The code to match and produce more x86 vector blends was enabled for all
architectures even though the transform may pessimize the code for other
architectures that do not provide a vector blend instruction.
Added an aarch64 testcase to check that a VZIP instruction is generated instead
of byte movs.
Differential Revision: https://reviews.llvm.org/D44118
llvm-svn: 327132
Previously we unpacked the even bytes of each input into the high byte of 16-bit elements then did an v8i16 arithmetic shift right by 8 bits to fill the upper bits of each word with sign bits. Then we did the v8i16 multiply and then masked to zero the upper 8-bits of each result. The similar was done for all the odd bytes. The results are then packed together with packuswb
Since we are masking each multiply result element to 8-bits, and those 8-bits are determined only by the lower 8-bits of each of the inputs, we don't need to fill the upper bits with sign bits. So we can just unpack into the low byte of each element and treat the upper bits as garbage. This is what gcc also does.
Differential Revision: https://reviews.llvm.org/D44267
llvm-svn: 327093
This instruction can be thought of as reading either the even elements of a vXi32 input or the lower half of each element of a vXi64 input. We currently use the vXi32 interpretation, but vXi64 matches better with its broadcast behavior in EVEX.
I'm looking at moving MULDQ/MULUDQ creation to a DAG combine so we can do it when AVX512DQ is enabled without having to go through Custom lowering. But in some of the test cases we failed to use a broadcast load due to the size difference. This should help with that.
I'm also wondering if we can model these instructions in native IR and remove the intrinsics and I think using a vXi64 type will work better with that.
llvm-svn: 326991
These patterns weren't checking the alignment of the load, but were using the aligned instructions. This will cause a GP fault if the data isn't aligned.
I believe these were introduced in r312450.
llvm-svn: 326967
The v8i32 conversion on AVX1 targets was only working after LowerMUL splits 256-bit vectors.
While I was there I've also made it so we don't have to check for AVX2 and BWI directly and instead just ask if the type is legal.
Differential Revision: https://reviews.llvm.org/D44190
llvm-svn: 326917
Summary:
Only IMUL16rri uses an extra P0156. IMUL32* and IMUL16rr only use
P1.
This was computed using https://github.com/google/EXEgesis/blob/master/exegesis/tools/compute_itineraries.cc
This can easily be validated by running perf on the following code:
```
int main(int argc, char**argv) {
int a = argc;
int b = argc;
int c = argc;
int d = argc;
for (int i = 0; i < LOOP_ITERATIONS; ++i) {
asm volatile(
R"(
.rept 10000
imull $0x2, %%edx, %%eax
imull $0x2, %%ecx, %%ebx
imull $0x2, %%eax, %%edx
imull $0x2, %%ebx, %%ecx
.endr
)"
: "+a"(a), "+b"(b), "+c"(c), "+d"(d)
:
:);
}
return a+b+c+d;
}
```
-> test.cc
perf stat -x, -e cycles --pfm-events=uops_executed_port:port_0:u,uops_executed_port:port_1:u,uops_executed_port:port_2:u,uops_executed_port:port_3:u,uops_executed_port:port_4:u,uops_executed_port:port_5:u,uops_executed_port:port_6:u,uops_executed_port:port_7:u test
Reviewers: craig.topper, RKSimon, gadi.haber
Subscribers: llvm-commits, gchatelet, chandlerc
Differential Revision: https://reviews.llvm.org/D43460
llvm-svn: 326877
The code checks Level == AfterLegalizeDAG which is the fourth and last of the possible DAG combine stages that we have.
There is a Level called AfterLegalVectorOps, but that's the third DAG combine and it doesn't always run.
A function called isAfterLegalVectorOps should imply it returns true in either of the DAG combines that runs after the legalize vector ops stage, but that's not what this function does.
llvm-svn: 326832
EAX can turn out to be alive here, when shrink wrapping is done
(which is allowed when using dwarf exceptions, contrary to the
normal case with WinCFI).
This fixes PR36487.
Differential Revision: https://reviews.llvm.org/D43968
llvm-svn: 326764
Almost none of these usages were FP specific. And we had no clear guideliness on when to use hasAVX vs hasFP256.
