When spilling a VGPR in `emitPrologue`, chain functions prefer to use
offsets to access the stack instead of the SP.
This patch fixes `emitEpilogue` to do the same. It also brings back some
test coverage that was lost in #93526, when WWM registers started being
shifted to the lowest available range (which meant that tests that were
originally spilling v8 would shift to spill v0, which is a scratch
register for chain functions and didn't get spilled).
Change-Id: Icb07fccd859b563cd45f74c25ae578ecb38bdeeb
In graphics shaders it is better overall to use NSA encoding for IMAGE
instructions, because the benefit of less constrained register
allocation outweighs the cost of larger encoding. In particular NSA form
often avoids the need for extra V_MOV_B32 instructions between IMAGE
instructions, which can allow the IMAGE instructions to be claused.
Note that in GFX12 there is no longer a bit in the encoding to choose
between NSA and non-NSA forms, so this only affects GFX10 and GFX11.
This is in preparation for adding a new optimization to the pass that
cares about the order of instructions. The existing optimization does
not care, so this just causes minor codegen differences.
into zext x
LegalizerHelper has two padding strategies: undef or zero.
see LegalizerHelper:273
see LegalizerHelper:315
This PR is about zero sugar and Coke Zero.
; CHECK-NEXT: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES %a(s32),
[[C]](s32)
Please continue padding merge values.
// %bits_8_15:(s8) = G_CONSTANT i8 0
// %0:(s16) = G_MERGE_VALUES %bits_0_7:(s8), %bits_8_15:(s8)
%bits_8_15 is defined by zero. For optimization, we pick zext.
// %0:_(s16) = G_ZEXT %bits_0_7:(s8)
The upper bits of %0 are zero and the lower bits come from %bits_0_7.
I was wrong last patch. I viewed the `Visited` set purely as a possible
recursion deterrent where functions calling a callee multiple times are
handled elsewhere. This doesn't consider cases where a function is
called multiple times by different callers still part of the same call
graph. New test shows the aforementioned case.
Reapplies #111004, fixes#115562.
Update VInterp instructions with true16 and fake16 formats.
This patch includes instructions:
v_interp_p10_f16_f32
v_interp_p2_f16_f32
v_interp_p10_rtz_f16_f32
v_interp_p2_rtz_f16_f32
dasm test vinterp-fake16.txt is removed and the testline are merged into
vinterp.txt which handles both true16/fake16 cases
It is not correct to lower "x&(-1>>(32-y))" to "bfe x, 0, y". When y
equals 32, "-1" is not shifted, so x&(-1>>(32-32) is still x, but "bfe
x, 0, 32" is 0. However, if we know y is at most of 5 bits (< 32), we
can still do the pattern matching.
Based on the suggestion from
https://github.com/llvm/llvm-project/pull/115543, we should not do the
pattern matching from x << (32-y) >> (32-y) to "bfe x, 0, y" at all.
This reverts commits a2bacf8ab58af4c1a0247026ea131443d6066602 and
bdf8e308b7.
This patch introduces a new generic target, `gfx9-4-generic`. Since it doesn’t support FP8 and XF32-related instructions, the patch includes several code reorganizations to accommodate these changes.
Currently all implicit-def instructions are part of
bb prolog. We should only include the wwm-register's
implicit definitions into the BB prolog. The other
vector class registers' implicit defs when exist at
the bb top might cause interference when pushed the
LR_split copy insertion downwards. The SplitKit is
very strict on altering the insertion points and will
assert such instances.
Only select to a VGPR if it's trivally used in VGPR only contexts.
This fixes mishandling frame indexes used in SGPR only contexts,
like inline assembly constraints.
This is suboptimal in the common case where the frame index
is transitively used by only VALU ops. We make up for this by later
folding the copy to VALU plus scalar op in SIFoldOperands.
The increase in fallbacks that was previously reported were not caused
by this change.
Original description:
This matches InstCombine and DAGCombine.
RISC-V only has an ADDI instruction so without this we need additional
patterns to do the conversion.
Some of the AMDGPU tests look like possible regressions. Maybe some
patterns from isel aren't imported.
We’re facing an issue (#113782) that is currently blocking #112403. However,
since #112403 involves extensive test changes, I’d prefer to land it as soon as
possible. This PR reorganizes the tests by moving test cases expected to fail
into a separate file. Additionally, it changes the `[15 x i32]` arguments to
`[13 x i32]` to bypass the issue.
Resource usage information would try to overwrite unnamed functions if
there are multiple within the same compilation unit. This aims to either
use the `MCSymbol` assigned to the unnamed function (i.e.,
`CurrentFnSym`), or, rematerialize the `MCSymbol` for the unnamed
function.
This reverts commit c31014322c0b5ae596da129cbb844fb2198b4ef4.
Based on the discussions in #112772, this pass is not needed after the
introduction of `llvm.threadlocal.address` intrinsic.
Fixes https://github.com/llvm/llvm-project/issues/112771.
Fixes a copy-paste typo.
The typo resulted in producing bad v_perm based operands for the v_dot4
combine. When adding a corresponding byte pair to the v_dot byte pair
chains, we must take note of the byte position in the corresponding
source nodes. These byte positions are used to ensure we extract the
correct DWord from the ultimate source, and formulate a correct
perm_mask from the extracted DWord.
With the typo, we the S0 byte would used the DWord offset for the
corresponding S1 byte. If this offset was not the same as the true DWord
offset for the S0 byte, we would extract and use the wrong byte for S0
in the v_dot.
Fixes https://github.com/llvm/llvm-project/issues/112941
Use a local pointer type to represent the named barrier in builtin and
intrinsic. This makes the definitions more user friendly
bacause they do not need to worry about the hardware ID assignment. Also
this approach is more like the other popular GPU programming language.
Named barriers should be represented as global variables of addrspace(3)
in LLVM-IR. Compiler assigns the special LDS offsets for those variables
during AMDGPULowerModuleLDS pass. Those addresses are converted to hw
barrier ID during instruction selection. The rest of the
instruction-selection changes are primarily due to the
intrinsic-definition changes.
