Duke4.net Forums: Polymer won't run in latest, does work with 5149 (Nvidia) - Duke4.net Forums

Jump to content

Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic

Polymer won't run in latest, does work with 5149 (Nvidia)

#1

Just getting back into eDuke after many years, was playing Megaton on Steam from time to time.

Downloaded the latest version, 5312, this works well in polymost but can't run polymer. When I switch to it in the game the game closes. When I check the polymer box in the launcher, the game doesn't launch. If I click fullscreen and polymer in the menu, it launches to a black screen for a few seconds and closes.

I'm using Windows 7 64bit, good recent rig with an Nvidia Geforce GTX 560ti, i5 processor, etc. Haven't added DukePlus or HRP or anything, just those two Duke .grp and .rts files from Megaton on steam.

I also tried version 5307 and it didn't work there either. Before giving up I figured I'd try downloading a random build from a few months ago, 5149, and it does work in that one. Polymer works fine, looks great with the color lightning and everything.

So, reporting the bug there, and if I can ask another question, should I just keep using 5149? Are there any really bad bugs in that version? Thanks. Also, does eduke do a "latest stable build" thing? As opposed to the frequent developer builds.

Here's the debug info from the crash log in the debug win64 version of 5312:

Caught exception 0xC0000005 at 0x0000000000000000: Access violation.
0x0000000000000000 : [unknown module] : [unknown file]
0x00000000005C9207 : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /var/www/synthesis/eduke32/eduke32/build/src/polymer.c (1037) : in function (polymer_loadboard)
0x00000000005C8847 : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /var/www/synthesis/eduke32/eduke32/build/src/polymer.c (790) : in function (polymer_init)
0x0000000000551F07 : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /var/www/synthesis/eduke32/eduke32/build/src/engine.c (11774) : in function (setgamemode)
0x0000000000431FB8 : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /var/www/synthesis/eduke32/eduke32/source/game.c (11799) : in function (app_main)
0x00000000005E1C8D : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /var/www/synthesis/eduke32/eduke32/build/src/sdlayer.c (438) : in function (WinMain)
0x00000000005E1471 : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /var/www/synthesis/eduke32/eduke32/build/src/mingw_main.c (93) : in function (main)
0x00000000004013E5 : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /home/ruben/mingw-w64/src/mingw-w64/mingw-w64-crt/crt/crtexe.c (317) : in function (__tmainCRTStartup)
0x00000000004014E8 : F:\Games\eduke32_win64_debug_20150725-5312\eduke32.debug.exe : /home/ruben/mingw-w64/src/mingw-w64/mingw-w64-crt/crt/crtexe.c (169) : in function (WinMainCRTStartup)
0x00000000771659CD : C:\Windows\system32\kernel32.dll : BaseThreadInitThunk
0x000000007739B981 : C:\Windows\SYSTEM32\ntdll.dll : RtlUserThreadStart
Wed Jul 29 22:28:41 2015
---------------


This post has been edited by PsychoGoatee: 29 July 2015 - 06:33 PM

0

User is offline   Mark 

#2

It was a couple of weeks ago so I don't remember for sure, but I think I ran up to version 5295 or 5298. Anything after that and I crashed with Polymer too.

EDIT: The crash to the desktop starts with 5298 32 bit version and leaves this at the end of my log.

Setting video mode 1680x1050 (32-bpp fullscreen)
Disabling desktop composition...
OpenGL information
NVIDIA Corporation GeForce 9800 GTX+/PCIe/SSE2 3.3.0
Opened "WGR2/textures" as cache file
Initializing Polymer subsystem...
g_errorLineNum: 0, g_tw: 0

EDIT AGAIN: A little OT but in the same WGR2 project, framerate in Mapster dropped. With Polymer running
4593 = 75fps
4777 = 48fps
5297 = 55fps
at map start. I didn't narrow it down to version yet.

This post has been edited by Mark.: 31 July 2015 - 11:15 AM

0

#3

View PostMark., on 29 July 2015 - 06:53 PM, said:

It was a couple of weeks ago so I don't remember for sure, but I think I ran up to version 5295 or 5298. Anything after that and I crashed with Polymer too.

Coincidentally, I just downloaded up the latest High Resolution Pack, which includes 5267 bundled with it. This version works for me too, so far 5149 and 5267 are the two that have worked with polymer for me.

I could give one of those later versions a try too if there are some really key bug fixes.