I might also remove hasInt256 too since its an alias for hasAVX2.
llvm-svn: 326682
rL322525 - mmx zero constant support
rL322553 - mmx i32 zero extended value
rL326497 - mmx i64 general constant handling
Not all constants are folded, we generate some on the GPRs (similar to SSE build vector) where appropriate
llvm-svn: 326673
We were previously doing this with isel patterns. Moving it to op legalization gives us chance to see the required bitcast earlier. And it lets us remove some isel patterns.
llvm-svn: 326669
These instructions are double-pumped, split into 2 128-bit ops and then passing through either FPU pipe.
Found while testing llvm-mca (D43951)
llvm-svn: 326597
64-bit MMX constant generation usually ends up lowering into SSE instructions before being spilled/reloaded as a MMX type.
This patch bitcasts the constant to a double value to allow correct loading directly to the MMX register.
I've added MMX constant asm comment support to improve testing, it's better to always print the double values as hex constants as MMX is mainly an integer unit (and even with 3DNow! its just floats).
Differential Revision: https://reviews.llvm.org/D43616
llvm-svn: 326497
Emulated TLS is enabled by llc flag -emulated-tls,
which is passed by clang driver.
When llc is called explicitly or from other drivers like LTO,
missing -emulated-tls flag would generate wrong TLS code for targets
that supports only this mode.
Now use useEmulatedTLS() instead of Options.EmulatedTLS to decide whether
emulated TLS code should be generated.
Unit tests are modified to run with and without the -emulated-tls flag.
Differential Revision: https://reviews.llvm.org/D42999
llvm-svn: 326341
An extract_element where the result type is larger than the scalar element type is semantically an any_extend of from the scalar element type to the result type. If we expect zeroes in the upper bits of the i8/i32 we need to mae sure those zeroes are explicit in the DAG.
For these cases the best way to accomplish this is use an insert_subvector to pad zeroes to the upper bits of the v1i1 first. We extend to either v16i1(for i32) or v8i1(for i8). Then bitcast that to a scalar and finish with a zero_extend up to i32 if necessary. We can't extend past v16i1 because that's the largest mask size on KNL. But isel is smarter enough to know that a zext of a bitcast from v16i1 to i16 can use a KMOVW instruction. The insert_subvectors will be dropped during isel because we can determine that the producing instruction already zeroed the upper bits of the k-register.
llvm-svn: 326308
While the description for the instruction does mention OR, its talking about how the individual classification test results are ORed together.
The incoming mask is used as a zeroing write mask. If the bit is 1 the classification is written to the output. The bit is 0 the output is 0. This equivalent to an AND.
Here is pseudocode from the intrinsics guide
FOR j := 0 to 1
i := j*64
IF k1[j]
k[j] := CheckFPClass_FP64(a[i+63:i], imm8[7:0])
ELSE
k[j] := 0
FI
ENDFOR
k[MAX:2] := 0
llvm-svn: 326306
These tables add 3000 lines to X86InstrInfo.cpp. And if we ever manage to auto generate them they'll be a separate file anyway.
Differential Revision: https://reviews.llvm.org/D43806
llvm-svn: 326225
Currently we assert that only non target specific opcodes can have
missing RegisterClass constraints in the MCDesc. The backend can have
instructions with register operands but don't have RegisterClass
constraints (say using unknown_class) in which case the instruction
defining the register will constrain it.
Change the assert to only fire if a def has no regclass.
https://reviews.llvm.org/D43409
llvm-svn: 326142
Agner's tables indicate that for SSE42+ targets (Core2 and later) we can reduce the FADD/FSUB/FMUL costs down to 1, which should fix the Himeno benchmark.
Note: the AVX512 FDIV costs look rather dodgy, but this isn't part of this patch.
Differential Revision: https://reviews.llvm.org/D43733
llvm-svn: 326133
There's still some shortcoming in our ability to combine binops of constants with different sizes separated by an extend. I'll try to look at that next.
llvm-svn: 326128
Summary:
We have an early DAG combine to turn these patterns into MOVMSK, but that combine doesn't work if the vXi1 type has more elements than the widest legal vXi8 type. Type legalization will eventually split it down to v16i1 or v32i1 and then the bitcast gets legalized to a truncstore and a scalar load. The truncstore will get lowered to a series of extracts and bit math.
This patch adds a custom legalization to use a sign extend and MOVMSK instead. This prevents the eventual scalarization.
Reviewers: spatel, RKSimon, zvi
Reviewed By: RKSimon
Subscribers: mgorny, llvm-commits
Differential Revision: https://reviews.llvm.org/D43593
llvm-svn: 326119