steam-frame/FEX-Emu--FEX
1
0 Fork 0
mirror of https://github.com/FEX-Emu/FEX.git synced 2026-10-06 13:00:15 +02:00
Code Issues Packages Projects Releases 66 Wiki Activity
14,478 Commits 11 Branches 70 Tags
5d8d052a77c5e330fecf24e0ad79feb46a49a812
Commit Graph
14478 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Ryan Houdek 5d8d052a77 Merge pull request #5640 from lioncash/move 
VectorOps: Avoid move in 256-bit VAddP/VFAddP if possible
 2026-07-01 12:41:03 -07:00
Tony Wasserka 7fa4d78269 Merge pull request #5635 from Sonicadvance1/178 
FEXOfflineCompiler: Fixes HostFeature detection under Win32
 2026-07-01 17:15:13 +02:00
Ryan Houdek 110313e7de Merge pull request #5638 from lioncash/swap 
IR: Add constant for swapping midsections of 256-bit vectors around
 2026-06-30 16:38:56 -07:00
Ryan Houdek 44e24c9e6b Merge pull request #5637 from mrpippy/unicodestring 
Windows/UnixLib: Fix loading with new MemoryWineLoadUnixLibByName mechanism
 2026-06-30 16:23:08 -07:00
Brendan Shanks 6c58fef220 Windows/UnixLib: Fix loading with new MemoryWineLoadUnixLibByName mechanism. 2026-06-30 15:37:55 -07:00
LC 3d593ce87d Merge pull request #5626 from Sonicadvance1/177 
Re-optimize FYL2X for reduced precision x87 path
 2026-06-30 18:15:17 -04:00
Ryan Houdek 36e1b5107e InstcountCI: Update 2026-06-30 15:01:46 -07:00
Ryan Houdek f89123f489 unittests/ASM: Allow up to 3-bits of precision loss 2026-06-30 15:00:16 -07:00
Paulo Matos b7280a765d asm_tests: Re-optimize FYL2X for reduced precision x87 path 2026-06-30 15:00:16 -07:00
Paulo Matos 37b010795e Re-optimize FYL2X for reduced precision x87 path 2026-06-30 15:00:16 -07:00
Ryan Houdek a4f89b79a3 Merge pull request #5636 from lioncash/move 
AVX: Remove unnecessary moves from VPSHUF{D, HW, LW}
 2026-06-30 12:35:13 -07:00
Ryan Houdek 5bde4d875a FEXOfflineCompiler: Fixes HostFeature detection under Win32 
CPUFeature detection is marginally different between Linux and Windows.
FOC was only using the Linux path which had two broken things happening
to it.
- Feature detection was incorrect and enabling/disabling features
  differently from wow64/arm64ec .dll files
- HostType was being set as Linux even though it was generating code for
  WINE

Ensure that when built for Win32 that it uses the correct feature
fetching.

One thing that is still incorrect is that 64-bit or 32-bit is determined
at compile time on win32, whereas the Linux side parses an ELF and
determines bitness at runtime. This doesn't fix that remaining problem
there.
 2026-06-30 11:39:02 -07:00
Ryan Houdek a79c471c31 Merge pull request #5634 from lioncash/shuffle 
unittests: Add selector tests for VPSHUF{D, HW, LW}
 2026-06-30 11:26:09 -07:00
Ryan Houdek 417bd8604c Merge pull request #5632 from lioncash/vpblendd_test 
unittests: Add test for stress-testing VPBLENDD selectors
 2026-06-29 22:13:05 -07:00
LC 3d289f4489 VectorOps: Avoid move in 256-bit VAddP/VFAddP if possible 
If the first source and destination alias, then it's fine to use the
register in destructive operations, since the source data doesn't need
to be preserved.

