steam-frame/FEX-Emu--FEX
1
0 Fork 0
mirror of https://github.com/FEX-Emu/FEX.git synced 2026-10-14 08:00:50 +02:00
Code Issues Packages Projects Releases 67 Wiki Activity
11,674 Commits 11 Branches 71 Tags
1700a73382d73f09b2fcbdb83b8b215cd1337db7
Commit Graph
626 Commits
Author SHA1 Message Date
Ryan Houdek 9f2f10a65f IR: Add new 128-bit vector move operation 
This allows us to slightly optimize some x87 behaviour.
 2025-03-12 22:21:28 -07:00
Ryan Houdek a37d6a3841 JIT: Optimize CAS 
Hey kid, want to see a sick trick?

Finally optimal codegen for 64-bit cmpxchg.
 2025-03-07 14:24:45 -08:00
Ryan Houdek 031afbfb18 Various: Adds missing SPDX file headers 
NFC
 2025-03-07 11:34:31 -08:00
Ryan Houdek 66841ce35c IR: Remove some static analysis warnings 
NFC
 2025-03-05 12:30:27 -08:00
Ryan Houdek 2bf87ff40f Various: Removes warnings about uninitialized variables 
NFC. These wouldn't even occur in practice.
 2025-03-04 19:59:26 -08:00
Paulo Matos b5241e0f60 Improve reciprocal estimate and tests 
3DNow Reciprocal estimations did not have enough accuracy. Tests were enabled
to check for accurate values of reciprocals.

* where needed, reciprocal accuracy was increased.
* 3DNow sqrt reciprocal fixed for negative values.
* New helper VFCopySign IR op added.

Fixes #4319.
 2025-02-27 15:07:14 +01:00
Ryan Houdek de6931b1f5 OpcodeDispatcher: Emulate SHA1RNDS4 with ARM sha extensions 
```diff
     "sha1rnds4 xmm0, xmm1, 10b": {
-      "ExpectedInstructionCount": 55,
+      "ExpectedInstructionCount": 10,
```

So I spent a few hours glaring at this instruction. Then spent a few
more glaring in to the sunset and then found the optimization.
 2025-02-22 04:57:58 -08:00
Ryan Houdek cdf6a16efc JIT: Implement ARM VSha1SU1 IR operation 2025-02-19 11:33:38 -08:00
Ryan Houdek bbcd4c168c JIT: Implement support for VSha256U1 operation 2025-02-18 18:03:31 -08:00
Ryan Houdek 7cd4d53fa9 Fixes some instances of auto usage with unintentional copy 
Just switch the uses over to `const auto&`
 2025-02-12 23:51:57 -08:00
Ryan Houdek 3ef695cf70 JIT: Optimize memory stores with zero 
Minor optimization but I've seen it around.
 2025-02-09 22:50:12 -08:00
Billy Laws 0e11a9b7ac FEXCore: Don't copy IR after compilation 
This wastes a significant amount of time when the copies IR is promptly
thrown away after compilation anyway.
 2025-02-04 16:52:16 +00:00
Billy Laws 68c77d12d4 FEXCore: Don't delete IRListView move constructor 2025-02-04 16:51:59 +00:00
Tony Wasserka ebb7137839 IRDumper: Allow const RA data 2025-01-31 15:20:23 +01:00
Paulo Matos 44c65c35c8 Revert "Enable RA of SVE Predicate Registers" 
This reverts commit fcbf0de05a.

The initial user of this code has been re-implemented in  b148cc6c.
This is not needed any longer so we're removing it.
 2025-01-29 11:56:19 +01:00
Paulo Matos 0bccb1ece5 Ensure predicate cache is reset when control flow leaves block 
Whenever the control float leaves the block, it might clobber the
predicate register so we reset the cache whenever that happens.

Fixes #4264
 2025-01-27 20:12:43 +01:00
Tony Wasserka da58e6a597 Fix warnings about unused variables 2025-01-21 12:07:33 +01:00
Tony Wasserka e54b9237c6 Drop use of assume-asserting logging macros 2025-01-21 12:01:33 +01:00
Paulo Matos 2d53867668 x87 fst/fld optimization for different addrmodes 
Includes tests and instcountci files and tests.
When the x87 optimizations were implement, we missed
optimizing different addressing modes. This commit addresses this issue.

Discussed in #4252.
 2025-01-14 16:20:33 +01:00
Ryan Houdek 8cfc016b3f Merge pull request #4265 from pmatos/RevertPredCache 
Revert pred cache
 2025-01-10 12:25:38 -08:00
Paulo Matos cbda688e29 Revert "Cache predicate register generation from pattern" 
This reverts commit 72a4063651.

