Identifies kernels which performs device side kernel enqueues and emit
metadata for the associated hidden kernel arguments. Such kernels are
marked with calls-enqueue-kernel function attribute by
AMDGPUOpenCLEnqueueKernelLowering pass and later on
hidden kernel arguments metadata HiddenDefaultQueue and
HiddenCompletionAction are emitted for them.
Differential Revision: https://reviews.llvm.org/D39255
llvm-svn: 316907
Summary:
Kill the thread if operand 0 == false.
llvm.amdgcn.wqm.vote can be applied to the operand.
Also allow kill in all shader stages.
Reviewers: arsenm, nhaehnle
Subscribers: kzhuravl, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D38544
llvm-svn: 316427
The range should be assumed to be the hardware maximum
if a workitem intrinsic is used in a callable function
which does not know the restricted limit of the calling
kernel.
llvm-svn: 316346
Fixes cfe/trunk/test/Misc/backend-resource-limit-diagnostics.cl
test after r315808
We may hit few other similar issues, but I want to discuss good
solution offline.
llvm-svn: 315830
- Do not allow amd_amdgpu_isa directives on non-amdgcn architectures
- Do not allow amd_amdgpu_hsa_metadata on non-amdhsa OSes
- Do not allow amd_amdgpu_pal_metadata on non-amdpal OSes
Differential Revision: https://reviews.llvm.org/D38750
llvm-svn: 315812
- Emit NT_AMD_AMDGPU_ISA
- Add assembler parsing for isa version directive
- If isa version directive does not match command line arguments, then return error
Differential Revision: https://reviews.llvm.org/D38748
llvm-svn: 315808
Reverting to investigate layering effects of MCJIT not linking
libCodeGen but using TargetMachine::getNameWithPrefix() breaking the
lldb bots.
This reverts commit r315633.
llvm-svn: 315637
Merge LLVMTargetMachine into TargetMachine.
- There is no in-tree target anymore that just implements TargetMachine
but not LLVMTargetMachine.
- It should still be possible to stub out all the various functions in
case a target does not want to use lib/CodeGen
- This simplifies the code and avoids methods ending up in the wrong
interface.
Differential Revision: https://reviews.llvm.org/D38489
llvm-svn: 315633
Summary:
The interpolation mode workaround ensures that at least one
interpolation mode is enabled in PSInputAddr. It does not also check
PSInputEna on the basis that the user might enable bits in that
depending on run-time state.
However, for amdpal os type, the user does not enable some bits after
compilation based on run-time states; the register values being
generated here are the final ones set in the hardware. Therefore, apply
the workaround to PSInputAddr and PSInputEnable together. (The case
where a bit is set in PSInputAddr but not in PSInputEnable is where the
frontend set up an input arg for a particular interpolation mode, but
nothing uses that input arg. Really we should have an earlier pass that
removes such an arg.)
Reviewers: arsenm, nhaehnle, dstuttard
Subscribers: kzhuravl, wdng, yaxunl, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D37758
llvm-svn: 315591
Summary:
Add LLVM_FORCE_ENABLE_DUMP cmake option, and use it along with
LLVM_ENABLE_ASSERTIONS to set LLVM_ENABLE_DUMP.
Remove NDEBUG and only use LLVM_ENABLE_DUMP to enable dump methods.
Move definition of LLVM_ENABLE_DUMP from config.h to llvm-config.h so
it'll be picked up by public headers.
Differential Revision: https://reviews.llvm.org/D38406
llvm-svn: 315590
MCObjectStreamer owns its MCCodeEmitter -- this fixes the types to reflect that,
and allows us to remove the last instance of MCObjectStreamer's weird "holding
ownership via someone else's reference" trick.
llvm-svn: 315531
- Use HSA metadata streamer directly from AMDGPUAsmPrinter
- Make naming consistent with PAL metadata
Differential Revision: https://reviews.llvm.org/D38746
llvm-svn: 315526
- Move PAL metadata definitions to AMDGPUMetadata
- Make naming consistent with HSA metadata
Differential Revision: https://reviews.llvm.org/D38745
llvm-svn: 315523
- Rename AMDGPUCodeObjectMetadata to AMDGPUMetadata (PAL metadata will be included in this file in the follow up change)
- Rename AMDGPUCodeObjectMetadataStreamer to AMDGPUHSAMetadataStreamer
- Introduce HSAMD namespace
- Other minor name changes in function and test names
llvm-svn: 315522
This adds debug tracing to the table-generated assembly instruction matcher,
enabled by the -debug-only=asm-matcher option.
