Page 3 of 4 FirstFirst 1234 LastLast
Results 21 to 30 of 35

Thread: A Developer's Perspective On Porting Games To Linux

  1. #21
    Join Date
    Jan 2013
    Location
    Ontario, Canada
    Posts
    92

    Default

    Quote Originally Posted by Grogan View Post
    Nothing else does that in Linux or Windows. (But yes, I'm sure it is the drivers allowing that to occur. Give something like a game (more or less) direct access to hardware and all bets are off.)
    Quoting myself, but thinking about this, I can't actually say nothing has ever done that. Call of Duty Ghosts, when it crashes, leads to an unrecoverable lockup if using beta and the current stable release of Catalyst drivers on my hardware. (HD 5870). That's the only thing in Windows here that has ever done that. COD Ghosts (a piece of shit that I'm far, far angrier about than DeadFall Adventures) still crashes/freezes with good Catalyst drivers, leaving my screen frozen (perhaps with a crash handler dialog behind it), but I can just ctrl-alt-del and kill the program with task manager and all is well. I can launch it again and it's good until next time it happens.

    Deadfall Adventures is the only thing that's ever caused me to hard reset in Linux, on this system.

  2. #22
    Join Date
    Jan 2013
    Location
    Ontario, Canada
    Posts
    92

    Default

    Quote Originally Posted by deanjo View Post
    No but hardware suppliers have to provide proper support for their product. Without that you can't expect others to magically fix something that is not even their problem to begin with.

    Hardware suppliers ultimately decide if their product is supported well or not. If you decide to support a hardware vendor with a history of bad support in hopes of things getting better well that's basically a Hail Mary. I wouldn't go out and buy a vehicle with a poor history of service and part availability hoping thing will get better, same applies with pretty much any other consumer product I purchase. If I did buy something on the hopes that the situation will improve, I would consider it a charity donation without expecting any improvement in the short term (or possibly ever) as a direct result of my donation.
    There's no denying that Nvidia is a better supported option on Linux, but that doesn't give people license to write things that only work with the proprietary Nvidia driver and take everyone else's money anyway. I'm not terribly angry about this, I appreciate the Linux ports even if they don't work for me, but I know that things that don't work well on non-nvidia hardware often use nvidia specific hacks. (This phenomenon isn't only confined to Linux... if they don't stick exactly to the directx apis (or use different code paths for ATI specific hacks) it can be rubbish. If Deadfall Adventures was marked as "early access" I missed it. Possible I wasn't paying attention, I was gung ho to buy it specifically to have another nice game to play on Linux. There hasn't been an update for the game in ages though.

    What also cannot be denied is that others manage to do it. I have many games that work well.

    The current radeon + Mesa 10.3-git actually works quite well for many other things. I can't say I've tried everything since the last updates but I'm pretty happy. I did play a few chapters of Metro Last Light without it segfaulting and disappearing. Very smoothly too. I have an impassable point though, where it crashes immediately on loading the next level. Near the end. It's been that way for ages, from fglrx through all iterations of the free radeon drivers and I don't expect it to be any different at this time. We'll see when I get back to that point.

    I don't want to buy another Nvidia card, though. They don't seem to last very long. (I've had poor luck with them... the last three I've owned have all died prematurely. One of them spent most of its life in the closet unused too. I used it for several months then shelved it. I sold it with a system and had to replace it a month later.) Also, this new found freedom of being able to build any kernel I want, any damned day I want, without having to contend with rubbish graphics driver module compilation is almost like being back in the good old days.

  3. #23

    Default

    Anyway, you have to report hard lockup to driver developers.

  4. #24
    Join Date
    Apr 2010
    Posts
    22

    Default

    I don't know if you skipped the slides in the presentation on purpose, or just ignore it for the sake of an argument.

    "AMD acknowledged the hard lockups but has yet to address this issue in the drivers".

    IT is a problem with the drivers. Period. They know about it - as does the game developer. It is also very unfortunate that this bug might be affecting gameplay - but until fixed in the drivers for ATI hardware, you have no other option but to wait.

  5. #25
    Join Date
    Nov 2009
    Location
    Italy
    Posts
    970

    Default

    Hopefully nobody will use Catalyst anymore in a couple of years: http://openbenchmarking.org/result/1...SO-1405083SO83
    ## VGA ##
    AMD: X1950XTX, HD3870, HD5870
    Intel: GMA45, HD3000 (Core i5 2500K)

  6. #26
    Join Date
    Jul 2012
    Posts
    159

    Default

    Use mesa drivers... and see fi compatibility problems will be fixed. AMD support is very inaccurate.

  7. #27
    Join Date
    Jun 2010
    Location
    ฿ 16LDJ6Hrd1oN3nCoFL7BypHSEYL84ca1JR
    Posts
    1,052

    Default

    Quote Originally Posted by Grogan View Post
    See, that's the thing, it's a hard lockup. There is no net... the machine neither responds to ping nor higher level requests.
    netconsole is not high level

  8. #28
    Join Date
    May 2007
    Location
    Third Rock from the Sun
    Posts
    6,584

    Default

    Quote Originally Posted by Grogan View Post
    There's no denying that Nvidia is a better supported option on Linux, but that doesn't give people license to write things that only work with the proprietary Nvidia driver and take everyone else's money anyway.
    Sure it does. There is no reason why a developer should have to remove a feature because of one vendors broken support. Cure the disease instead of treating the symptom.

  9. #29
    Join Date
    Oct 2008
    Posts
    3,176

    Default

    Quote Originally Posted by deanjo View Post
    Sure it does. There is no reason why a developer should have to remove a feature because of one vendors broken support. Cure the disease instead of treating the symptom.
    Broken support for a standard feature is one thing.

    Using nvidia specific hacks that don't work elsewhere and aren't part of any standard is another. That would be like blaming nvidia for lacking mantle support and crashing when a game tried to use it.

    There are different applications that fall under each of those categories.

  10. #30
    Join Date
    May 2007
    Location
    Third Rock from the Sun
    Posts
    6,584

    Default

    Quote Originally Posted by smitty3268 View Post
    Broken support for a standard feature is one thing.

    Using nvidia specific hacks that don't work elsewhere and aren't part of any standard is another. That would be like blaming nvidia for lacking mantle support and crashing when a game tried to use it.

    There are different applications that fall under each of those categories.
    The AMD cards are not crashing because of a "nvidia hack" they are crashing on oGL code, an open standard.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •