces the dynamic stack boolean field to code-object-v3
and above under the code properties of the kernel descriptor and under
the kernel metadata map of NT_AMDGPU_METADATA. This field corresponds to
the is_dynamic_callstack field of amd_kernel_code_t.
Differential Revision: https://reviews.llvm.org/D128344
Further improve liveness copying for CC register post optimization
by mirroring live internal splits.
The fixes a bug in register allocation when CC register liveness
is extended across a branches instead of split.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D129557
Since the divergence-driven instruction selection has been enabled for AMDGPU,
all the uniform instructions are expected to be selected to SALU form, except those not having one.
VGPR to SGPR copies appear in MIR to connect values producers and consumers. This change implements an algorithm
that evolves a reasonable tradeoff between the profit achieved from keeping the uniform instructions in SALU form
and overhead introduced by the data transfer between the VGPRs and SGPRs.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D128252
D114999 added code to kill an immediate def if it was folded into its
only use by convertToThreeAddress. This patch updates LiveVariables when
that happens in order to fix verification failures exposed by D129213.
Differential Revision: https://reviews.llvm.org/D129661
D25618 added a method to verify the instruction predicates for an
emitted instruction, through verifyInstructionPredicates added into
<Target>MCCodeEmitter::encodeInstruction. This is a very useful idea,
but the implementation inside MCCodeEmitter made it only fire for object
files, not assembly which most of the llvm test suite uses.
This patch moves the code into the <Target>_MC::verifyInstructionPredicates
method, inside the InstrInfo. The allows it to be called from other
places, such as in this patch where it is called from the
<Target>AsmPrinter::emitInstruction methods which should trigger for
both assembly and object files. It can also be called from other places
such as verifyInstruction, but that is not done here (it tends to catch
errors earlier, but in reality just shows all the mir tests that have
incorrect feature predicates). The interface was also simplified
slightly, moving computeAvailableFeatures into the function so that it
does not need to be called externally.
The ARM, AMDGPU (but not R600), AVR, Mips and X86 backends all currently
show errors in the test-suite, so have been disabled with FIXME
comments.
Recommitted with some fixes for the leftover MCII variables in release
builds.
Differential Revision: https://reviews.llvm.org/D129506
The SI machine scheduler inherits from ScheduleDAGMI.
This patch adds support for a few features that are implemented
in ScheduleDAGMI (or its base classes) that were missing so far
because their support is implemented in overridden functions.
* Support cl::opt -view-misched-dags
This option allows to open a graphical window of the scheduling DAG.
* Support cl::opt -misched-print-dags
This option allows to print the scheduling DAG in text form.
* After constructing the scheduling DAG, call postprocessDAG()
to apply any registered DAG mutations.
Note that currently there are no mutations defined in AMDGPUTargetMachine.cpp
in case SIScheduler is used.
Still add this to avoid surprises in the future in case mutations are added.
Differential Revision: https://reviews.llvm.org/D128808
This reverts commit e2fb8c0f4b as it does
not build for Release builds, and some buildbots are giving more warning
than I saw locally. Reverting to fix those issues.
D25618 added a method to verify the instruction predicates for an
emitted instruction, through verifyInstructionPredicates added into
<Target>MCCodeEmitter::encodeInstruction. This is a very useful idea,
but the implementation inside MCCodeEmitter made it only fire for object
files, not assembly which most of the llvm test suite uses.
This patch moves the code into the <Target>_MC::verifyInstructionPredicates
method, inside the InstrInfo. The allows it to be called from other
places, such as in this patch where it is called from the
<Target>AsmPrinter::emitInstruction methods which should trigger for
both assembly and object files. It can also be called from other places
such as verifyInstruction, but that is not done here (it tends to catch
errors earlier, but in reality just shows all the mir tests that have
incorrect feature predicates). The interface was also simplified
slightly, moving computeAvailableFeatures into the function so that it
does not need to be called externally.
The ARM, AMDGPU (but not R600), AVR, Mips and X86 backends all currently
show errors in the test-suite, so have been disabled with FIXME
comments.
Differential Revision: https://reviews.llvm.org/D129506
The availability of LiveIntervals affects kill flags in the output, so
declare the use to avoid strange effects where the output of this pass
is different depending on what other passes are scheduled after it.
