steam-frame/FEX-Emu--FEX
1
0 Fork 0
mirror of https://github.com/FEX-Emu/FEX.git synced 2026-10-14 20:00:46 +02:00
Code Issues Packages Projects Releases 67 Wiki Activity
6,306 Commits 11 Branches 71 Tags
170c9ee9e46c5f59ff73c46cabcaf330cae8a460
Commit Graph
2541 Commits
Author SHA1 Message Date
Ryan Houdek 170c9ee9e4 LookupCache: Switch to fextl 2023-03-30 16:28:33 -07:00
Ryan Houdek 1eb36b8b31 Convert a ton of things over to fextl 2023-03-30 16:28:33 -07:00
Ryan Houdek 43e6d398b6 X86Dispatcher: Move xbyak to custom types 2023-03-30 08:49:26 -07:00
Mai 88dba60bee Merge pull request #2579 from Sonicadvance1/invalid_instruction_log 
Core: Add a new log message for unsupported instruction
 2023-03-29 22:52:08 -04:00
Ryan Houdek d615ae9c6a Core: Add a new log message for unsupported instruction 
The previous log in the frontend is super useful when an instruction
decoding wasn't supported.
Now that most of AVX is covered, a game will crash on SIGILL (and
usually catch it) and close without any indication.

Now if the instruction is decoded but it is invalid for the
configuration, still output a message as a good indicator that the game
is using instructions that the host doesn't support.

Will let us still pick up on games crashing due to lack of SVE very
easily.
 2023-03-29 14:38:16 -07:00
Lioncache 73d250c555 ARMEmitter: Handle SVE2 integer add/subtract wide category 2023-03-29 17:02:18 -04:00
Lioncache 337f8b06a3 ARMEmitter: Convert SVE2 integer multiply long to wide helper 
Unifies the emitted ops under the same underlying emitter function.
 2023-03-29 16:52:37 -04:00
Lioncache 87fa545bd0 ARMEmitter: Convert SVE2 integer add/subtract long to wide helper 
The generic helper will be used to implement the remaining unimplemented
category from this group
 2023-03-29 16:52:34 -04:00
Ryan Houdek 7747ac8de8 Merge pull request #2576 from lioncash/mul 
ARMEmitter: Handle SVE Integer Multiply-Add - Unpredicated group
 2023-03-29 13:14:29 -07:00
Lioncache deb1c9e933 ARMEmitter: Handle SVE mixed sign dot product category 2023-03-29 15:43:07 -04:00
Lioncache 672a88395d ARMEmitter: Handle SVE2 saturating multiply-add high category 2023-03-29 15:38:38 -04:00
Lioncache 88524ce718 ARMEmitter: Handle SVE2 saturating multiply-add long category 2023-03-29 15:33:49 -04:00
Lioncache bb153054f9 ARMEmitter: Handle SVE2 integer multiply-add long category 2023-03-29 15:28:32 -04:00
Lioncache 22a7a49042 ARMEmitter: Handle SVE2 complex integer multiply-add 2023-03-29 15:19:09 -04:00
Lioncache 9876f3eb5c ARMEmitter: Handle SVE2 saturating multiply-add interleaved long category 2023-03-29 15:08:30 -04:00
Lioncache b9e4ce4029 ARMEmitter: Handle SVE integer dot product (unpredicated) category 2023-03-29 14:55:06 -04:00
Lioncache 0c048772e0 ARMEmitter: Handle CDOT (vectors) 2023-03-29 14:54:35 -04:00
Lioncache 830c1884d1 OpcodeDispatcher: Handle store variants of VMASKMOVPD/VMASKMOVPS 
And with that, we support all of the AVX1-only instructions.

The remaining instructions for full AVX1 support is now just the SSE4.2
string instructions.
 2023-03-29 14:03:23 -04:00
Lioncache 5abf9de8a5 IR: Add VStoreVectorMasked IR op 
Will be used to implement the store variants of VPMASKMOV and
VMASKMOVP{D, S}
 2023-03-29 14:03:20 -04:00
Lioncache 25960fe6b1 OpcodeDispatcher: Handle load variants of VMASKMOVP{D, S} 2023-03-28 10:35:23 -04:00
Lioncache eb8626c1f7 IR: Add VLoadVectorMasked IR op 
Will be used to implement the load variants of VMASKMOVP{D, S} and
VPMASKMOV{D, Q}

Particularly useful, since with SVE this behavior can be collapsed into
two instructions (CMPGT followed by the relevant LD1 load instruction)
 2023-03-28 01:57:25 -04:00
Lioncache ef7853ca4a ARMEmitter: Fix treating 32-bit elements as 64-bit with ld1w 
These conditionals were accidentally inverted and were treating 32-bit
elements as 64-bit ones, when this is unintended.

