Repository navigation
RESIDENT EVIL 7 biohazard Gold Edition (PPSA04404) #21992
NiekGevers
started this conversation in
Game bring-up
Replies: 1 comment
|
Same message shows up on the RE Village demo (MAIDEN) thread, #9093, with #22271 as the issue and #9950 as the PR. So this looks like an RE Engine-wide blocker, not RE7-specific. If your |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Existing thread
Title
RESIDENT EVIL 7 biohazard Gold Edition (PPSA04404)
Content version tested
01.000.004
AnyPS5 commit
mainbc84729655, built here, with these merged on top: #21776 (98d48523d0), #21948 (b015b6b844), #21719 (5a21f30384), #15941 (416bc9ad90), and the local change under Current blocker.System tested
Windows 11 Home (build 26300), MSYS2 UCRT64 GCC 16.2.0 (Release), Ryzen 5 220, GeForce RTX 5050 Laptop GPU (NVIDIA 621.14), 16 GB RAM.
Furthest point
In game
What is shown
APS5_PROFILE_DRAW=1):DISPATCH_DIRECTabout 1800 times at 2.6 ms each (prologue 1.0 ms, device call 1.1 ms, lock wait 0.3 ms), andDRAW_INDEX_INDIRECT_MULTIabout 1000 times at 3.3 ms each.Current blocker
On
mainthe title stops at the first load after New Game (about 220 s):The owner (local diagnostic print, not pushed) is a 128x64, 4-sample
VK_FORMAT_D32_SFLOATtarget without stencil, with depth-only HTILE that is not pipe-aligned. #9950 implements publication for written D32 surfaces but conflicts withmain. Onmain,PublishToGuestalready publishes a depth plane plus an HTILE plane, so locally onlycapturePublicationand thepublishablecondition were changed to admit this layout:The stops met after that, one at a time, and what got past each one:
external GPU writes over native multisample depth are not implemented, from a GPU store into memory after the real HTILE:mainguards 0x8000 bytes of HTILE for this target, while the HTILE that is not pipe-aligned is 0x1000. Fixed by fix(agc): size HTILE that is not pipe-aligned by its own meta blocks #21948.external GPU writes over native multisample depth are not implementedagain, inside the HTILE: after the owner was published (no host write pending, no clear), a dispatch bound its HTILE (0x211f300000+0x8000) as a writable buffer. Fixed by fix(agc): retire native multisample depth owners that guest writes replace #21719.native multisample depth publication cannot distinguish an unwritten byte from a same-value guest store, at about 198 s, in 1 of 2 runs without fix(agc): publish a reused native multisample depth target with the guest's undecided bytes, then retire its owner #15941. Fixed by fix(agc): publish a reused native multisample depth target with the guest's undecided bytes, then retire its owner #15941 (1 run, which reached the opening).Still needed: #21948, #21719, #15941, and #9950 or the local change above.
How it was run
src/eboot.elfiseboot.bin.esbakandsrc/sce_module/libc.prxissce_module/libc.prx.esbakfrom the dump.libs/holds the built.prxfiles and the MinGW runtime DLLs, andapp0/holds the game files plus the relinker'ssce_module/libc.prx.guest.prx.Environment:
ANYPS5_GPU=RTX,APS5_PROFILE_DRAW=1,APS5_LOG_TIMESTAMPS=1,ANYPS5_HOST_REPORT=1.All reactions