The changes in the target AsmParsers are to add an MCInstrInfo reference under
a consistent name, so that we can use it from table-generated code. This was
already being used this way for targets that use deprecation warnings, but 5
targets did not have it, and Hexagon had it under a different name to the other
backends.
llvm-svn: 315445
MCObjectStreamer owns its MCAsmBackend -- this fixes the types to reflect that,
and allows us to remove another instance of MCObjectStreamer's weird "holding
ownership via someone else's reference" trick.
llvm-svn: 315410
opt-bisect/optnone disable the AMDGPUUniformAnnotateValues pass.
The heuristic in the custom selector for brcond deferred the
branch uniformity check to the pattern, which would fail.
llvm-svn: 315360
This patch adds a post-linking pass which replaces the function pointer of enqueued
block kernel with a global variable (runtime handle) and adds
runtime-handle attribute to the enqueued block kernel.
In LLVM CodeGen the runtime-handle metadata will be translated to
RuntimeHandle metadata in code object. Runtime allocates a global buffer
for each kernel with RuntimeHandel metadata and saves the kernel address
required for the AQL packet into the buffer. __enqueue_kernel function
in device library knows that the invoke function pointer in the block
literal is actually runtime handle and loads the kernel address from it
and puts it into AQL packet for dispatching.
This cannot be done in FE since FE cannot create a unique global variable
with external linkage across LLVM modules. The global variable with internal
linkage does not work since optimization passes will try to replace loads
of the global variable with its initialization value.
Differential Revision: https://reviews.llvm.org/D38610
llvm-svn: 315352
functions.
This makes the ownership of the resulting MCObjectWriter clear, and allows us
to remove one instance of MCObjectStreamer's bizarre "holding ownership via
someone else's reference" trick.
llvm-svn: 315327
Summary:
Atomic buffer operations do not work (and trap on gfx9) when the
components are unaligned, even if their sum is aligned.
Previously, we generated an offset of 4156 without an SGPR by
splitting it as 4095 + 61 (immediate + inline constant). The
highest offset for which we can do this correctly is 4156 = 4092 + 64.
Fixes dEQP-GLES31.functional.ssbo.atomic.*
Reviewers: arsenm
Subscribers: kzhuravl, wdng, yaxunl, dstuttard, tpr, llvm-commits, t-tye
Differential Revision: https://reviews.llvm.org/D37850
llvm-svn: 315302
ELFObjectWriter's constructor.
Fixes the same ownership issue for ELF that r315245 did for MachO:
ELFObjectWriter takes ownership of its MCELFObjectTargetWriter, so we want to
pass this through to the constructor via a unique_ptr, rather than a raw ptr.
llvm-svn: 315254
Old expansion was 20 VGPRs, 78 SGPRs and ~380 instructions.
This expansion is 11 VGPRs, 12 SGPRs and ~120 instructions.
Passes OpenCL conformance test_integer_ops quick_[u]long_math
Differential Revision: https://reviews.llvm.org/D38607
llvm-svn: 315081
This is a follow-up to https://reviews.llvm.org/D38138.
I fixed the capitalization of some functions because we're changing those
lines anyway and that helped verify that we weren't accidentally dropping
any options by using default param values.
llvm-svn: 314930
Summary:
For the amdpal OS type:
We write an AMDGPU_PAL_METADATA record in the .note section in the ELF
(or as an assembler directive). It contains key=value pairs of 32 bit
ints. It is a merge of metadata from codegen of the shaders, and
metadata provided by the frontend as _amdgpu_pal_metadata IR metadata.
Where both sources have a key=value with the same key, the two values
are ORed together.
This .note record is part of the amdpal ABI and will be documented in
docs/AMDGPUUsage.rst in a future commit.
Eventually the amdpal OS type will stop generating the .AMDGPU.config
section once the frontend has safely moved over to using the .note
records above instead of .AMDGPU.config.
Reviewers: arsenm, nhaehnle, dstuttard
Subscribers: kzhuravl, wdng, yaxunl, llvm-commits, t-tye
Differential Revision: https://reviews.llvm.org/D37753
llvm-svn: 314829
These are problematic because they apply to everything,
and can easily clobber whatever more specific predicate
you are trying to add to a function.