This post has been edited by PsychoGoatee: 29 July 2015 - 07:18 PM

0

User is offline   Plagman 

  • Former VP of Media Operations

#4

Can both of you run 'glinfo' in the console and attach the full resulting log?
0

User is offline   Mark 

#5

Yup. Hang on.

Since eduke and mapster both crash before the console comes up I can't provide anything. Or do you mean run glinfo on the last working version for me? Here is the log from the last working version.

Attached File(s)



This post has been edited by Mark.: 30 July 2015 - 01:43 PM

0

User is offline   Plagman 

  • Former VP of Media Operations

#6

Run it with the new build, but in Polymost mode.
0

User is offline   Mark 

#7

ok.get back to you shortly.

here it is

Attached File(s)



This post has been edited by Mark.: 30 July 2015 - 05:03 PM

0

#8

Alrighty, here's the log.

Attached File(s)



This post has been edited by PsychoGoatee: 30 July 2015 - 05:21 PM

0

User is offline   Plagman 

  • Former VP of Media Operations

#9

For the 560Ti it should fix itself if you upgrade to the very latest drivers. For the 9800GTX I'm not so sure, but it's worth a shot.
0

User is offline   Mark 

#10

Too bad for me then. I'm already running the latest Vista 32 bit drivers from 2013 for my old card. I'm assuming a later driver will mean nothing for my card. Only newer cards or OS's.

So for now I will go back to version 4593 of Eduke for the best framerate. Thanks for the help.

Is it advisable to run the same version of Eduke and Mapster or could I safely use a newer version of Mapster?

This post has been edited by Mark.: 31 July 2015 - 01:43 PM

0

#11

Similarly, I use the latest version that I can, after that point for some reason new drivers made the card freeze up on my system etc. I've had surprisingly good lucking playing everything coming out on steam etc like Killing Floor 2 at least. Thanks anyway.

That said, I generally play in Polymost anyway.

This post has been edited by PsychoGoatee: 31 July 2015 - 01:59 PM

0

User is offline   Mark 

#12

On the positive side, after the bottoming out of fps with version 4777 things have started to increase with newer versions. But still not to the pre-4593 level. But this may all have to do with my system. Not many people are noticing a slowdown or their newer systems are running so much better than mine so the slowdown doesn't matter.

This post has been edited by Mark.: 31 July 2015 - 02:18 PM

0

User is offline   Mark 

#13

I've got good news and bad news. The good news is I found a newer driver for my old card that I missed before when searching. The latest builds of Eduke32 no longer crash to desktop. The bad news is framerate still sucks compared to the older version from my previous posts. I also checked to see if there was any performance boost with the new drivers on the older Eduke32 version, but sadly no change.

This post has been edited by Mark.: 02 August 2015 - 01:34 PM

0

User is offline   Plagman 

  • Former VP of Media Operations

#14

If you're referring to the other post with the framerate comparisons with and without lighthacks in E1L1, that would account for the difference. The shots depicting high performance have no shadow maps and lights around the map, so you weren't comparing apples to apples. It's possible older versions had a bug with loading lights from maphacks and subsequently ignored them? If you're talking about another framerate comparison, then I can look at that. There wasn't really anything over the past year that should have changed Polymer perf one way or another.
0

User is offline   Mark 

#15

I was doing the majority of my testing with the WGR2 project so maphacks don't enter into the equation. Hendricks did some mojo to batch create a whole string of svn builds and I narrowed it down to a specific version. Then the vanilla maps entered the discussion and things got convoluted even more. But I'm back to the original project where I noticed the fps drop. I'll have to find my notes. IIRC it was between 4599 and 4639. The synthesis repository has a gap between 4593 and 4696 so this puts it smack in the middle of that.

This post has been edited by Mark.: 02 August 2015 - 06:29 PM

0

User is offline   Mark 

#16

Another small test. I turned off lighting from the console using r_pr_lighting 0

4593 = 135 fps
4777 = 85 fps
5312 = 95 fps

This post has been edited by Mark.: 03 August 2015 - 11:36 AM

0

#17

I've now encountered what appears to be the same problem, although I have Intel HD4000 graphics not nvidia (I've seen the HD4000 is used in quite a lot of machines). With r5265 I can do "setrendermode 4" and all is well (well, actually there are a few warnings in the eduke log file). However later versions e.g. r5308 simply bomb out immediately.

