Video Card Bugs

[Unbroken]

Upon first execution of the game .exe, I was awarded with a large crash.dmp file, found here. Appears to be the same issue Sven outlined in the video card bugs thread, as I have an Intel HD 4000 integrated card in my laptop.

When I ran the .exe with admin privileges, it suddenly works. A third time without admin privs and it still works. The only thing I did between the first and subsequent tries was to run the updater briefly and allow it to send logs, etc.

[sven]

Upon first execution of the game .exe, I was awarded with a large crash.dmp file, found here. Appears to be the same issue Sven outlined in the video card bugs thread, as I have an Intel HD 4000 integrated card in my laptop.

Many thanks for the upload. I'll take a look at that tomorrow morning.

When I ran the .exe with admin privileges, it suddenly works. A third time without admin privs and it still works. The only thing I did between the first and subsequent tries was to run the updater briefly and allow it to send logs, etc.

Interesting. If you run it now, without admin privileges, does it still work? (There's a chance that the crash-on-launch may be something that only happens, sometimes, the first time the game runs after a clean install. Frustratingly, we have a number of Intel HD 4000s in the beta, but, reports of this crash have been very rare.)

[Unbroken]

Interesting. If you run it now, without admin privileges, does it still work? (There's a chance that the crash-on-launch may be something that only happens, sometimes, the first time the game runs after a clean install. Frustratingly, we have a number of Intel HD 4000s in the beta, but, reports of this crash have been very rare.)

I played the game briefly to start seeing how everything works and quit a few minutes later. Tried to run it again now and it crashed. Here's the windows appcrash data:

Code: Select all

Problem signature:
  Problem Event Name:   APPCRASH
  Application Name:   sis.exe
  Application Version:   0.0.0.0
  Application Timestamp:   55d37bce
  Fault Module Name:   ig7icd32.dll
  Fault Module Version:   9.17.10.2867
  Fault Module Timestamp:   50635085
  Exception Code:   c0000005
  Exception Offset:   0033ea9a
  OS Version:   6.1.7601.2.1.0.768.3
  Locale ID:   4105
  Additional Information 1:   0a9e
  Additional Information 2:   0a9e372d3b4ad19135b953a78882e789
  Additional Information 3:   0a9e
  Additional Information 4:   0a9e372d3b4ad19135b953a78882e789

Right after that, I re-ran the game as admin, and it's working.

[sven]

Right after that, I re-ran the game as admin, and it's working.

Hrm. Well, I looked at the crash, at it does seem to be an access violation while trying to generate the fancy dynamic textures used in the planet renderings. It's plausible that running in admin mode is giving you a slightly different memory use policy, which, might begin to explain why that's helping.

At this point, my basic hunch is that on machines with no dedicated video memory, I really should be using lower-res versions of most of the art assets. And that should be easy enough to do.

[sven]

Right after that, I re-ran the game as admin, and it's working.

Hrm. Well, I looked at the crash, at it does seem to be an access violation while trying to generate the fancy dynamic textures used in the planet renderings. It's plausible that running in admin mode is giving you a slightly different memory use policy, which, might begin to explain why that's helping.

At this point, my basic hunch is that on machines with no dedicated video memory, I really should be using lower-res versions of most of the art assets. And that should be easy enough to do.

That said, I just took another look at your logs, and if I'm reading these right, you're using a fairly old version of the HD 4000 driver. Could you try updating it, and see if that avoids the crash? (I'd suggest using either the official Intel driver update utility, or checking your laptop manufacturer's support page.)

[Unbroken]

That said, I just took another look at your logs, and if I'm reading these right, you're using a fairly old version of the HD 4000 driver. Could you try updating it, and see if that avoids the crash? (I'd suggest using either the official Intel driver update utility, or checking your laptop manufacturer's support page.)

I can try. I recall trying to update almost all my drivers a few months ago and I thought they were the latest versions.

I also had the game devour all my memory. When it finally ran out after about 45min of play, it was using 1.7 gigs of memory according to the task manager. I'd happily upload crash.dmp, except that it too is about 1.7 gigs in size - I unfortunately lack the time as of this post to upload the monstrosity. When it crashed, I was running through a few fights against some slavers and utilizing the old MOO2 trick of point-blank missile spam to achieve easy wins.

Here's what Windows had to say:

Problem signature:
Problem Event Name: APPCRASH
Application Name: sis.exe
Application Version: 0.0.0.0
Application Timestamp: 55d37bce
Fault Module Name: KERNELBASE.dll
Fault Module Version: 6.1.7601.18933
Fault Module Timestamp: 55a69ec4
Exception Code: 00004667
Exception Offset: 0000c42d
OS Version: 6.1.7601.2.1.0.768.3
Locale ID: 4105
Additional Information 1: 0a9e
Additional Information 2: 0a9e372d3b4ad19135b953a78882e789
Additional Information 3: 0a9e
Additional Information 4: 0a9e372d3b4ad19135b953a78882e789

[sven]

I also had the game devour all my memory. When it finally ran out after about 45min of play, it was using 1.7 gigs of memory according to the task manager. I'd happily upload crash.dmp, except that it too is about 1.7 gigs in size - I unfortunately lack the time as of this post to upload the monstrosity.