Differential Revision: https://reviews.llvm.org/D129555
Remove the bound in the definition, since it's not guaranteed/could
provide a false sense of security (I'd be inclined to go further and
change this to a pointer parameter, since that's what it really is - but
figured I'd preserve some of the author's intent here)
This change replaces the C++ predicates with the HasNoUse builtin
predicate that would enable the no-ret atomic op selection in
GlobalISel.
Differential Revision: https://reviews.llvm.org/D125213
This patch removes the predicate for return atomic ops and uses
AddedComplexity to distinguish its selection from its no return variant.
This will produce better matchers that doesn't unnecessarily check for
the negated predicate if the initial predicate failed. Also, it
simplifies the enabling of no return atomic ops selection in GlobalISel.
Differential Revision: https://reviews.llvm.org/D128241
This patch adds the support for `fmax` and `fmin` operations in `atomicrmw`
instruction. For now (at least in this patch), the instruction will be expanded
to CAS loop. There are already a couple of targets supporting the feature. I'll
create another patch(es) to enable them accordingly.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D127041
This patch removes a bit of code duplication and
moves the v_cmpx optimization out of the
runOnMachineFunction pass.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D129086
Merge tests and fixes from D128110 and D128315 on top of already
committed D128800.
Original author: arsenm
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D128882
Modifies the GCNDPPCombine pass to enable DPP formation for the new DPP
instruction in gfx11, namely VOP3 encoded instructions with DPP and VOPC
with DPP.
Depends on D128656
Reviewed By: #amdgpu, rampitec
Differential Revision: https://reviews.llvm.org/D128682
We form VOPD instructions in the GCNCreateVOPD pass by combining
back-to-back component instructions. There are strict register
constraints for creating a legal VOPD, namely that the matching operands
(e.g. src0x and src0y, src1x and src1y) must be in different register
banks. We add a PostRA scheduler
mutation to put possible VOPD components back-to-back.
Depends on D128442, D128270
Reviewed By: #amdgpu, rampitec
Differential Revision: https://reviews.llvm.org/D128656
Tell the matcher what we are looking for instead of matching everything
and then discarding the result if doesn't fit.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D128171
Update intrinsics to use n x f16 and n x i16 instead
of 32-bit types. This may avoid the need for a bitcast
and is probably less confusing.
Depends on making v16f16 and v16i16 types legal.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D128951
GFX11 has a new message type MSG_DEALLOC_VGPRS which can be used to
release a shader's VGPRs. Sending this at the end of a shader (just
before the s_endpgm) can help overall system performance in cases where
the s_endpgm would have to wait for outstanding VMEM stores to complete
before releasing the VGPRs.
Differential Revision: https://reviews.llvm.org/D128442
Follow up to D127894, new liveness update code needs to handle
the case where S_ANDN2 input must be extended through loops when
V_CNDMASK_B32 has been hoisted.
Reviewed By: arsenm
Differential Revision: https://reviews.llvm.org/D128800
D106023 excluded 16-bit instructions from rematerialization, with the
justification that we can't rematerialize instructions that preserve
the high bits (plus the instructions which do are a confusing mess
between different subtargets). This doesn't make sense to me as a
problem since cases where we would rely on the high bit behavior would
still need to be represented as a register value constraint with a
tied operand. It's not a hidden side effect and should still be
rematerializable.
Without this, the new test case would fail with:
AMDGPUInstPrinter.cpp:545: void llvm::AMDGPUInstPrinter::printImmediate64(uint64_t, const llvm::MCSubtargetInfo &, llvm::raw_ostream &): Assertion `isUInt<32>(Imm) || Imm == 0x3fc45f306dc9c882' failed.
Differential Revision: https://reviews.llvm.org/D128435
The new resumable mca::Pipeline capability introduced in this patch
allows users to save the current state of pipeline and resume from the
very checkpoint.
It is better (but not require) to use with the new IncrementalSourceMgr,
where users can add mca::Instruction incrementally rather than having a
fixed number of instructions ahead-of-time.
Note that we're using unit tests to test these new features. Because
integrating them into the `llvm-mca` tool will make too many churns.
Differential Revision: https://reviews.llvm.org/D127083
VOPD is a new encoding for dual-issue instructions for use in wave32.
This patch includes MC layer support only.