Caused #4264
 2025-01-10 12:52:11 +01:00
Paulo Matos 159ed07e68 Print arg type f80Bit 2025-01-10 09:14:49 +01:00
LC a6c67ca749 Merge pull request #4251 from Sonicadvance1/ir_numelements_to_elementsize 
IR: Change convention from number of elements to elementsize
 2025-01-04 18:39:24 -05:00
Ryan Houdek a47ed105e7 x87StackOptimizationPass: Minor opt to f80 fchs and fabs 
It's faster to load the f80 sign mask from our named vector constants
than synthesizing the values. Changes a 4 instruction sequence to
synthesize to be 1 load.
 2025-01-03 13:47:10 -08:00
Ryan Houdek 1ecfa3253d IR: Change convention from number of elements to elementsize 
The IR stores elementsize, where the json was wanting number of
elements. While the IR Emitter function declaration always wanted
element size. This was causing us to do a little dance from ElementSize
-> Number of elements -> ElementSize. Just pass the ElementSize directly
instead of this bogus little dance.
 2025-01-03 11:01:03 -08:00
Billy Laws efd6e95059 OpcodeDispatcher: Match x86 overflow behaviour for F2I conversions 
ARM behaviour here is to saturate on overflow or NaN inputs, whereas
X86 returns a sentinel value of 2^(bitsize-1), explicitly emulate this.
 2024-12-30 00:42:55 +00:00
Alyssa Rosenzweig 29405f2690 ConstProp: fix 32-bit masking behaviour 
if we want to replace a node with one of its sources, we need to zero extend if
the source is 64-bit and the destination is 32-bit.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-12-13 10:44:56 -05:00
Alyssa Rosenzweig 51f505acca ConstProp: drop some unused headers 
ycm complained.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
 2024-12-13 10:44:56 -05:00
Ryan Houdek d00d476a0a IR/Passes: Adds missing comment that clang-format keeps complaining about locally 
NFC
 2024-12-11 15:03:28 -08:00
Paulo Matos 72a4063651 Cache predicate register generation from pattern 2024-12-06 10:15:38 +01:00
Paulo Matos 1d3ce30e50 Generate SVE for 80bit load/stores when possible 
Fixes #4166.
 2024-12-06 10:15:29 +01:00
Paulo Matos fcbf0de05a Enable RA of SVE Predicate Registers 2024-12-02 18:35:31 +01:00
Ryan Houdek 9b6cc8f7e0 IR: Convert OpSize over to enum class 
NFC

Do the final mopping up to convert the OpSize enum to an enum class!
 2024-10-29 16:52:16 -07:00
Ryan Houdek 82f936cb6d IR: Converts base IR operations to store OpSize sizes 
NFC

Finally converts the IR operations themselves to store the OpSize for
the IR operation size and element sizes.

This also finally, FINALLY, converts that remaining `_Constant` helper
to stop using a size field that is specified in bits rather than bytes
like all the other IR op handlers. That thing was so confusing and now
it's gone.
 2024-10-28 21:26:59 -07:00
Ryan Houdek 034b62292b Passes/x86StackOptimization: Fixes implicit conversion of OpSize 2024-10-28 19:26:02 -07:00
Ryan Houdek 6cca007817 IR: Fix some missing OpSize conversions 
Missed these in the previous PR.
 2024-10-28 16:25:36 -07:00
Ryan Houdek 65ddae1b71 IR: Change F80VBSLStack to use IR::OpSize 2024-10-28 02:25:17 -07:00
Ryan Houdek 063f524084 IR: Change F80CVTToInt to use IR::OpSize 2024-10-28 02:24:25 -07:00
Ryan Houdek 2dd0a82059 IR: Change F80CVTTo to use IR::OpSize 2024-10-28 02:23:47 -07:00
Ryan Houdek b810070e9f IR: Change F80CVTInt to use IR::OpSize 2024-10-28 02:07:20 -07:00
Ryan Houdek 4c7ac17f7d IR: Change F80CVT to use IR::OpSize 2024-10-28 02:06:48 -07:00
Ryan Houdek 84767c8b20 IR: Change PushStack to use IR::OpSize 2024-10-28 02:05:35 -07:00
Ryan Houdek eccfb53bd5 IR: Change StoreStackMemory to use IR::OpSize 2024-10-28 02:01:47 -07:00
Ryan Houdek f4e930262f IR: Change PCLMUL to use IR::OpSize 2024-10-28 01:50:24 -07:00
Ryan Houdek d26d9e7e03 IR: Change CRC32 to use IR::OpSize 2024-10-28 01:50:24 -07:00
Ryan Houdek 51c1998d70 IR: Change VAESDecLast to use IR::OpSize 2024-10-28 01:50:24 -07:00
Ryan Houdek 87f818249d IR: Change VAESDec to use IR::OpSize 2024-10-28 01:50:23 -07:00
Ryan Houdek d81f92f5e2 IR: Change VAESEncLast to use IR::OpSize 2024-10-28 01:50:23 -07:00
Ryan Houdek 29ffe02afe IR: Change VAESEnc to use IR::OpSize 2024-10-28 01:50:23 -07:00
Ryan Houdek dfe4076fe4 IR: Change Vector_F64ToI32 to use IR::OpSize 2024-10-28 01:50:23 -07:00