On Tue, Jun 25, 2002 at 09:41:16PM +0100, Sergey V. Udaltsov wrote:
>
>> Strange... When the 3 * screen size is checked an error message is produced and DRI
>> is given as disabled in the X log.
>Sorry for misinformation. After detailed investigation, I found the
>complain. It just was not prefixed with [drm]:))) So I need 13440K:(((((
>
>> Even yesterday Leif and I briefly discussed this on the IRC meeting and
>> unfortunately ther seem to be some limitations in Mach64 itself
>:((( Why can't you say something nice sometimes?
>> regarding this, i.e., as far as we know it's not possible to put none of
>> the front back and depth buffer on AGP. So unless we dismiss the back
>> buffer I don't see how this restrition will go away. (I don't know how
>> Windows copes with this neither - it's something I still have to check).
>Is there a way to check in with Windows drivers? It would be very

As I said before I still have to investigate. I can't give more answers
until I actually do it. (PS: reboot my machine to windows is not
something I do often or even like to do..)

>interesting... Well, about "back buffer" - what would be the penalty of
>dismissing it?
>

No double buffering, i.e., you would see stuff getting drawed over the
previous frame. If the fps are high, that might get unnoticed, but not
otherwise.

>> >> Yes. Either start X with gdb or attach gdb after X starts but before
>> >> changing resolution from a remote terminal, e.g.:
>> >> Then reproduce the segfault, i.e., change the resolution in this case,
>> >> the gdb command line should reapper. Type 'bt' and post the result.
>> >OK. I will try.
>At the moment, I don't have network and second computer but I managed to
>find in X log - crash is caused by signal 11. And the situation when it
>appears a bit strange. I can safely switch VTs when I run just "X" from
>VT1. Even from login screen of gdm I can switch to VT1. But after I log
>in into gdm - after this point switching to VT1 causes signal 11. Well,
>tomorrow I'll try to use gdb remotely... About changing resolutions:
>well, I can do it in most cases. But when vmware changes the resolution
>(in full screen mode - AFAIK it does it using DGA, isn't it?) - X
>crashes the same way.
>
>> Yes, there are some situations, but they don't depend on the available
>> memory and/or screen resolution. So if is this what you were talking
>> about then the difference in fps come solely from less texture trashing.
>No, I mean they depend on using different GL effects (anti-aliasing,
>multi-texturing etc). So in some cases I have HW 3d (and it is
>reasonably fast) - but in some "bad" cases glclock switches to SW - and
>goes VERY slow.
>
>BTW, playing with different resolutions (Using Ctl-Alt-+/-) I found that
>fps in glxgears really depend on resolution. Not too much but still:
>800x600 - 267
>1024x768 - 259
>1280x1024 - 251
>Same size, 16bpp. A bit funny, isn't it?

This phenomenon was already discussed once here. It has to do with
competition over the video memory bandwith between the GPU and the DAC.

Jos� Fonseca


-------------------------------------------------------
This sf.net email is sponsored by: Jabber Inc.
Don't miss the IM event of the season | Special offer for OSDN members! 
JabConf 2002, Aug. 20-22, Keystone, CO http://www.jabberconf.com/osdn
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel

Reply via email to