Currently instructions use SubtargetPredicate/PredicateControl
to apply this to patterns applied to an instruction definition,
but not to free standing Pats. Add a wrapper around Pat
so the special PredicateControls requirements can be appended
to the final predicate list like how Mips does it.
llvm-svn: 314742
We have a single library build without relaxation options.
When inlined library functions remove fast math attributes
from the functions they are integrated into.
This patch sets relaxation attributes on the functions after
linking provided corresponding relaxation options are given.
Math instructions inside the inlined functions remain to have
no fast flags, but inlining does not prevent fast math
transformations of a surrounding caller code anymore.
Differential Revision: https://reviews.llvm.org/D38325
llvm-svn: 314568
The hardware will only forward EXEC_LO; the high 32 bits will be zero.
Additionally, inline constants do not work. At least,
v_addc_u32_e64 v0, vcc, v0, v1, -1
which could conceivably be used to combine (v0 + v1 + 1) into a single
instruction, acts as if all carry-in bits are zero.
The llvm.amdgcn.ps.live test is adjusted; it would be nice to combine
s_mov_b64 s[0:1], exec
v_cndmask_b32_e64 v0, v1, v2, s[0:1]
into
v_mov_b32 v0, v3
but it's not particularly high priority.
Fixes dEQP-GLES31.functional.shaders.helper_invocation.value.*
llvm-svn: 314522
Implement shouldCoalesce() to help regalloc avoid running out of GR128
registers.
If a COPY involving a subreg of a GR128 is coalesced, the live range of the
GR128 virtual register will be extended. If this happens where there are
enough phys-reg clobbers present, regalloc will run out of registers (if
there is not a single GR128 allocatable register available).
This patch tries to allow coalescing only when it can prove that this will be
safe by checking the (local) interval in question.
Review: Ulrich Weigand, Quentin Colombet
https://reviews.llvm.org/D37899https://bugs.llvm.org/show_bug.cgi?id=34610
llvm-svn: 314516
Summary:
This commit adds comments on how the AMDPAL OS type overloads the
existing AMDGPU_ calling conventions used by Mesa, and adds a couple of
new ones.
Reviewers: arsenm, nhaehnle, dstuttard
Subscribers: mehdi_amini, kzhuravl, wdng, yaxunl, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D37752
llvm-svn: 314502
Summary:
Added support for scratch (including spilling) for OS type amdpal:
generates code to set up the scratch descriptor if it is needed.
With amdpal, the scratch resource descriptor is loaded from offset 0 of
the global information table. The low 32 bits of the address of the
global information table is passed in s0.
Added amdgpu-git-ptr-high function attribute to hard-wire the high 32
bits of the address of the global information table. If the function
attribute is not specified, or is 0xffffffff, then the backend generates
code to use the high 32 bits of pc.
The documentation for the AMDPAL ABI will be added in a later commit.
Subscribers: arsenm, kzhuravl, wdng, nhaehnle, yaxunl, dstuttard, t-tye
Differential Revision: https://reviews.llvm.org/D37483
llvm-svn: 314501
Summary:
This operating system type represents the AMDGPU PAL runtime, and will
be required by the AMDGPU backend in order to generate correct code for
this runtime.
Currently it generates the same code as not specifying an OS at all.
That will change in future commits.
Patch from Tim Corringham.
Subscribers: arsenm, nhaehnle
Differential Revision: https://reviews.llvm.org/D37380
llvm-svn: 314500
This was intended to be no-functional-change, but it's not - there's a test diff.
So I thought I should stop here and post it as-is to see if this looks like what was expected
based on the discussion in PR34603:
https://bugs.llvm.org/show_bug.cgi?id=34603
Notes:
1. The test improvement occurs because the existing 'LateSimplifyCFG' marker is not carried
through the recursive calls to 'SimplifyCFG()->SimplifyCFGOpt().run()->SimplifyCFG()'.
The parameter isn't passed down, so we pick up the default value from the function signature
after the first level. I assumed that was a bug, so I've passed 'Options' down in all of the
'SimplifyCFG' calls.
2. I split 'LateSimplifyCFG' into 2 bits: ConvertSwitchToLookupTable and KeepCanonicalLoops.
This would theoretically allow us to differentiate the transforms controlled by those params
independently.
3. We could stash the optional AssumptionCache pointer and 'LoopHeaders' pointer in the struct too.
I just stopped here to minimize the diffs.
4. Similarly, I stopped short of messing with the pass manager layer. I have another question that
could wait for the follow-up: why is the new pass manager creating the pass with LateSimplifyCFG
set to true no matter where in the pipeline it's creating SimplifyCFG passes?