Using the debugger, I've found the line causing the error is:

bglBufferStorage(GL_ELEMENT_ARRAY_BUFFER, prindexringsize * sizeof(GLuint), NULL, prindexringmapflags | GL_DYNAMIC_STORAGE_BIT);


bglBufferStorage is null. If I do a glinfo, one of the lines is "Buffer Storage: not supported". Well, it explains that one then. (example log file attached). What am I missing by no longer being able to use mode 4 ?

TTFN,
Jon

Attached File(s)


0

User is offline   Mark 

#18

Technically I have no idea why the crash. I had been running what had been considered the latest drivers for my card in Vista from 2013. After doing some research for "the best drivers for" my card I found a link claiming that driver version xxxx from 2015 works fine. So I installed it and my crash went away. IIRC my debug log still shows that buffer storage not supported. I would confirm that but I reverted back to version 4593 where I get the best framerate.

This post has been edited by Mark.: 12 August 2015 - 11:09 AM

0

User is offline   Mblackwell 

  • Evil Overlord

#19

It's probably slower if glBufferStorage isn't available. But you should really upgrade your GPU since it's 7 years old. I believe with the current updates it basically means that Polymer is GL4.0+ and the 9 series is GL3.3.
0

User is offline   Mark 

#20

So backwards compatibility is being traded for future optimizing? Well for now I'll have to make do with what I have because I would also need to upgrade my Power Supply to run a newer video card. Not really worth it for this old machine. But hopefully it means I'll have a mile wide grin on my face when I get a new system and try out Polymer for the first time on it. :D

Although I just happen to remember that the slowdowns I documented happened before any recent changes to Polymer code. :D PLagman said he did nothing back during that time frame that would account for the slowdown.

But I agree, upgrading can only help.

This post has been edited by Mark.: 12 August 2015 - 04:09 PM

0

#21

View PostMblackwell, on 12 August 2015 - 03:27 PM, said:

It's probably slower if glBufferStorage isn't available.


Almost certainly with older versions of Eduke (e.g. r5265 from early June 2015) where their log files give me:


Initializing Polymer subsystem...
PR : Board loaded.
PR : FBO #1 initialization failed.
PR : FBO #2 initialization failed.
PR : FBO #3 initialization failed.
PR : FBO #4 initialization failed.
PR : FBO #5 initialization failed.
PR : Initialization complete in 281 ms.
Rendering method changed to great justice (Polymer)


Later versions of Eduke just blow up.

View PostMblackwell, on 12 August 2015 - 03:27 PM, said:

I believe with the current updates it basically means that Polymer is GL4.0+ and the 9 series is GL3.3.


It now requires GL4.4+ ; looking at the specs for glBufferStorage this function is not a 'core' function until OpenGL 4.4, which is pretty new

TTFN,
Jon
0

User is offline   Mblackwell 

  • Evil Overlord

#22

Yeah I just was looking into that last night. It means Polymer always blows up on OSX, since the latest driver stack only supports up to 4.1 in most cases.
0

User is offline   J432 

#23

I have lag problems with polymer too plus crashing in latest version. I play BloodCM and with polymost there are ROR graphics glitches, so neither option is good. My card is ATI HD4670 and I'm planning card upgrade, but think it still would be nice if some software solution was found for older card owners.

This post has been edited by J432: 07 September 2015 - 04:52 AM

0

User is offline   Scrooge 

#24

Well, thanks to Plagman in commit 5344, my 4 year-old mobile card is supporting Eduke32 once again in Polymer ! Thank you !
glinfo:
OpenGL information
 NVIDIA Corporation GeForce GT 540M/PCIe/SSE2 4.2.0
 Maximum anisotropy:      16.0
 BGRA textures:           supported
 Non-power-of-2 textures: supported
 Clamp-to-edge:           supported
 Multitexturing:          supported
 Frame Buffer Objects:    supported
 Texure compression:      supported
 Multisampling:           supported
 NVIDIA multisample hint: supported
 ARBfp fragment programs: supported
 Depth textures:          supported
 Shadow textures:         supported
 Rectangle textures:      supported
 env_combine:             supported
 Vertex Buffer Objects:   supported
 Shader Model 4:          supported
 Occlusion queries:       supported
 GLSL:                    supported
 Debug Output:            not supported
 Buffer Storage:          not supported

2

Share this topic:


Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic


All copyrights and trademarks not owned by Voidpoint, LLC are the sole property of their respective owners. Play Ion Fury! ;) © Voidpoint, LLC

Enter your sign in name and password


Sign in options