izzy/UnrealEngineUWP
0
0
Fork 0
mirror of https://github.com/izzy2lost/UnrealEngineUWP.git synced 2026-03-26 18:15:20 -07:00
Code Issues Packages Projects Releases Wiki Activity
166,760 Commits 23 Branches 254 Tags
00a43bb11c45cfb2e07fa0e79d0eb7bb7ef04c19
Commit Graph

615 Commits

Author SHA1 Message Date
danny chapman
05de68d4bb Small typo/fix for SpringDamper 
#rb trivial

[CL 16775013 by danny chapman in ue5-main branch]
 2021-06-24 12:09:41 -04:00
Andrew Davidson
24a7315c8f LWC: Replaced Promote/DemoteArrayType with a single ConvertArrayType now that conversion is possible via explicit cast. 
#preflight 60d33a1bd9586b000131ef75

[CL 16769796 by Andrew Davidson in ue5-main branch]
 2021-06-24 05:03:46 -04:00
aurel cordonnier
056021b0ff Fix for LWC compile error in RandomStream 
#fyi jordan.hoffman, andy.davidson
#rnx

[CL 16768226 by aurel cordonnier in ue5-main branch]
 2021-06-23 22:42:32 -04:00
aurel cordonnier
69aa1ba676 [Backout] - CL16766887 
#fyi jordan.hoffmann
Original CL Desc
-----------------------------------------------------------------
Fixed floating point trucation introduced in LWC builds for CL 16765056
   " Fix: RandomPointInBoundingBox is available globally but not available on the random stream
    #jira UE-14027 "

[CL 16767679 by aurel cordonnier in ue5-main branch]
 2021-06-23 21:19:04 -04:00
jordan hoffmann
99bf9f4988 Fix: RandomPointInBoundingBox is available globally but not available on the random stream 
#jira UE-14027

[CL 16766887 by jordan hoffmann in ue5-main branch]
 2021-06-23 20:27:12 -04:00
jordan hoffmann
53fc361d47 Fix: RandomPointInBoundingBox is available globally but not available on the random stream 
#jira UE-14027
#rb justin.hare

[CL 16765056 by jordan hoffmann in ue5-main branch]
 2021-06-23 19:00:46 -04:00
dmytro vovk
55d0033dbd Added missing double overload for VectorGetComponentDynamic 
#rb none
#jira none

[CL 16724432 by dmytro vovk in ue5-main branch]
 2021-06-20 17:48:33 -04:00
Andrew Davidson
41de59d4b3 LWC: Avoid replacing double types with float precision versions on save. 
#rb zak.middleton
#preflight 60ca80992ab21800013d03ea

[CL 16701618 by Andrew Davidson in ue5-main branch]
 2021-06-17 04:38:26 -04:00
danny chapman
88a01461c0 Remove the approx version of spring damper. Just deleting some very newly introduced options so hopefully not problematic. 
#rb trivial

[CL 16660436 by danny chapman in ue5-main branch]
 2021-06-14 13:18:32 -04:00
dmytro vovk
95e2e5b221 Fixed compiling QAGame with LWC on 
#jira UE-117930
#rb none
#fyi Andy.Davidson

[CL 16652707 by dmytro vovk in ue5-main branch]
 2021-06-12 19:14:42 -04:00
danny chapman
d4d126898b Trivial/compiler warning fixes 
#rb trivial

[CL 16651548 by danny chapman in ue5-main branch]
 2021-06-12 06:33:41 -04:00
danny chapman
3ff5775f85 Add SpringInterpQuaternion script function and ControlRig node. Also add an initialization option to the spring interp functions. COntrol rig nodes support the current input being optional, and expose strength rather than stiffness. 
#jira UE-117575
#rb aaron.cox,cesar.castro

[CL 16649881 by danny chapman in ue5-main branch]
 2021-06-11 20:07:12 -04:00
dmytro vovk
23eb0bd262 Added missing NEON double vector functions and fixed VectorMaskBits 
#jira UE-117058
#rb none

[CL 16592846 by dmytro vovk in ue5-main branch]
 2021-06-08 16:28:31 -04:00
Zak Middleton
dbd4f3ea3a #ue5 - LWC: add platform flag for AVX2 feature level, and enable AVX2 on next-gen consoles. 
[CL 16575354 by Zak Middleton in ue5-main branch]
 2021-06-07 15:05:37 -04:00
Zak Middleton
eb33c4e621 #ue5 - LWC: add RESTRICT keyword to VectorDeinterleave, to protect against duplicated inputs for the two out vars. 
#jira none
#fyi Dmytro.Vovk

