I came back to this thread wanting to post my reply but i can see that michael has already illustrated my point making this post pratically useless, so I will refrain from further spamming of this thread after this last post
..
Quote:With lower expectations comes less nagging. And what of ATI?
we seem to be coming back to the issue of dedicated graphics vs integrated graphics again, i would presume that intel would have atleast the same level of nagging as nvidia as the currently own 90% of the graphics industry throught their integrated chipsets, while nvidia continues to occupy the dedicated space. But maybe this world is guarded by other laws that are unbeknown to us, anyway its all speculation. I must admit though that your continued refusal to acknowledge nvidias failings in this space is quite impressive, I really hope they're paying you well..
As for ATI, I choose to focus my second post on your criticism of intel, but I would stille say that ATI drivers are progressing on the issue, whereas nvidia is standing still.
Quote:So shame on NVIDIA for not supporting MPEG decoding hardware acceleration. If only CPUs were fast enough to do this...
Why disable working functions on excisting hardware? Sure modern cpu's can keep up, but what about people using older hardware, and why waste cpu cycles on something that can already be handled else where? Seems nvidia only cares about the latest and the greatest, it's just and odd thing to do if you ask me.
Quote:Yes, he had to "resort" to 2.0 shaders because, surprise surprise, there is no API out there that will do it. Not even from Intel! Also, because we like things to just work here at XBMC, it wouldn't be good enough if NVIDIA supported something and Intel didn't.
Quite true, one can only dream of a cross platform unified video acceleration api, but it's not very likely to happen.
Quote:XBMC is completely unusable on some Intel chips if you enable VSYNC. Can't navigate the UI, can't play a video.
Yes and so is a geforceFX5200, same thing cpu goes to 100% every things frozen or incredibly slow and video results in a horrible green mess. Cant really see what the difference here is, older nvidia card that don't support the openGL 2.0 ARB spec, seems this is another case of comparing nvidias latest and greatest to intels legacy chipsets..
Quote:OpenGL says so. At least last I checked. And from wikipedia: "In August 2006, Intel added support to the open-source X.Org/XFree86 drivers for the latest 965 series that include the GMA (X)3000 core.[40] These drivers were developed for Intel by Tungsten Graphics." FOR Intel BY Not Intel.
Wow this is retarded, I just pointed to an Intel page with the names of developers currently employed by Intel to develop linux drivers, and you give me an old story from wikipedia (that modern bastion of truth). Incase theres any doubt checkout the xf86 developemt changelog and compare it with what wikipedia and intel are saying:
http://gitweb.freedesktop.org/?p=xorg/dr...bfa03d0bf3
And by the way what is wrong with hiring and external firm to develop an initial driver release? Many companies do this in open source, even ATI and this doesn't seem to stop them from producing functioning drivers.
Quote:I meant having to put a hack into XBMC. Are Intel's drivers OS on Windows also? If not, we can't just fix the driver as the behavior would still be wrong in Windows/OS X. Open source is wonderful, I support it, XBMC supports it, but it sometimes means shit. If it's not open source on all platforms, it's worthless. How receptive is the team to external patches? If they don't take them and integrate them, it's worthless.
As far as i know team XBMC has already put patches for nvidia driver errors in XBMC's codebase (the geforceFX5200 error being one of them). Reality is every driver manufacturer has errors, what matters to the platform is the response time between releases and lately nvidia has neglected linux as shown in previous posts. From a multiplatform-developers standpoint your comments about open source make sense, but from a (linux) user perspective, it doesnt, windows has and always will be the home of closed source software so your basically asking for the impossible. As far as intels responses to patches I cant speak, but their linux driver development is integrated into the main xorg development tree, so they follow their bug-reporting / patch application process, so one can atleast guess it would rival that of team-xbmc.
Quote:No, I actually recommend integrated graphics from NVIDIA. My problem stems from that VSYNC issue.
Hmm 5 seconds on google revealed that the issue dissapeard with the 2.2.1 driver:
"http://www.phoronix.com/scan.php?page=news_item&px=NjI5NA
There may be hope with Intel X 2.2.1 Driver
In Ubuntu 8.04 this problem is gone :-D (Tested with 32bit)"
Quote:As far as XBMC is concerned, there's no point to open source graphics unless they're open on all platforms. AFAIK, Intel hasn't opened its drivers for Windows and OS X, so we can't do anything with the Linux drivers either. Fixes have to go into XBMC.
Again that maybe true from a developers standpoint but not really from the users, especially when there's difference in linux support from nvidia and intel.
Quote:Bleak outlook? Are they not able to play videos and show music visualizations? Are the viz and videos slower than they are on Intel? Less detailed?
Once again, ATI is left out of the conversation.
No currently everything works in xbmc as it does on intel, but as I have demonstrated nvidias falling behind on their linux support, so as a linux user you might want to choose your product accordingly. And yes ATI was left out of this discussion for reasons noted earlier