Tiny saving, but reduces overall register use in some cases.
 2026-06-30 00:23:30 -04:00
Ryan Houdek d138c854f3 Merge pull request #5631 from lioncash/vpblendd 
instcountci/VEX_map3: Add missing third param to VPBLENDD
 2026-06-29 20:11:57 -07:00
Ryan Houdek e26a792b70 Merge pull request #5630 from lioncash/comiss 
Vector: Only signify 128-bit vector loads in UCOMISxOp
 2026-06-29 14:40:44 -07:00
Ryan Houdek 7e2d3b07c0 Merge pull request #5629 from lioncash/comment 
instcountci/VEX_map1: Remove obsolete comments
 2026-06-29 13:55:23 -07:00
Ryan Houdek 394a6f28db Merge pull request #5628 from lioncash/movmsk 
AVX: Reduce codegen for 256-bit VMOVMSKPD/VMOVMSKPS
 2026-06-29 13:42:17 -07:00
Ryan Houdek 72e01274c9 Merge pull request #5627 from lioncash/vpblendw 
instcountci: Add a few more cases for VPBLENDW/VPBLENDD and VSHUFPD/VSHUFPS
 2026-06-29 13:03:49 -07:00
Ryan Houdek 9f2e982944 Merge pull request #5617 from simon902/vcvtps2ph 
OpcodeDispatcher: Fix missing zeroing for vcvtps2ph
 2026-06-29 10:40:45 -07:00
LC c6b0f360fe AVX: Make use of table swapping constant to trim down relevant ops 
Now we can get rid of excessive overhead, with the ability to improve
this further in the future.
 2026-06-29 12:51:30 -04:00
LC ef19242be3 IR: Add constant for swapping midsections of 256-bit vectors around 
This'll let us eliminate some pretty gnarly inserts while
ZIP1Q/UZP1Q/etc are still unavailable.
 2026-06-29 12:50:53 -04:00
LC 2cb4f8b6f5 AVX: Remove unnecessary moves from VPSHUF{D, HW, LW} 
These aren't necessary anymore, since these use operations
that already zero extend.
 2026-06-29 08:43:11 -04:00
LC 59919a0b0c unittests: Expand VPBLENDW selector test to cover 128-bit paths 
Ensures both codepaths are tested. Technically also acts as a good
test for selector masking as well.
 2026-06-29 07:03:16 -04:00
LC bf1857ecc7 unittests: Expand VPBLENDD selector test to cover 128-bit paths 
Ensures both codepaths are tested. Technically also acts as a good
test for selector masking as well.
 2026-06-29 06:56:37 -04:00
LC b40920596a unittests: Add selector stress tests for VPSHUFD 
Ensures any added optimization paths result in the same output.
 2026-06-29 06:56:34 -04:00
LC 3b14c322e8 unittests: Add selector stress tests for VPSHUFLW/VPSHUFHW 
Ensures any added optimization paths result in the same output.
 2026-06-29 06:45:18 -04:00
LC 3e60aa5738 Merge pull request #5610 from Sonicadvance1/171 
FEXCore: Pass host type that changes codegen to FEXCore
 2026-06-29 02:04:31 -04:00
Simon Scherer 6ca2d27c82 unittests/ASM: Add vcvtps2ph_zeroing test to Disabled_Tests_Simulator 2026-06-29 08:00:58 +02:00
Simon Scherer d2f26d1969 InstcountCI: Update 2026-06-29 07:55:05 +02:00
LC fc25443827 unittests: Move full_vpblendw_imm test into the VEX folder 
Keeps all of the selector tests in the same location.
 2026-06-28 23:17:19 -04:00
LC 986800885a unittests: Add test for stress-testing VPBLENDD selectors 
Drops a test in like the one for VPERMQ to ensure that, even if different
optimization paths are introduced, the behavior remains consistent.
 2026-06-28 23:13:39 -04:00
LC 52c6ab1cec instcountci/VEX_map3: Add missing third param to VPBLENDD 
Ensures all registers are non-aliasing, which makes for better unideal
output for observation.
 2026-06-28 21:38:43 -04:00
Ryan Houdek 83a989c6dc Merge pull request #5625 from lioncash/perm 
AVX: Handle two field insertions in VPERMQ
 2026-06-28 17:08:48 -07:00
LC 32b96c259b Merge pull request #5623 from Sonicadvance1/176 
Windows/UnixLib: Adds remaining helpers
 2026-06-28 18:44:35 -04:00
Ryan Houdek 954581c750 Windows/UnixLib: Adds remaining helpers 
Centralizes all the nasty behaviour that will end up breaking when WINE
eventually turns on userspace syscall dispatch. Pushes all of the logic
in to the UnixLib. Support both paths until everyone is migrated to
supporting the UnixLib, then we can delete the bit of code duplication
between the PE side and UnixLib side.

