-------------- Original message ----------------------
From: "Julian Thomas" 

> However, when I removed the card, it appears that it is not the Matrox
> G450 but some form of NVIDIA RIVA.  Still looking for reccos for a
> replacement.  AGP 1.5v.
>  
> -- 
>  Julian Thomas:   jt{at}jt-mj.net    http://jt-mj.net

Hello
  Though many may already know this it might be worthy of mention that
nVidia Riva TNT2 is the last version of gpu/chipset made by nVidia that is
fully supported by free drivers.  There is an IBM driver and also the free
version of scitech both function very well, even in 3D which is still even
by todays' standards fairly respectable with the exception of very modern
3D accellerated games.  However the Matrox cards do have superior 2D
function by a quite noticeable margin and is IMHO a considerably better
choice for OS2/eCS.

  Incidentally, since there has been some discussion recently about total
compatibility as it applies to various sockets in which a few have bemoaned
the "fact" that "gone are the days where if it fits, it
works" and now specifically with AGP slots, I'd like to advance my
opinion that this is still true in most cases.  I've found very few AGP
incompatibilities.  There are really only two main problems of concern 1)
higher voltage  ie-damage possibility and 2) lower voltage ie- no function.
 Most cards are voltage protected, having some kind of VR onboard (as well
as polarity protection) so that damage does not occur in the case of higher
supply voltages.  In the case of lower supply voltages than what the card
requires the mobo manuals are pretty good at defining just how backwardly
compatible that mobo's AGP slot is (eg whether AGPx1 support exists or not)
since the vast majority of cards/slots handle AGPx2 and up quite well. 
There is a slight break at AGPx4 but this affects extremel

y few slots. I have spent  considerable time verifying AGP compatibility
prior to a card purchase (which, granted, was a PITA) for backwards
compatibility of a new card to an old slot but the result was that it
appears AGP, given taking into account the advertised compatibility range,
has very few problems. AFAIK there is little opportunity for such problems
in PCI Express.  Are people here having different results? having
compatibility issues with AGP?  I'd like to know about basic
incompatibility issues if they are firsthand knowledge. Please post
examples here if this thread seems suitable to the group.

  I am aware that there are a few problems, once again with voltage, in CPU
sockets but many of those are overcome by bios updates and this often
applies even between desktop and server systems but  to a lesser degree . 
Unfortunately there seems to be no rule of thumb as to which companies will
pursue continued bios updates and availability and it is much more
difficult these days to swap bioses with those of consistent
IO/chipset/features.  Still, in the majority of cases AFAIK it is still
true that "if it fits it works", at least under some conditions.
Even some chipsets that seem to have difficulty working with ill-matched
RAM at default timings will usually work with less aggressive settings. 
The biggest problems with ram compatbility are dual,interleaved sets that
must match each other as well as the mobo and even this is minimized in the
more stable chipsets and better manufacturers.

  These problems are not some new conspiracy to make things more difficult
since most often a little research can prevent problems.  It certainly
isn't all that new since there have been such problems at least as far back
as during Slot 1 days, when SECC2, early P3 and Xeon processors, came out
that would fit into Slot 1 but not run and P2's would fit in those slots
and not run.  
Jimmy


 
Yahoo! Groups Links

 To visit your group on the web, go to:
    http://groups.yahoo.com/group/os2hardware/

 To unsubscribe from this group, send an email to:
    os2hardware-unsubscribe{at}yahoogroups.com

 Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 




---