steam-frame/FEX-Emu--FEX
1
0 Fork 0
mirror of https://github.com/FEX-Emu/FEX.git synced 2026-10-10 04:03:44 +02:00
Code Issues Packages Projects Releases 67 Wiki Activity
10,299 Commits 11 Branches 71 Tags
96055cbde704300a279c024feb91c76fd9cc80cf
Commit Graph
362 Commits
Author SHA1 Message Date
Alyssa Rosenzweig c70e44cf05 OpcodeDispatcher: extract Push helper 
Mirrors the Pop helper we added. This cleans up a bunch

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-14 09:37:06 -04:00
Alyssa Rosenzweig 900c62fa7b OpcodeDispatcher: add Pop helpers 
hide away the allocate dance

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-14 09:37:06 -04:00
Alyssa Rosenzweig cab02be637 IR: remove unused pair create/extract 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-14 09:37:06 -04:00
Alyssa Rosenzweig 5631ff4fd5 OpcodeDispatcher: optimize variable shifts 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:43 -04:00
Alyssa Rosenzweig a4545f493e OpcodeDispatcher: add IncrementByCarry helper 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:43 -04:00
Alyssa Rosenzweig f138d7d9b8 OpcodeDispatcher: optimize JA/JNA 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig 3bb9d44bf5 OpcodeDispatcher: optimize SetCFDirect 
generally same # of instructions, but potentially fewer cycles:

old:
        "rmif x4, #63, #nzCv",
        "cfinv"

new:
        "xor x20, x4, #1",
        "rmif x20, #63, #nzCv"

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig 06e6a1b19e OpcodeDispatcher: optimize logic ops 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig 9621eca677 OpcodeDispatcher: fix setrflag masking 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig 1dc22e33ae OpcodeDispatcher: inline some flag calculations 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig 0ef8aaebeb OpcodeDispatcher: optimize BZHI 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig d17c427e47 OpcodeDispatcher: optimize non-flagm2 comiss 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig 79c745929a FEXCore: invert CF internally 
Flag day change to the ABI.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:06:42 -04:00
Alyssa Rosenzweig 8330cc6876 OpcodeDispatcher: defer carry inverts 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 13:05:21 -04:00
Alyssa Rosenzweig 16e6163677 OpcodeDispatcher: drop unused NZCVIndexMask 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 11:26:12 -04:00
Alyssa Rosenzweig bbbc0dc9dc OpcodeDispatcher: fix weird formatting 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-08-10 11:26:12 -04:00
Mai 4882f10536 Merge pull request #3888 from Sonicadvance1/avx128_optimize_blends 
AVX128: Optimize blends
 2024-08-07 17:08:19 -04:00
Billy Laws be4777110c OpcodeDispatcher: Don't apply the address-size flag to segment addresses 
The address-size flag only applies to the offset from the segment base,
rather than the segment address itself.
 2024-07-31 18:14:32 +00:00
Ryan Houdek f8ef6feff9 AVX128: Optimize blends 
Optimizes the AVX128 blends by reusing the prior SSE4.1 implementation.
Only difference is the destination register isn't reused as a source
register.

One confusing thing is that Felix Cloutier's documentation has a typo on
the 256-bit VPBLENDW instruction where it had the top 128-bit lane
reusing the destination instead of sources. So I wrote a unittest to
ensure correctness.

Fixes #3796
 2024-07-23 19:24:19 -07:00
Ryan Houdek 3c5b59d985 AVX128: Implement support for scalar FMA with AFP 
Now that I have AFP supporting hardware I felt better implementing this
since I can run unit tests.

Fixes #3793
 2024-07-22 12:58:19 -07:00
Paulo Matos a1378f94ce X87 Code Refactoring and Optimization Pass 2024-07-22 08:44:45 +02:00
Alyssa Rosenzweig d20b46e46f IR: drop LoadFlag/StoreFlag ops 
pointless, we can just load/store the context now.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-21 15:49:09 -04:00
Ryan Houdek b0bd8a62a2 AVX128: Improve VPERMILPS/PD and VPSHUFD 
VPSHUFD and VPERMILPS are aliases of each other.

Reuses the implementation path from the PSHUFD implementation which has
a few swizzles and then a table lookup.

VPERMILPD is a very simple swizzle per 128-bit lane.

