Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Resolved [Bug Report][OpenGL-Accurate Date] Dragon Ball Z Infinite World [NTSC-U]
#1
# PCSX2 versions used: 
  • pcsx2-v1.3.1-1207-gd7391ba-windows-x86 (latest dev build)
# Emulation settings: "Safest" preset, no hacks.
# Plugins used: GSdx AVX2 using the OpenGL renderer, DX11 renderer and software mode
# Plugin settings: CRC Hack Level "None", toggling the Accurate Date option the OpenGL renderer
# Description: 
This bug seems to be exclusive to the OpenGL renderer, specifically only when the Accurate Date option is used.
When using the OGL renderer with the Accurate Date option, cell shaded 3D models appear to have some extra shine spots applied to them.
Turning that option off will make those buggy spots disappear.

# CPU: Intel® Core™ i5-4690K CPU @ 3.50GHz
# GPU: NVIDIA GeForce GTX 970 (Driver: NVIDIA 355.98)

# Screenshots:

OpenGL renderer with Accurate Date on, the bug should be fairly obvious when looking at the heads of the left and right 3D models.
[Image: jaMtGCE.jpg]

OpenGL renderer, Accurate Date off. The shiny spots are gone, which may indicate that this is an Accurate Date bug only.
[Image: AMAvxWZ.png]

Software mode, can't tell the difference from this and the previous screenshot.
[Image: eigNOQ2.png]

GS dump is attached to this post.


Attached Files
.7z   DBZInfWorld- GS-AccurateDateBug.7z (Size: 1,87 MB / Downloads: 76)

Sponsored links

#2
Not all games need Accurate DATE, so if it works fine with it off... leave it off Tongue2
[Image: vwah44]
Gaming: Intel i7 3770k @ 4.2Ghz | R9 290 | 16GB RAM | 480GB(240GB+240GB RAID0) SSD | 3 TB HDD | 1 TB HDD | 500GB HDD
Server: AMD FX 6300 @ 4.4Ghz | GTX 670 | 16GB RAM | 240GB SSD | 320GB HDD
PCSX2 General Troubleshooting FAQ
#3
I'm not sure about this -- would be better to notify gregory about this issue just in case.
We're supposed to be working as a team, if we aren't helping and suggesting things to each other, we aren't working as a team.
- Refraction

[Image: 84t1dRu.png]
#4
I'm sure the accurate options shouldn't cause errors, at worst it should be the same. I think there's also a couple of games which look worse with accurate blending on too.
[Image: ref_sig_anim.gif]
Like our Facebook Page and visit our Facebook Group!
#5
(10-05-2015, 04:58 PM)refraction Wrote: I'm sure the accurate options shouldn't cause errors, at worst it should be the same.  I think there's also a couple of games which look worse with accurate blending on too.

It isn't that easy. A rendering is composed of N draw calls. Fixing one might actually break the rendering completely.

Anyway it smells like a bug. Could be a mix of complex feature, or a bad rounding.

What happen if you add this option in the ini file
Quote:override_GL_ARB_shader_image_load_store = 0

It will disable the detection of atomic texture and therefore one of the accurate_date algo.
#6
(10-07-2015, 10:10 AM)gregory Wrote: It isn't that easy. A rendering is composed of N draw calls. Fixing one might actually break the rendering completely.

Anyway it smells like a bug. Could be a mix of complex feature, or a bad rounding.

What happen if you add this option in the ini file

It will disable the detection of atomic texture and therefore one of the accurate_date algo.

I can confirm that the bug goes away. If that is set to 1 and I switch to HW mode, I get the bug. If I set it back to 0 and cycle between SW and HW modes, the bug is there again.
#7
Ok. Cool, it remains 2 possibilities.
Either the algorithm badly interact with an other rendering option. (The algo does the rendering 2 times, it expects the 2nd rendering pixel processing to be the same as the first one). Or potentially there is another bug in the driver. I need to replay it on my side.
#8
(10-07-2015, 02:31 PM)gregory Wrote: Ok. Cool, it remains 2 possibilities.
Either the algorithm badly interact with an other rendering option. (The algo does the rendering 2 times, it expects the 2nd rendering pixel processing to be the same as the first one). Or potentially there is another bug in the driver. I need to replay it on my side.

If more testing or GS dumps are needed, I'm willing to help.
#9
(10-07-2015, 04:09 PM).r5 Wrote: If more testing or GS dumps are needed, I'm willing to help.

Well, could you buy me some time credits Tongue2
#10
(10-07-2015, 04:13 PM)gregory Wrote: Well, could you buy me some time credits Tongue2

Err...not sure if my GPU supports that extension, but on the bright side I just updated my NVIDIA driver to revision 358.50 which includes support for these ones:
  1. GL_ARB_ES3_2_compatibility
  2. GL_ARB_fragment_shader_interlock
  3. GL_ARB_gpu_shader_int64
  4. GL_ARB_parallel_shader_compile
  5. GL_ARB_post_depth_coverage
  6. GL_ARB_sample_locations
  7. GL_ARB_shader_atomic_counter_ops
  8. GL_ARB_shader_ballot
  9. GL_ARB_shader_clock
  10. GL_ARB_shader_viewport_layer_array
  11. GL_ARB_sparse_texture2
  12. GL_ARB_sparse_texture_clamp
  13. GL_ARB_texture_filter_minmax




Users browsing this thread: 1 Guest(s)