08-04-2017, 11:24 PM
The option "Use GPU Texture Scaling" does not work under Vulkan correct? I wanted to test on Vulkan but yea is a black screen unless that option is enabled.
The Legend of Heroes: Trails of Cold Steel [BLUS31572]
|
08-04-2017, 11:24 PM
The option "Use GPU Texture Scaling" does not work under Vulkan correct? I wanted to test on Vulkan but yea is a black screen unless that option is enabled.
08-05-2017, 06:36 AM
(08-04-2017, 11:24 PM)BLACKSAMA Wrote: The option "Use GPU Texture Scaling" does not work under Vulkan correct? I wanted to test on Vulkan but yea is a black screen unless that option is enabled. I'm not sure if the 'Use GPU Texture Scaling' does anything or not under Vulkan. The 'Write Color Buffers' option is needed if using Vulkan in this game to be able to render just about anything to the screen. There's a massive FPS hit observed with using 'Write Color Buffers' in either video renderer though.
Video driver: (nvidia-384) 384.59-0ubuntu0~gpu16.04.1 / CPU: AMD FX-8300 / OS Ubuntu 16.04
08-05-2017, 05:39 PM
(08-05-2017, 06:36 AM)Green Wrote:(08-04-2017, 11:24 PM)BLACKSAMA Wrote: The option "Use GPU Texture Scaling" does not work under Vulkan correct? I wanted to test on Vulkan but yea is a black screen unless that option is enabled. The option Write Color Buffers also works in Open GL, it seems to fix everything in both backends but it absolutely destroys any resemblances of performance. As reference Open GL does not work without Use GPU Scaling, everything is black just like Vulkan, so it is the Use GPU Scaling option that makes this game playable, I was checking the tooltip and yea the option does not work on Vulkan says "for now only works in OpenGL".
Something is definitely up with latest build is the third time it goes unresponsive and controls go haywire, music keeps going as if nothing, but it either does not accept any input or it goes crazy with the inputs. I only have this and Steel II so I am not sure if is a general problem or only Steel.
Edit: last revision is broken, emulator randomly goes unresponsive, problem originated between this https://github.com/RPCS3/rpcs3/pull/3142 and this https://github.com/RPCS3/rpcs3/pull/3037 investigating.
08-09-2017, 03:52 AM
Code: S {PPU[0x100001b] Thread (_libsail-control) [0x00d7d0c4]} sys_prx: Loaded module: /dev_flash/sys/external/libdivxdec.sprx I was recently testing the precompiled Linux AppImage builds (rpcs3-v0.0.3-2017-08-08-e9b020b-7553_linux64) and had ran into the above error. This is something that I've had to previously workaround on my system in by building with option -DUSE_SYSTEM_FFMPEG=ON and with a newer version of FFMPEG. Otherwise to get around this error you currently must rename the game's movie folder to something else so that the opening movie sequence are bypassed. RPCS3.log.7z (Size: 53.74 KB / Downloads: 144)
Video driver: (nvidia-384) 384.59-0ubuntu0~gpu16.04.1 / CPU: AMD FX-8300 / OS Ubuntu 16.04
08-10-2017, 07:27 AM
(08-09-2017, 03:52 AM)Green Wrote: Thanks for reporting. Should get fixed with this PR: https://github.com/RPCS3/rpcs3/pull/3190 Retest when merged.
Asus N55SF, i7-2670QM (~2,8 ghz under typical load), GeForce GT 555M (only OpenGL)
08-10-2017, 01:31 PM
Code: S {PPU[0x100001b] Thread (_libsail-control) [0x00d7d0c4]} PPU: LLVM: Loaded module v2-libsail_avi.sprx-8187437EC9EA899D-bdver1.obj Above is the test result for rpcs3-v0.0.3-2017-08-10-69a5f3d-7575_linux64.AppImage on Ubuntu 16.04. The game still hangs while attempting to load the first opening video sequence. Attached are two log files for comparison. RPCS3_3190.log.7z (Size: 54.46 KB / Downloads: 139) RPCS3_shared_ffmpeg.log.7z (Size: 139 KB / Downloads: 147)
Video driver: (nvidia-384) 384.59-0ubuntu0~gpu16.04.1 / CPU: AMD FX-8300 / OS Ubuntu 16.04
08-17-2017, 12:33 PM
(08-10-2017, 01:31 PM)Green Wrote: Maybe this PR: https://github.com/RPCS3/rpcs3/pull/3244
Asus N55SF, i7-2670QM (~2,8 ghz under typical load), GeForce GT 555M (only OpenGL)
08-17-2017, 03:58 PM
(08-17-2017, 12:33 PM)ssshadow Wrote: Maybe this PR: https://github.com/RPCS3/rpcs3/pull/3244 Still no dice. eb22e4e-178_linux64 also hangs wit the same error at starting the opening movie sequence. The following is taken from my build output with using -DUSE_SYSTEM_FFMPEG=ON (Which does play the movie sequences.) Code: -- RPCS3: using shared ffmpeg The attached log file is for eb22e4e-178_linux64. RPCS3_eb22e4e.log.7z (Size: 58.56 KB / Downloads: 134)
Video driver: (nvidia-384) 384.59-0ubuntu0~gpu16.04.1 / CPU: AMD FX-8300 / OS Ubuntu 16.04
As of the current Master branch, the Camp Menu loading time is instantaneous now. The FPS performance hit since Commit d9a6930 remains present. In the current Master branch, the Windows 10 FPS performance is between 5 to 8 FPS slower than Ubuntu 16.04 in OpenGL when compared with similar config settings & same hardware. (Commit d9a6930 largely affects Ubuntu, but the commit before that places Ubuntu at 10 to 14 FPS greater than Windows 10.)
F {SPU[0x200000c] Thread (_libadec_at3CellSpursKernel0)} MEM: Access violation reading location 0x0 is still a very common occurrence. (This makes really long boss fights unplayable, as advancing further hangs the game.)
Video driver: (nvidia-384) 384.59-0ubuntu0~gpu16.04.1 / CPU: AMD FX-8300 / OS Ubuntu 16.04
|
« Next Oldest | Next Newest »
|