Helpful that everything that gets punched through the UnixLib is
optional, so worst case some optional bits can break for a while.
 2026-06-28 12:58:24 -07:00
LC 24da43f823 Merge pull request #5613 from Sonicadvance1/175 
Windows/UnixLib: Adds support for Hardware TSO support
 2026-06-28 15:56:09 -04:00
LC 600e2ddecf Vector: Only signify 128-bit vector loads in UCOMISxOp 
Mainly a correctness change more than anything. The COMISX and
UCOMISX group of operations only ever load 128 bits when given
a vector source.

No change in codegen, but ensures this doesn't change if any backing
handling changes.
 2026-06-28 15:40:14 -04:00
Ryan Houdek 9740f488cf Merge pull request #5624 from lioncash/psad 
AVX: Slightly trim codegen for VPSADW 256-bit case
 2026-06-28 12:26:37 -07:00
LC f81467fd30 instcountci/VEX_map1: Remove obsolete comments 
Since the registers are non-aliasing, this is about the best we can do
now. These are just holdovers from the initial bring-up of the 256-bit
SVE path where optimization wasn't as strong a concern as getting everything
in place and running properly.
 2026-06-28 14:49:24 -04:00
LC 2ec2c39cf1 AVX: Lessen codegen for VMOVMSKPD 
Performs the same thing as what we've done for VMOVMSKPS. Though,
the effect isn't as drastic, given we're only operating on a max
of four elements as opposed to 8.
 2026-06-28 14:06:32 -04:00
Ryan Houdek c85e426cc5 Merge pull request #5622 from lioncash/perm 
AVX: Handle trivial UZP/ZIP operations in VPERMQ
 2026-06-28 11:01:49 -07:00
LC 41fc57f46c AVX: Lessen codegen for 256-bit VMOVMSKPS 
Currently we can trivially split this up and join the results, which
is much nicer than iterating all the elements individually and shifting
their sign bit over.
 2026-06-28 13:52:19 -04:00
LC a215bb9709 instcountci: Add a few more cases for VPBLENDW/VPBLENDD and VSHUFPD/VSHUFPS 
Just provides a little more comprehensive output
 2026-06-28 13:08:03 -04:00
Ryan Houdek b23fa30099 Merge pull request #5618 from wsxarcher/fixsmcfullvector 
Core: Start a new block for the next op in full SMC check
 2026-06-28 09:25:23 -07:00
Ryan Houdek 848c4b2d68 Merge pull request #5621 from wsxarcher/fixfexfolder 
Do not exec FEX if it is a folder in FEXBash
 2026-06-28 09:22:14 -07:00
Ryan Houdek 4f995cbc1c Merge pull request #5620 from wsxarcher/fixuafbash 
Fix UAF of PS1
 2026-06-28 09:08:51 -07:00
wsxarcher b21c49352e Core: Start a new block for the next op in full SMC check 2026-06-28 18:07:28 +02:00
Ryan Houdek 3470dd1e7b Merge pull request #5619 from lioncash/dpps 
AVX: Handle trivial cases better for VDPPS
 2026-06-28 08:55:34 -07:00