one-byte NOPs. If the processor actually executes those NOPs, as it sometimes
does with aligned bundling, this can have a performance impact. From my
micro-benchmarks run on my one machine, a 15-byte NOP followed by twelve
one-byte NOPs is about 20% worse than a 15 followed by a 12. This patch
changes NOP emission to emit as many 15-byte (the maximum) as possible followed
by at most one shorter NOP.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176464 91177308-0d34-0410-b5e6-96231b3b80d8
'R' An address that can be sued in a non-macro load or store.
This patch includes a positive test case.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176452 91177308-0d34-0410-b5e6-96231b3b80d8
* Only apply divide bypass optimization when not optimizing for size.
* Fixed bug caused by constant for 0 value of type Int32,
used dividend type to generate the constant instead.
* For atom x86-64 apply the divide bypass to use 16-bit divides instead of
64-bit divides when operand values are small enough.
* Added lit tests for 64-bit divide bypass.
Patch by Tyler Nowicki!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176442 91177308-0d34-0410-b5e6-96231b3b80d8
This matters for example in following matrix multiply:
int **mmult(int rows, int cols, int **m1, int **m2, int **m3) {
int i, j, k, val;
for (i=0; i<rows; i++) {
for (j=0; j<cols; j++) {
val = 0;
for (k=0; k<cols; k++) {
val += m1[i][k] * m2[k][j];
}
m3[i][j] = val;
}
}
return(m3);
}
Taken from the test-suite benchmark Shootout.
We estimate the cost of the multiply to be 2 while we generate 9 instructions
for it and end up being quite a bit slower than the scalar version (48% on my
machine).
Also, properly differentiate between avx1 and avx2. On avx-1 we still split the
vector into 2 128bits and handle the subvector muls like above with 9
instructions.
Only on avx-2 will we have a cost of 9 for v4i64.
I changed the test case in test/Transforms/LoopVectorize/X86/avx1.ll to use an
add instead of a mul because with a mul we now no longer vectorize. I did
verify that the mul would be indeed more expensive when vectorized with 3
kernels:
for (i ...)
r += a[i] * 3;
for (i ...)
m1[i] = m1[i] * 3; // This matches the test case in avx1.ll
and a matrix multiply.
In each case the vectorized version was considerably slower.
radar://13304919
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176403 91177308-0d34-0410-b5e6-96231b3b80d8
This patch eliminates the need to emit a constant move instruction when this
pattern is matched:
(select (setgt a, Constant), T, F)
The pattern above effectively turns into this:
(conditional-move (setlt a, Constant + 1), F, T)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176384 91177308-0d34-0410-b5e6-96231b3b80d8
- ISD::SHL/SRL/SRA must have either both scalar or both vector operands
but TLI.getShiftAmountTy() so far only return scalar type. As a
result, backend logic assuming that breaks.
- Rename the original TLI.getShiftAmountTy() to
TLI.getScalarShiftAmountTy() and re-define TLI.getShiftAmountTy() to
return target-specificed scalar type or the same vector type as the
1st operand.
- Fix most TICG logic assuming TLI.getShiftAmountTy() a simple scalar
type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176364 91177308-0d34-0410-b5e6-96231b3b80d8
dispatch code. As far as I can tell the thumb2 code is behaving as expected.
I was able to compile and run the associated test case for both arm and thumb1.
rdar://13066352
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176363 91177308-0d34-0410-b5e6-96231b3b80d8
v2: based on Michels patch, but now allows copying of all registers sizes.
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176346 91177308-0d34-0410-b5e6-96231b3b80d8
It's much easier to specify the encoding with tablegen directly.
Signed-off-by: Christian König <christian.koenig@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176344 91177308-0d34-0410-b5e6-96231b3b80d8
This function will be used later when the capability to search delay slot
filling instructions in successor blocks is added. No intended functionality
changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176325 91177308-0d34-0410-b5e6-96231b3b80d8
The work done by the post-encoder (setting architecturally unused bits to 0 as
required) can be done by the existing operand that covers the "#0.0". This
removes at least one use of the discouraged PostEncoderMethod uses.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176261 91177308-0d34-0410-b5e6-96231b3b80d8
If an otherwise weak var is actually defined in this unit, it can't be
undefined at runtime so we can use normal global variable sequences (ADRP/ADD)
to access it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176259 91177308-0d34-0410-b5e6-96231b3b80d8
This fixes an issue where trying to assemlbe valid ADR instructions would cause
LLVM to hit a failed assertion.
Patch by Keith Walker.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176189 91177308-0d34-0410-b5e6-96231b3b80d8
There's no need to generate a stack frame for PPC32 SVR4 when there are
no local variables assigned to the stack, i.e., when no red zone is needed.
(PPC64 supports a red zone, but PPC32 does not.)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176124 91177308-0d34-0410-b5e6-96231b3b80d8
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176105 91177308-0d34-0410-b5e6-96231b3b80d8
Make it possible to map between e32 and e64 encoding opcodes.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176104 91177308-0d34-0410-b5e6-96231b3b80d8