// Create an early function pass manager to cleanup the output of the
// frontend.
EarlyFPM.addPass(SimplifyCFGPass());
-->
/// \brief Construct a pass with the default thresholds
/// and switch optimizations.
SimplifyCFGPass::SimplifyCFGPass()
: BonusInstThreshold(UserBonusInstThreshold),
LateSimplifyCFG(true) {} <-- switches get converted to lookup tables and loops may not be in canonical form
If this is unintended, then it's possible that the current behavior of dropping the 'LateSimplifyCFG'
setting via recursion was masking this bug.
Differential Revision: https://reviews.llvm.org/D38138
llvm-svn: 314308
We can have a v_mac with an immediate src0.
We can still fold if it's an inline immediate,
otherwise it already uses the constant bus.
llvm-svn: 313852
These write to the low and high half of the destination
register and leave the other 16-bits unchanged. This is true
for most 16-bit instructions on gfx9, but we don't use that
now.
llvm-svn: 313812
Also add some tests that should be able to use v_mad_mixhi_f16,
but do not yet. This is trickier because we don't really model
the partial update of the register done by 16-bit instructions.
llvm-svn: 313806
Also starts selecting global loads for constant address
in some cases. Some end up selecting to mubuf still, which
requires investigation.
We still get sub-optimal regalloc and extra waitcnts inserted
due to not really tracking the liveness of the separate register
halves.
llvm-svn: 313716
The pre-RA scheduler does load/store clustering, but post-RA
scheduler undoes it. Add mutation to prevent it.
Differential Revision: https://reviews.llvm.org/D38014
llvm-svn: 313670
Because the stack growth direction and addressing is done
in the same direction, modifying SP at the beginning of the
call sequence was incorrect. If we had a stack passed argument,
we would end up skipping that number of bytes before pushing
arguments, leaving unused/inconsistent space.
The callee creates fixed stack objects in its frame, so
the space necessary for these is already logically allocated
in the callee, so we just let the callee increment SP if
it really requires it.
llvm-svn: 313279
Using SplitCSR for the frame register was very broken. Often
the copies in the prolog and epilog were optimized out, in addition
to them being inserted after the true prolog where the FP
was clobbered.
I have a hacky solution which works that continues to use
split CSR, but for now this is simpler and will get to working
programs.
llvm-svn: 313274
MachineScheduler when clustering loads or stores checks if base
pointers point to the same memory. This check is done through
comparison of base registers of two memory instructions. This
works fine when instructions have separate offset operand. If
they require a full calculated pointer such instructions can
never be clustered according to such logic.
Changed shouldClusterMemOps to accept base registers as well and
let it decide what to do about it.
Differential Revision: https://reviews.llvm.org/D37698
llvm-svn: 313208
These two instructions are normally selected, but when the
two address pass converts mac into mad we end up with the
mad where we could have one of these.
Differential Revision: https://reviews.llvm.org/D37389
llvm-svn: 312928
A mrt exp with vm=1 must be in exact (non-WQM) mode, as it also exports
the exec mask as the valid mask to determine which pixels to render.
This commit marks any exp as needing to be in exact mode.
Actually, if there are multiple mrt exps, only one needs to have vm=1,
and only that one needs to be in exact mode. But that is an optimization
for another day.
Differential Revision: https://reviews.llvm.org/D36305
llvm-svn: 312915
We have a lot of operand definition work essentially producing
every valid permutation of operands to workaround builiding
operand lists based on the instruction features. Apparently tablegen
already has a mostly undocumented operator to concat dags which
simplies this.
Convert one simple place to use this. The BUF instruction definitions
have much more complicated logic that can be totally rewritten now.
llvm-svn: 312822
The various scalar bit operations set SCC,
so one is erased or moved it needs to be recomputed.
Not sure why the existing tests don't fail on this.
llvm-svn: 312819
Summary:
Mesa still uses a hack where empty inline assembly is used as a kind of
optimization barrier. This exposed a problem where not enough wait states
were inserted, because the hazard recognizer implicitly assumed that each
inline assembly "instruction" has at least one wait state.