A VOPD instruction is constituted of an X component (for which there are
13 possible opcodes) and a Y component (for which there are the 13 X
opcodes plus 3 more). Most of the complexity in defining and parsing
a VOPD operation arises from the possible different total numbers of
operands and deferred parsing of certain operands depending on the
constituent X and Y opcodes.
Reviewed By: dp
Differential Revision: https://reviews.llvm.org/D128218
waitcnt vmcnt instructions are currently generated in loop bodies before using
values loaded outside of the loop. In some cases, it is better to flush the
vmcnt counter in a loop preheader before entering the loop body. This patch
detects these cases and generates waitcnt instructions to flush the counter.
Reviewed By: foad
Differential Revision: https://reviews.llvm.org/D115747
In GFX11 ShaderType is determined by the hardware and should no longer
be written into bits[3:2] of the ds_ordered_count offset field.
Differential Revision: https://reviews.llvm.org/D128196
Running iwyu-diff on LLVM codebase since fb67d683db detected a few
regressions, fixing them.
The impact on preprocessed output is negligible: -4k lines.
The granularity of SPI_SHADER_PGM_RSRC2_PS.EXTRA_LDS_SIZE changed
in GFX11. It is now in units of 256 dwords instead of 128 dwords.
COMPUTE_PGM_RSRC2.LDS_SIZE is unaffected. It is still in units of
128 dwords.
Differential Revision: https://reviews.llvm.org/D128179
The instructions that generate the source of dual source blend export
should run in strict-wqm. That is if any lane in a quad is active,
we need to enable all four lanes of that quad to make the shuffling
operation before exporting to dual source blend target work correctly.
Differential Revision: https://reviews.llvm.org/D127981
LDS_PARAM_LOAD and LDS_DIRECT_LOAD use EXEC per quad
(if any pixel is enabled in the quad, data is written
to all 4 pixels/threads in the quad).
Tag LDS_PARAM_LOAD and LDS_DIRECT_LOAD as using strict_wqm
to enforce this and avoid lane clobbering issues.
Note that only the instruction itself is tagged.
The implicit uses of these do not need to be set WQM.
The reduces unnecessary WQM calculation of M0.
Differential Revision: https://reviews.llvm.org/D127977
Detect LDS direct WAR/WAW hazards and compute values for
wait_vdst (va_vdst) parameter. Where appropriate this
raises wait_vdst from the default 0 to allow concurrent
issue of LDS direct with VALU execution.
Also detect LDS direct versus VMEM source VGPR hazards
and insert vm_vsrc=0 waits using s_waitcnt_depctr.
Differential Revision: https://reviews.llvm.org/D127963
`llvm::max(Align, MaybeAlign)` and `llvm::max(MaybeAlign, Align)` are
not used often enough to be required. They also make the code more opaque.
Differential Revision: https://reviews.llvm.org/D128121
This is a temporary measure to avoid generating incorrect code until the
compiler understands the new way that GFX11 encodes 16-bit operands in
VOP instructions.
Differential Revision: https://reviews.llvm.org/D128054
This includes:
- New llvm.amdgcn.image.msaa.load.* intrinsics
- NSA changes, because MIMG-NSA is now limited to 3 dwords
- Split CD forms of IMAGE_SAMPLE instructions out into separate
test files since they are no longer supported in GFX11
Differential Revision: https://reviews.llvm.org/D127837
Pre gfx1030 null for sdst is different.
c97436f8b6 [AMDGPU] Use null for dead sdst operand - requires a change to make
it not apply to pre gfx1030
Differential Revision: https://reviews.llvm.org/D127869
The sched_barrier builtin allow the scheduler's behavior to be shaped by users
when very specific codegen is needed in order to create highly optimized code.
This patch adds more granular control over the types of instructions that are
allowed to be reordered with respect to one or multiple sched_barriers. A mask
is used to specify groups of instructions that should be allowed to be scheduled
around a sched_barrier. The details about this mask may be used can be found in
llvm/include/llvm/IR/IntrinsicsAMDGPU.td.
Reviewed By: rampitec
Differential Revision: https://reviews.llvm.org/D127123
On gfx10+ null register can be used as both 32 and 64 bit operand.
Define a 64 bit version of the register to use during codegen.
Differential Revision: https://reviews.llvm.org/D127527
Compared to permlane16, permlane64 has no BC input because it has no
boundary conditions, no fi input because the instruction acts as if FI
were always enabled, and no OLD input because it always writes to every
active lane.