This looks very much like the behavior we've seen on an Intel HD 3000 driver [1]. We've seen signs of a memory leak on the HD 4600 as well -- so in some ways, this is expected. What has me scratching my head is why some of the other HD 4000 beta testers *aren't* seeing it. Let me know if you have any luck with those driver updates. From what I can see in my logs, the date stamp on your graphics drivers is 2012 -- but, if they're locked to an old manufacturer spec, there may not be a lot that can be done about that.

[Unbroken]

According to the update utility, my display drivers can't be updated to the newer generic version as they've originally been customized by the manufacturer (Dell in this case).

As far as SiS' memory consumption goes, after I load the game it already uses 800 megs of memory. Whenever the game needs to call up a lot of new textures (such as upon meeting a new race), it'll freeze for a couple seconds as it places all of them into memory. I'll try to keep a closer eye on exactly what actions cause the memory usage to balloon out of control.

[Arioch]

I would suggest updating the drivers directly, rather than relying on Dell's update utility. In my experience Dell does a very poor job of keeping its list of "approved" drivers up to date.

I like Dell computers, and I'm on my third Dell, but their custom software is garbage.

[Unholyfroggod]

Only been playing for an hour or so but I'm clicking in and out of menus etc like crazy to try to cause errors. Good news is I have a Radeon HD 5870 1Gig in a desktop and I have found NO errors or glitches yet
Will keep playing (as soon as my daughter is done with the computer)

[Arioch]

Great to hear!

[Mithramuse]

Hi there! New tester here... I have the Intel HD Graphics 4000, and have now run out of memory twice (crash to desktop), and had one in-game error that was almost certainly a planet texture (the planet "spot" was a big black square, and the error cleared when I moved the focus away from the system with that planet). The in-game issue uploaded... but I don't remember the game name (836? Sorry, will do better!). The first crash failed to create a dump log, but the second has; I'll get that uploaded somewhere if you do want to look at it, but reading back in this thread it sounds like this is a bit lower in priority at the moment for you.

I'm running Win7 with HD4000 drivers from 2/1/2012 (version 8.15.10.2639); will wait on poking for upgrades in case you want me to try and force more errors (I also need to go back and poke that star system again, that generated the black square texture, to make sure it'll happen again).

Looking good all in all. Will try to do more to hunt bugs and the like, instead of just oooooh'ing and aaaah'ing my way around.

--Mithramuse

[sven]

I'm running Win7 with HD4000 drivers from 2/1/2012 (version 8.15.10.2639); will wait on poking for upgrades in case you want me to try and force more errors (I also need to go back and poke that star system again, that generated the black square texture, to make sure it'll happen again).

Thanks for the report. It's certainly a known issue -- and right now, we have at least 3 active testers who are playing through these bugs. I think they're all essentially the same root issue. Not quite sure when I'll have time to put together a comprehensive fix -- right now, my best guess on an ETA would be "sometime after we do the diplomacy system, but, before we do a Steam EA".

edited by sven: Though it's not a perfect solution, we're certainly seeing better stability on Intel HD Graphics cards running with the 64-bit sis build. Whether this is because the memory leak is just less likely to be fatal in a larger virtual memory pool, or because there's actually a bug in the 32-bit Intel OpenGL drivers remains a bit of a mystery. (Right now, the numbers I'm seeing suggest that the drivers *do* have a bug in 32-bit mode, but that even without the bug, SiS is using enough virtual memory on Intel HD Graphics systems that running in 64-bit mode is probably helpful, regardless.)

Looking good all in all. Will try to do more to hunt bugs and the like, instead of just oooooh'ing and aaaah'ing my way around.

Glad to hear you've been having fun

[SirDamnALot]

The welcome screen (the one with the big planets spinning and "new game" option) makes my GPU goes bananas with workload.
Rest of the game seems fine (galaxy map, tactical map) but this screen gives me near constant 90% load.

gpu_load.jpg (14.27 KiB) Viewed 60224 times

Blue: ingame, Purple: Welcome screen/Main Menu
Memory Load increases also, but the overall load level is far less.
I'll keep an eye out for other occations that happens. Maybe the framerate is running rampant and not capped at the monitor refresh rate (60hz)?
Fullscreen seems a little worse than windowed, to no surprise

System specs:
OS: Microsoft Windows 7 Professional
CPU: Intel(R) Core(TM) i7-3770K CPU @ 3.50GHz, 3801 MHz, 4 Kern(e), 8 logische(r) Prozessor(en)
GPU: AMD Radeon R9 200 Series (Sapphire Vapor-X R9 290 OC)
Mem: 16 GB

[sven]

The welcome screen (the one with the big planets spinning and "new game" option) makes my GPU goes bananas with workload.
Rest of the game seems fine (galaxy map, tactical map) but this screen gives me near constant 90% load.

Interesting. My own Geforce GTX 460 typically has a < 30% load on the start screen, and it's a considerably wimpier piece of hardware. Something odd is certainly happening here. My hunch would be that the issue is caused by some sort of interaction with planet shader code.

Would you try selecting a planet in game then hitting ctrl+Y. That should jump you into a test screen that shows only a rendering of the selected planet. If the planet shader is causing excessive load on your card, the gpu stats while this screen is open would prove it. (Note that different planets may have varying shader costs, a ringed world, like Tendao or Rastaban II, is probably a good "worst case" test.)

Thanks again. (If it's not the planet shader, my next guess would be the sun shader -- I don't have a hotkey setup to test that, but, historically, there have been a lot of odd behaviors with the animated suns on AMD hardware.)