This patch expands the support of lowerInterleavedload to {8|16|32}x8i stride 3.
LLVM creates suboptimal shuffle code-gen for AVX2. In overall, this patch is a specific fix for the pattern (Strid=3 VF={8|16|32}) and we plan to include the store (deinterleved side).
The patch goal is to optimize the following sequence:
a0 b0 c0 a1 b1 c1 a2 b2
c2 a3 b3 c3 a4 b4 c4 a5
b5 c5 a6 b6 c6 a7 b7 c7
into
a0 a1 a2 a3 a4 a5 a6 a7
b0 b1 b2 b3 b4 b5 b6 b7
c0 c1 c2 c3 c4 c5 c6 c7
Reviewers
1. zvi
2. igor
3. guyblank
4. dorit
5. Ayal
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@312722 91177308-0d34-0410-b5e6-96231b3b80d8
This patch expands the support of lowerInterleavedStore to 16x8i stride 4.
LLVM creates suboptimal shuffle code-gen for AVX2. In overall, this patch is a specific fix for the pattern (Strid=4 VF=16) and we plan to include more patterns in the future.
The patch goal is to optimize the following sequence:
At the end of the computation, we have ymm2, ymm0, ymm12 and ymm3 holding
each 16 chars:
c0, c1, , c16
m0, m1, , m16
y0, y1, , y16
k0, k1, ., k16
And these need to be transposed/interleaved and stored like so:
c0 m0 y0 k0 c1 m1 y1 k1 c2 m2 y2 k2 c3 m3 y3 k3 ....
Differential Revision: https://reviews.llvm.org/D35829
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@310252 91177308-0d34-0410-b5e6-96231b3b80d8
This patch expands the support of lowerInterleavedStore to 32x8i stride 4.
LLVM creates suboptimal shuffle code-gen for AVX2. In overall, this patch is a specific fix for the pattern (Strid=4 VF=32) and we plan to include more patterns in the future. To reach our goal of "more patterns". We include two mask creators. The first function creates shuffle's mask equivalent to unpacklo/unpackhi instructions. The other creator creates mask equivalent to a concat of two half vectors(high/low).
The patch goal is to optimize the following sequence:
At the end of the computation, we have ymm2, ymm0, ymm12 and ymm3 holding
each 32 chars:
c0, c1, , c31
m0, m1, , m31
y0, y1, , y31
k0, k1, ., k31
And these need to be transposed/interleaved and stored like so:
c0 m0 y0 k0 c1 m1 y1 k1 c2 m2 y2 k2 c3 m3 y3 k3 ....
Reviewers:
dorit
Farhana
RKSimon
guyblank
DavidKreitzer
Differential Revision: https://reviews.llvm.org/D34601
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@309086 91177308-0d34-0410-b5e6-96231b3b80d8
r306334 fixed a bug in AArch64 dealing with wide interleaved accesses having
pointer types. The bug also exists in ARM, so this patch copies over the fix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@307409 91177308-0d34-0410-b5e6-96231b3b80d8
This patch refactors and strengthens the type checks performed for interleaved
accesses. The primary functional change is to ensure that the interleaved
accesses have valid element types. The added test cases previously failed
because the element type is f128.
Differential Revision: https://reviews.llvm.org/D31817
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@299864 91177308-0d34-0410-b5e6-96231b3b80d8
This patch teaches (ARM|AArch64)ISelLowering.cpp to match illegal vector types
to interleaved access intrinsics as long as the types are multiples of the
vector register width. A "wide" access will now be mapped to multiple
interleave intrinsics similar to the way in which non-interleaved accesses with
illegal types are legalized into multiple accesses. I'll update the associated
TTI costs (in getInterleavedMemoryOpCost) as a follow-on.
Differential Revision: https://reviews.llvm.org/D29466
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@296750 91177308-0d34-0410-b5e6-96231b3b80d8
There are no vldN/vstN f16 variants, even with +fullfp16.
We could use the i16 variants, but, in practice, even with +fullfp16,
the f16 sequence leading to the i16 shuffle usually gets scalarized.
We'd need to improve our support for f16 codegen before getting there.
Reject f16 interleaved accesses. If we try to emit the f16 intrinsics,
we'll just end up with a selection failure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@294818 91177308-0d34-0410-b5e6-96231b3b80d8
The interleaved access pass is an IR-to-IR transformation that runs before code
generation. It matches interleaved memory operations to target-specific
intrinsics (that are later lowered to load and store multiple instructions on
ARM/AArch64). We place tests for similar passes (e.g., GlobalMergePass) under
test/Transforms. This patch moves the InterleavedAccessPass tests out of
test/CodeGen and into target-specific directories under
test/Transforms/InterleavedAccess.
Although the pass is an IR pass, many of the existing tests were llc tests
rather opt tests. For example, the tests would check for ldN/stN instructions
generated by llc rather than the intrinsic calls the pass actually inserts.
Thus, this patch updates all tests to be opt tests that check for the inserted
intrinsics. We already have separate CodeGen tests that ensure we lower the
interleaved access intrinsics to their corresponding ldN/stN instructions. In
addition to migrating the tests to opt, this patch also performs some minor
clean-up (to ensure consistent naming, etc.).
Differential Revision: https://reviews.llvm.org/D29184
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@293309 91177308-0d34-0410-b5e6-96231b3b80d8
X86. The pass optimizes as a unit the entire wide load + shuffles pattern
produced by interleaved vectorization. This initial patch optimizes one pattern
(64-bit elements interleaved by a factor of 4). Future patches will generalize
to additional patterns.
Patch by Farhana Aleen
Differential revision: http://reviews.llvm.org/D24681
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@284260 91177308-0d34-0410-b5e6-96231b3b80d8