Reviewers: arsenm
Subscribers: kzhuravl, wdng, yaxunl, dstuttard, tpr, llvm-commits, t-tye
Differential Revision: https://reviews.llvm.org/D37205
llvm-svn: 312635
When packet size equals packet align and is power of 2, transform
__read_pipe* and __write_pipe* to specialized library function.
Differential Revision: https://reviews.llvm.org/D36831
llvm-svn: 312598
- Make SIMemOpInfo a class
- Add accessor methods to SIMemOpInfo
- Move get*Info methods to SIMemOpInfo
Differential Revision: https://reviews.llvm.org/D37395
llvm-svn: 312541
- Rename MemOpInfo -> SIMemOpInfo
- Move SIMemOpInfo class out of SIMemoryLegalizer class
Differential Revision: https://reviews.llvm.org/D37394
llvm-svn: 312540
Summary:
This fixes a bug that was exposed on gfx9 in various
GL45-CTS.shaders.loops.*_iterations.select_iteration_count_fragment tests,
e.g. GL45-CTS.shaders.loops.do_while_uniform_iterations.select_iteration_count_fragment
Reviewers: arsenm
Subscribers: kzhuravl, wdng, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D36193
llvm-svn: 312337
build_vector is a more useful canonical form when
pattern matching packed operations, so turn shift
into high element into a build_vector.
Should show no change for now.
llvm-svn: 312282
The majority of the time spent in the pass checking
for the register reads. Rather than searching all of
the defined registers for uses in each instruction,
use a set of defined registers and check the operands
of the instruction.
This process still is algorithmically not great,
but with the additional trick of skipping the analysis
for addresses with one use, this brings one slow
testcase into a reasonable range.
llvm-svn: 312206
These aren't really packed instructions, so the default
op_sel_hi should be 0 since this indicates a conversion.
The operand types are scalar values that behave similar
to an f16 scalar that may be converted to f32.
Doesn't change the default printing for op_sel_hi, just
the parsing.
llvm-svn: 312179
The merge is only possible if the base address register is the
same for the two instructions. If there is only the one use,
there's no point in doing an expensive forward scan checking
for memory interference looking for a merge candidate.
This gives a signficant improvement in one extreme testcase.
The code to do the scan is still algorithmically terrible,
so this is still the slowest pass in that example.
llvm-svn: 312096
If denorms are not flushed we can use max instead of multiplication
by 1. For double that is simply faster, while for float and half
it is shorter, because mul uses constant bus and VOP3.
Differential Revision: https://reviews.llvm.org/D36856
llvm-svn: 312095
Under -cl-fast-relaxed-math we could use native_sqrt, but f64 was
allowed to produce HSAIL's nsqrt instruction. HSAIL is not here
and we stick with non-existing native_sqrt(double) as a result.
Add check for f64 to not return native functions and also remove
handling of f64 case for fold_sqrt.
Differential Revision: https://reviews.llvm.org/D37223
llvm-svn: 311900
Summary:
This is step towards separating the GCN and R600 tablegen'd code.
This is a little awkward for now, because the R600 functions won't have the
MCSubtargetInfo parameter, so we need to have AMDMGPUInstPrinter
delegate to R600InstPrinter, but once the tablegen'd code is split,
we will be able to drop the delegation and use R600InstPrinter directly.
Reviewers: arsenm
Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, llvm-commits
Differential Revision: https://reviews.llvm.org/D36444
llvm-svn: 311128
This reverts commit r310425, thus reapplying r310335 with a fix for link
issue of the AArch64 unittests on Linux bots when BUILD_SHARED_LIBS is ON.
Original commit message:
[GlobalISel] Remove the GISelAccessor API.
Its sole purpose was to avoid spreading around ifdefs related to
building global-isel. Since r309990, GlobalISel is not optional anymore,
thus, we can get rid of this mechanism all together.
NFC.
----
The fix for the link issue consists in adding the GlobalISel library in
the list of dependencies for the AArch64 unittests. This dependency
comes from the use of AArch64Subtarget that needs to know how
to destruct the GISel related APIs when being detroyed.
Thanks to Bill Seurer and Ahmed Bougacha for helping me reproducing and
understand the problem.
llvm-svn: 310969
The pass does simplifications of well known AMD library calls.
If given -amdgpu-prelink option it works in a pre-link mode which
allows to reference new library functions which will be linked in
later.
In addition it also used to process traditional AMD option
-fuse-native which allows to replace some of the functions with
their fast native implementations from the library.
The necessary glue to pass the prelink option and translate
-fuse-native is to be added to the driver.
Differential Revision: https://reviews.llvm.org/D36436
llvm-svn: 310731
Konstantin Zhuravlyov