This forum uses cookies
This forum makes use of cookies to store your login information if you are registered, and your last visit if you are not. Cookies are small text documents stored on your computer; the cookies set by this forum can only be used on this website and pose no security risk. Cookies on this forum also track the specific topics you have read and when you last read them. Please confirm whether you accept or reject these cookies being set.

A cookie will be stored in your browser regardless of choice to prevent you being asked this question again. You will be able to change your cookie settings at any time using the link in the footer.

Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Release of External Software Rendering for JPCSP
#31
While TOE looks much better, I still see some bugs in the combat scenes. I think a picture is more speaking than a description. I don't know if it is an issue with our renderer or with the game itself though.

   
Reply
#32
(04-28-2014, 09:09 AM)pal1000 Wrote: I decided to give a try to this experimental, brand new full software renderer and ignoring the fact that is terribly slow I found some problems with it. JPCSP crashes and drags Java with it when entering a random battle in FFIV CC (ULUS120560) in the latest JPCSP and software renderer revisions. Also, the text from menus is missing except for the game startup menu.

I attached the game save data and the crash dumps. I started The After Years and played until I got control on the world map.

I used the 32-bit SSE 4.1 variant running with JPCSP r3506 x86 and Java 8 Update 5 x86. I can't use AVX2 being unsupported by my CPU and it is also a 64-bit only variant and thus would require 64-bit Java and JPCSP. It's too much hassle to deal with making the effort unsuitable.
@gid15: we'd better to let dump a vertex decoder binary file with the base address in the name whenever one is generated. That way, if a crash occurs, we'll see if the exception occurs inside and why.
Reply
#33
(05-08-2014, 03:33 PM)onelight Wrote: I try SEE2,SEE3 version , but jpcsp close if 3d models is exist
Also with the cube demo?
If it happens with the cube demo, please provide a log file at the TRACE level. It shouldn't be too large for the cube demo.
If the crash only happens in another application, please try the generate a log file at the DEBUG level...

Thank you!
Always include a complete log file at INFO level in your reports. Thanks! How to post a log
Reply
#34
(05-09-2014, 06:44 AM)gid15 Wrote:
(05-08-2014, 03:33 PM)onelight Wrote: I try SEE2,SEE3 version , but jpcsp close if 3d models is exist
Also with the cube demo?
If it happens with the cube demo, please provide a log file at the TRACE level. It shouldn't be too large for the cube demo.
If the crash only happens in another application, please try the generate a log file at the DEBUG level...

Thank you!
it no happens with the cube demo
test Phantasy Star Portable 2 - NPJH50043
JPCSP close before show character
   
SSE3-log

.zip   SEE3-log.zip (Size: 1.07 MB / Downloads: 123)

noSSE
   

Use Video Resize 1X is triple speed the Video Resize 2X. it can run in 100% speed but too small.
how about use Video Resize 2X but 1X psp resolution.
Reply
#35
(05-08-2014, 03:05 PM)hlide Wrote:
(05-08-2014, 01:24 PM)nickblame Wrote: What am I doing wrong Confused
(previous versions did work)

I have the same issue if I have no external software-ge-renderer dll file in jpcsp/lib/windows-amd64 (only for 64-bit version)

if i put software-ge-renderer.dll or software-ge-renderer-avx2.dll it works.

and be sure not to use the wrong 32/64-bit version.

Okay you were right. Turns out I am completely retarded and this time arround, I was going into the plugins directory instead of the lib -for some mysterious reason- Huh blunder Angry

I get 13-14 fps with AVX and a little less in SSE41 on Tekken 6 during battle. Not bad!!
Reply
#36
(05-09-2014, 06:44 AM)gid15 Wrote:
(05-08-2014, 03:33 PM)onelight Wrote: I try SEE2,SEE3 version , but jpcsp close if 3d models is exist
Also with the cube demo?
If it happens with the cube demo, please provide a log file at the TRACE level. It shouldn't be too large for the cube demo.
If the crash only happens in another application, please try the generate a log file at the DEBUG level...

Thank you!
@gid15: when I build (only tested on 64-bit):
1) without SSE, i have no crash with TOE, 007 and PBP
2) with SSE4.1, i have a crash
3) with AVX2, i have a crash

If i disable jitasm on vertex decoder, those crashes are gone.

The reason is there is a load instruction which tries to use a 64-bit address as a signed 32-bit offset which is alway truncated. As long as this address ranges between 0x00000000 and 0x7FFFFFFF, it works but outside, chance to get an exception. So we'd better to disable jitasm vertex decoder until this issue is gone. This is my fault, i added some functions in jitasm to be able to pass immediate address to a load/store instruction but it happens to be more complicated to do in 64-bit because of that 32-bit offset limitation Sad.

P.S.: oh gosh, I just forgot to answer the right topic: jitasm-ed vertex_decoder was in use when __use_sse__ >= __use_sse2__ when I had a phenom X4 but this decoder now uses "dpps" which is a SSE4.1 intruction !
Reply
#37
Another issue with "Dracula - Undead Awakening (NPEZ00124)" :

   
Reply
#38
Test Gundam vs Gundam Next PLUS - NPJH50107
External Software Rendering fix 16 bit color depth bug, and the speed is good (40% to 60%) if you use 1X video size
External Software Rendering 32 bit color
       
opengl 16 bit color
   

but bring another color bug
External Software Rendering
   
opengl
   
Reply
#39
(06-09-2014, 06:47 AM)onelight Wrote: but bring another color bug
Can you elaborate what the color bug is exactly ?
Reply
#40
(06-10-2014, 12:16 AM)hlide Wrote:
(06-09-2014, 06:47 AM)onelight Wrote: but bring another color bug
Can you elaborate what the color bug is exactly ?

   
Reply


Forum Jump:


Users browsing this thread: 3 Guest(s)