[CL 16574022 by Zak Middleton in ue5-main branch]
 2021-06-07 13:42:27 -04:00
Zak Middleton
75107c146f #ue5 - LWC: AVX/AVX2 optimizations for VectorShuffle(), VectorSwizzle(), and VectorReplicate(). Instruction latencies on AVX or AVX2 dropped from ~6 to ~1-2 in many cases, and are hand-tuned between whether AVX1 or AVX2 are available. 
Instruction latencies and uop counts from documentation manuals and community testing (referenced in the source), plus comparing assembly output on godbolt led to many shuffle and swizzle template overrides and specializations.

Also optimized access to the XY and ZW portions of the AVX m256 union to sub-optimal MSVC compiler output that turns them in to memory ops rather than simple casts or 1-latency extracts.

Added more shuffle/swizzle validation tests to UnrealMathTests to stress specific in-lane and cross-lane permutes.

#jira UE-107989
#codereview Dmytro.Vovk

[CL 16554362 by Zak Middleton in ue5-main branch]
 2021-06-03 18:06:52 -04:00
yoan stamant
4984a31651 [ColorList] fixed some typos and unified method descriptions 
#rb trivial
#preflight 60b6bacc4d507d000193d132

#ROBOMERGE-SOURCE: CL 16525041 in //UE5/Private-Frosty/...
#ROBOMERGE-BOT: STARSHIP (Private-Frosty -> Main) (v826-16501804)

[CL 16525042 by yoan stamant in ue5-main branch]
 2021-06-01 19:44:59 -04:00
Zak Middleton
3df9b646e3 #ue5 - LWC: Fix build break (missing comment //) 
#jira none
#rb none

[CL 16490450 by Zak Middleton in ue5-main branch]
 2021-05-27 15:50:19 -04:00
Zak Middleton
50e11c8c0b #ue5 - LWC: Various improvements especially for AVX/AVX2: 
- Improved AVX loading of two separate m128d values using native instruction.
- Improved VectorGetComponent() in MSVC for some specializations because it generates suboptimal code with the generic approach. Biggest benefit to VectorGetComponent(V, 0) which is the most common.
- Implemented VectorDot3Scalar() for FPU and SSE for when only the scalar value is needed, which can avoid some permute operations. Added tests to UnrealMathTests.
- Added a few more specializations for AVX1 for VectorSwizzleTemplate<>.

#jira UE-107989
#codereview Dmytro.Vovk

[CL 16489826 by Zak Middleton in ue5-main branch]
 2021-05-27 15:16:37 -04:00
Zak Middleton
7dd1c9fe5d #ue5 - LWC: AVX requires 32-byte alignment for __m256d, but at least now we don't use the SVML compile flag erroneously. 
#jira none
#fyi Andrew.Davidson

[CL 16487740 by Zak Middleton in ue5-main branch]
 2021-05-27 13:22:02 -04:00
Zak Middleton
1a357dee3f #ue5 - LWC: SSE double alignment changed to always be 16-byte (instead of 32 depending on build flags). We're not using any AVX instructions which require 32 byte alignment. 
#jira none
#fyi Andrew.Davidson

[CL 16487687 by Zak Middleton in ue5-main branch]
 2021-05-27 13:16:33 -04:00
Andrew Davidson
b428f9a785 LWC: ShooterGame build fixes 
#rb trivial
#preflight 60af814c241ce50001fda59c

[CL 16482698 by Andrew Davidson in ue5-main branch]
 2021-05-27 08:07:26 -04:00
Andrew Davidson
bc00b05ca9 LWC: Fix chaos precision loss warnings 
#rb cedric.cailllaud
#jira UE-116286
#preflight 60af7c773dce58000179a058

[CL 16482537 by Andrew Davidson in ue5-main branch]
 2021-05-27 07:51:43 -04:00
Zak Middleton
806df69598 #ue5 - LWC: Avoid using SVML on clang on windows. 
#jira none
#rb none

[CL 16477647 by Zak Middleton in ue5-main branch]
 2021-05-26 18:58:49 -04:00
dmytro vovk
aeb8b111fd Optimized NEON double version of VectorDot3\4 and added Vector3Scalar version 
[CL 16477262 by dmytro vovk in ue5-main branch]
 2021-05-26 18:25:22 -04:00