Fixes #3797
Fixes #3784
 2024-07-18 04:10:58 -07:00
Alyssa Rosenzweig 1e709d1150 OpcodeDispatcher: add RecordX87 helper 
calls will be generated.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-16 09:07:35 +02:00
Ryan Houdek d79b7fcc49 Merge pull request #3808 from alyssarosenzweig/rclse/3 
Try to delete RCLSE again
 2024-07-12 20:38:06 -07:00
Ryan Houdek b9a6caea8d Merge pull request #3844 from Sonicadvance1/fix_vmovq 
AVX128: Fixes vmovq loading too much data
 2024-07-12 17:07:32 -07:00
Ryan Houdek 8021dc10a1 OpcodeDispatcher: Force noinline for the function call in the Bind helper 
Clang was inlining a few of the functions it was calling. So force it
never to inline since this is supports to be a little shim trampoline
only.
 2024-07-11 19:00:42 -07:00
Ryan Houdek 7e8d734e43 AVX256: Initial fixes just to get my unittest working 
This is the initial split to decouple AVX256 composed operations from
their MMX/SSE counterparts. This is to work around the subtle
differences with AVX/SSE zext/insert behaviour.
 2024-07-11 18:43:31 -07:00
Alyssa Rosenzweig 294f10fdd0 OpcodeDispatcher: reg cache mmx 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-11 13:21:14 -04:00
Tony Wasserka b9829ed316 OpcodeDispatcher: Replace even more hand-written wrapper templates 2024-07-11 16:19:15 +02:00
Tony Wasserka 4ccec17676 OpcodeDispatcher: Replace more hand-written wrapper templates 2024-07-11 16:19:15 +02:00
Tony Wasserka f45082043b OpcodeDispatcher: Replace hand-written wrapper templates with a generic utility 2024-07-11 16:19:14 +02:00
Alyssa Rosenzweig a4f8bbff02 OpcodeDispatcher: reg cache avx high 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-10 11:36:18 -04:00
Alyssa Rosenzweig cf5ab05b90 OpcodeDispatcher: reg cache AbridgedFTW 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-10 11:36:18 -04:00
Alyssa Rosenzweig 3a2ce240f9 OpcodeDispatcher: reg cache DF 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-10 11:36:18 -04:00
Alyssa Rosenzweig 1f01dd53f7 OpcodeDispatcher: reg cache fprs 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-10 11:36:18 -04:00
Alyssa Rosenzweig 72d41d70b6 OpcodeDispatcher: introduce GPR-only reg cache 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-10 11:36:18 -04:00
Alyssa Rosenzweig 2949bc211d OpcodeDispatcher: thunk through FlushRegisterCache 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-10 11:36:18 -04:00
Tony Wasserka 441187470e OpcodeDispatcher: Avoid monomorphization of some AVX functions 2024-07-10 17:01:30 +02:00
Tony Wasserka 59fd13cc2f OpcodeDispatcher: Avoid monomorphization of even more functions 2024-07-10 17:01:30 +02:00
Tony Wasserka c9e7bfdf16 OpcodeDispatcher: Avoid monomorphization of more functions 2024-07-10 17:01:30 +02:00
Tony Wasserka 2d700c381e OpcodeDispatcher: Avoid monomorphization of large functions 2024-07-10 17:01:30 +02:00
Ryan Houdek 72d6c8ebd6 Merge pull request #3820 from alyssarosenzweig/ir/drop-deferred 
Drop deferred flag infrastructure
 2024-07-09 17:06:25 -07:00
Ryan Houdek ec7c8fd922 AVX128: Optimize QPS/QD variant of gather loads! 
SVE has a special version of their gather instruction that gets similar
behaviour to x86's VGATHERQPS/VPGATHERQD instructions.

The quirk of these instructions that the previous SVE implementation
didn't handle and required ASIMD fallback, was that most gather
instructions require the data element size and address element size to
match. This x86 instruction uses a 64-bit address size while loading 32-bit
elements. This matches this specific variant of the SVE instruction, but
the data is zero-extended once loaded, requiring us to shuffle the data
after it is loaded.

This isn't the worst but the implementation is different enough that
stuffing it in to the other gather load will cause headaches.

Basically gets 32 instruction variants to use the SVE version!

Fixes #3827
 2024-07-08 17:19:18 -07:00
Ryan Houdek c5a0ae7b34 IR: Adds new QPS gather load variant! 2024-07-08 17:19:18 -07:00
Ryan Houdek 0d4414fdd0 AVX128: Removes templated AddrElementSize and add as argument 
NFC
 2024-07-06 18:32:35 -07:00
Alyssa Rosenzweig adc709db2f OpcodeDispatcher: drop remnants of deferred flags 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-05 17:22:41 -04:00
Alyssa Rosenzweig 395573720d OpcodeDispatcher: drop pointless flag defers for shifts 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-05 16:24:54 -04:00
Alyssa Rosenzweig 0e62759d24 OpcodeDispatcher: stop deferring logical 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-05 16:24:54 -04:00
Alyssa Rosenzweig 926b6c3117 OpcodeDispatcher: don't defer mul flags 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-07-05 16:24:54 -04:00