Also add missing tests to ensure this doesn't slip through in the
future.
 2023-03-28 00:55:04 -04:00
Ryan Houdek 55d65f3aea Merge pull request #2570 from lioncash/mov 
Arm64/VectorOps: Remove a few unnecessary EORs from comparisons in SVE path
 2023-03-27 13:35:48 -07:00
Ryan Houdek 2c2abc550b Merge pull request #2569 from lioncash/psadbw 
OpcodeDispatcher: Handle VMPSADBW
 2023-03-27 13:31:52 -07:00
Lioncache a1dc132f03 Arm64/VectorOps: Eliminate unnecessary EOR and MOV in FP compares 
We can use the zeroing variant of MOVPRFX to perform the same behavior.
 2023-03-27 16:16:28 -04:00
Lioncache ef31e0c7c7 OpcodeDispatcher: Handle VMPSADBW 2023-03-27 16:00:24 -04:00
Lioncache eecd016ba8 Arm64/VectorOps: Eliminate unnecessary EOR and MOV in VCMP{EQ,GT} 
We can use the zeroing version of MOVPRFX to perform the same behavior.
 2023-03-27 15:38:03 -04:00
Lioncache b2ec6d5208 OpcodeDispatcher: Move MPSADBW implementation into helper 
This will be used for implementing the AVX variant of this instruction.
 2023-03-27 12:40:28 -04:00
Lioncache 416a7b825d ARMEmitter: Move SVE2IntegerSaturatingAddSub over to using SVE2IntegerPredicated helper 
Deduplicates some code.
 2023-03-27 12:21:42 -04:00
Lioncache c3511ffa48 ARMEmitter: Move SVEIntegerPairwiseArithmetic over to using SVE2IntegerPredicated helper 
Deduplicates some code.
 2023-03-27 12:21:42 -04:00
Lioncache b9c3277b09 ARMEmitter: Move SVE2IntegerHalvingPredicated over to using SVE2IntegerPredicated helper 
Deduplicates some code.
 2023-03-27 12:21:38 -04:00
Lioncache 09f259c458 ARMEmitter: Handle SVE2 saturating/rounding bitwise shift left (predicated) category 2023-03-27 11:58:02 -04:00
Lioncache 7a2d23e189 ARMEmitter: Handle SVE2 integer unary operations (predicated) category 2023-03-27 11:36:06 -04:00
Lioncache 1e10561892 ARMEmitter: Handle SVE2 integer pairwise add accumulate long category 2023-03-27 11:22:39 -04:00
Ryan Houdek 9183cf144f Merge pull request #2547 from Sonicadvance1/standard_string_theory_versus_fex 
Convert most std::string over to fextl
 2023-03-24 04:37:40 -07:00
Ryan Houdek 043a03547f Resolve review comments 2023-03-24 04:09:03 -07:00
Ryan Houdek 7022b3b825 Review c_str() changes 2023-03-23 12:45:14 -07:00
Lioncache dfeee6a1ed Arm64/VectorOps: Fix comment typo in VUShr 
Meant to write USHL here, rather than SSHL
 2023-03-21 22:25:34 -04:00
Lioncache 91d6dc0528 Arm64/VectorOps: Handle 64-bit elements in VSShr 
Makes it consistent with all of the other variable shift IR ops.

Now this won't explode if we ever implement VPSRAVQ from AVX-512
 2023-03-21 22:23:12 -04:00
Ryan Houdek 54e784742c Merge pull request #2558 from lioncash/shift 
OpcodeDispatcher: Handle VPSLLVD/VPSLLVQ
 2023-03-21 15:15:40 -07:00
Ryan Houdek 22936144ce Merge pull request #2557 from lioncash/pred 
ARMEmitter: Handle SVE Inc/Dec by Predicate Count category
 2023-03-21 15:12:39 -07:00
Lioncache fea0162096 OpcodeDispatcher: Handle VPSLLVD/VPSLLVQ 2023-03-21 17:08:03 -04:00
Lioncache 8287375117 x86_64/VectorOps: Implement VUShl IR op 
This will be used to implement VPSLLVD/VPSLLVQ
 2023-03-21 16:57:48 -04:00
Lioncache 0106f03b44 Arm64/VectorOps: Implement VUShl IR op 
This will be used to implement VPSLLVD/VPSLLVQ
 2023-03-21 16:55:33 -04:00
Lioncache 3e2d1cf8c0 OpcodeDispatcher: Factor out variable shift code into helper 
We can slightly modify the table flags for VPSRAVD in order to
consolidate the variable shift code into one place.

Useful, since we still need to implement VPSLLVD and VPSLLVQ.

Also in the event we start supporting AVX-512, then the code can just be
extended from a single point.
 2023-03-21 16:25:20 -04:00
Lioncache 392438d70a ARMEmitter: Handle SVE inc/dec register by predicate count group 2023-03-21 16:16:13 -04:00
Lioncache 0930a10026 ARMEmitter: Handle SVE inc/dec vector by predicate count group 2023-03-21 16:16:13 -04:00
Lioncache cb8ab3d328 ARMEmitter: Handle SVE saturating inc/dec register by predicate count group 2023-03-21 16:16:13 -04:00
Lioncache a670c1d762 ARMEmitter: Handle SVE saturating inc/dec vector by predicate count group 2023-03-21 16:16:11 -04:00
Ryan Houdek 91f056bd2d Merge pull request #2556 from lioncash/shift 
OpcodeDispatcher: Handle VPSRLVD/VPSRLVQ
 2023-03-21 12:29:19 -07:00