Also use the new intrinsic in the atomic optimizer pass.
Differential Revision: https://reviews.llvm.org/D127662
GFX11 uses different pseudos for these because of a new constraint
on which operands' registers can overlap.
Differential Revision: https://reviews.llvm.org/D127659
This uses rotating reminder of division by 3 to select another
temp vgpr each next time in a sequence of several agpr copies.
Therefore, temp vgpr selection depends on the generated agpr
number. This number could change with any unrelated change to
the register definitions.
Stabilize the selection by using a real agpr number.
Differential Revision: https://reviews.llvm.org/D127524
The encoding of COMPUTE_TMPRING_SIZE.WAVESIZE and
SPI_TMPRING_SIZE.WAVESIZE has changed in GFX11: it is now in units
of 64 dwords instead of 256 dwords, and the field has been widened
from 13 bits to 15 bits.
Depends on D126989
Reviewed By: rampitec, arsenm, #amdgpu
Differential Revision: https://reviews.llvm.org/D127248
sources to SALU and VALU instructions.
Contributors:
Baptiste Saleil <baptiste.saleil@amd.com>
Patch 20/N for upstreaming of AMDGPU gfx11 architecture
Depends on D126989
Reviewed By: rampitec, foad, #amdgpu
Differential Revision: https://reviews.llvm.org/D127143
Add a basic implementation of isExtractSubvectorCheap that only
considers extracts at offset 0.
Differential Revision: https://reviews.llvm.org/D127385
Add new intrinsic and codegen support for the s_sendmsg_rtn_b32 and
s_sendmsg_rtn_b64 instructions.
Differential Revision: https://reviews.llvm.org/D127315
In GFX10 dlc controlled L1 cache bypass. In GFX11 it has been repurposed
to control MALL NOALLOC, and glc controls L1 as well as L0 cache bypass.
Update the documentation and SIMemoryLegalizer accordingly. Set dlc for
nontemporal and volatile accesses.
Differential Revision: https://reviews.llvm.org/D127405
Changes for GFX11:
- Clauses may not mix instructions of different types, and there are
more types. For example image instructions with and without a sampler
are now different types.
- The max size of a clause is explicitly documented as 63 instructions.
Previously it was implicitly assumed to be 64. This is such a tiny
difference that it does not seem worth making it conditional on the
subtarget.
- It can be beneficial to clause stores as well as loads.
Differential Revision: https://reviews.llvm.org/D127391
Supports encoding existing instrutions on gfx11 and MC support for the new VOPC
dpp instructions.
Patch 19/N for upstreaming of AMDGPU gfx11 architecture
Depends on D126978
Reviewed By: rampitec, #amdgpu
Differential Revision: https://reviews.llvm.org/D126989
Nic Curtis done the experiments to prove it is faster than a
separate mul and add.
Fixes: SWDEV-332806
Differential Revision: https://reviews.llvm.org/D127253
- VOP3 and SDWA forms of V_CMPX were not handled
- Hazard only exists if the compare defines EXEC (i.e. V_CMPX)
forwarded to the permlane.
Differential Revision: https://reviews.llvm.org/D127344
Clang-format InstructionSimplify and convert all "FunctionName"s to
"functionName". This patch does touch a lot of files but gets done with
the cleanup of InstructionSimplify in one commit.
This is the alternative to the less invasive clang-format only patch: D126783
Reviewed By: spatel, rengolin
Differential Revision: https://reviews.llvm.org/D126889
The generic legalizer framework is still used to reduce the problem
to scalar multiplication with the bit size a multiple of 32.
Generating optimal code sequences for big integer multiplication is
somewhat tricky and has a number of target-specific intricacies:
- The target has V_MAD_U64_U32 instructions that multiply two 32-bit
factors and add a 64-bit accumulator. Most partial products should
use this instruction.
- The accumulator is mapped to consecutive 32-bit GPRs, and partial-
product multiply-adds can feed the accumulator into each other
directly. (The register allocator's support for that is somewhat
limited, but that only matters for 128-bit integers and larger.)
- OTOH, on some hardware, V_MAD_U64_U32 requires the accumulator
to be stored in an even-aligned pair of GPRs. To avoid excessive
register copies, it makes sense to compute odd partial products
separately from even partial products (where a partial product
src0[j0] * src1[j1] is "odd" if j0 + j1 is odd) and add both
halves together as a final step.
- We can combine G_MUL+G_ADD into a single cascade of multiply-adds.
- The target can keep many carry-bits in flight simultaneously, so
combining carries using G_UADDE is preferable over G_ZEXT + G_ADD.
- Not addressed by this patch: When the factors are sign-extended,
the V_MAD_I64_I32 instruction (signed version!) can be used.
It is difficult to address these points generically:
1) Finding matching pairs of G_MUL and G_UMULH to find a wide
multiply is expensive. We could add a G_UMUL_LOHI generic instruction
and conditionally use that in the generic legalizer, but by itself
this wouldn't allow us to use the accumulation capability of
V_MAD_U64_U32. One could attempt to find matching G_ADD + G_UADDE
post-legalization, but this is also expensive.
2) Similarly, making sense of the legalization outcome of a wide
pre-legalization G_MUL+G_ADD pair is extremely expensive.
3) How could the generic legalizer possibly deal with the
particular idiosyncracy of "odd" vs. "even" partial products.
All this points in the direction of directly emitting an ideal code
sequence during legalization, but the generic legalizer should not
be burdened with such overly target-specific concerns. Hence, a
custom legalization.
Note that the implemented approach is different from that used by
SelectionDAG because narrowing of scalars works differently in
general. SelectionDAG iteratively cuts wide scalars into low and
high halves until a legal size is reached. By contrast, GlobalISel
does the narrowing in a single shot, which should be better for
compile-time and for the quality of the generated code.
This patch leaves three gaps open:
1. When the factors are uniform, we should execute the multiplication on
the SALU. Register bank mapping already ensures this.
However, the resulting code sequence is not optimal because it doesn't
fully use the carry-in capabilities of S_ADDC_U32. (V_MAD_U64_U32
doesn't have a carry-in.) It is very difficult to fix this after the
fact, so we should really use a different legalization sequence in
this case. Unfortunately, we don't have a divergence analysis and so
cannot make that choice.
(This only matters for 128-bit integers and larger.)
2. Avoid unnecessary multiplies when sources are known to be zero- or
sign-extended. The challenge is that the legalizer does not currently
have access to GISelKnownBits.
3. When the G_MUL is followed by a G_ADD, we should consider combining
the two instructions into a single multiply-add sequence, to utilize
the accumulator of V_MAD_U64_U32 fully. (Unless the multiply has
multiple uses and the implied duplication of the multiply is an
overall negative). However, this is also not true when the factors
are uniform: in that case, it is generally better to *not* combine
the two operations, so that the multiply can be done on the SALU.
Again, we don't have a divergence analysis available and so cannot
make an informed choice.
Differential Revision: https://reviews.llvm.org/D124844
Includes dpp versions of VOP3P instructions.
Patch 18/N for upstreaming of AMDGPU gfx11 architecture
Depends on D126917
Reviewed By: rampitec, #amdgpu
Differential Revision: https://reviews.llvm.org/D126978
The reverted dependent commit is now relanded, so reland this.
Includes dpp instructions and vop1/vop2 promoted to vop3
Patch 17/N for upstreaming of AMDGPU gfx11 architecture
Depends on D126483
Reviewed By: rampitec, #amdgpu
Differential Revision: https://reviews.llvm.org/D126917
There was an issue with encoding wide (>64 bit) instructions on
BigEndian hosts, which is fixed in D127195. Therefore reland this.
gfx11 adds the ability to use dpp modifiers on vop3 instructions.
This patch adds machine code layer support for that. The MCCodeEmitter
is changed to use APInt instead of uint64_t to support these wider
instructions.
Patch 16/N for upstreaming of AMDGPU gfx11 architecture
Differential Revision: https://reviews.llvm.org/D126483
MIR support is totally unusable for AMDGPU without this, since the set
of reserved registers is set from fields here.
Add a clone method to MachineFunctionInfo. This is a subtle variant of
the copy constructor that is required if there are any MIR constructs
that use pointers. Specifically, at minimum fields that reference
MachineBasicBlocks or the MachineFunction need to be adjusted to the
values in the new function.
I can't remove the function just yet as it is used in the generated .inc files.
I would also like to provide a way to compare alignment with TypeSize since it came up a few times.
Differential Revision: https://reviews.llvm.org/D126910
Austin Kerbow