|
CrystalCowboy
Ace
Posts: 1,648
Join Date: May, 2004
|
 |
« on: April 09, 2008, 03:07:27 PM » |
|
Let's hope this one is more substantial than their past efforts:
Linux-VIA love match announced
VIA HAS HERALDED its strong desire to collaborate with the open sores community with the announcement of its official VIA Linux Website due to open later this month...
Maybe they see the Intel Atom coming to eat their lunch, and want to carve out a refuge. |
|
|
|
|
Logged
|
|
|
|
|
Reflex
Ace
Posts: 11,331
Join Date: Dec, 2001
|
|
« Reply #1 on: April 09, 2008, 06:19:09 PM » |
|
I don't see Atom eating their lunch, chances are it will end up in completely different types of devices. After all its an old school in-order CPU and Via sells modern out of order CPU's.
|
|
|
|
|
Logged
|
|
|
|
|
Zut50
Ace
Posts: 4,265
Join Date: Nov, 2003
|
|
« Reply #2 on: April 09, 2008, 10:22:26 PM » |
|
VIA open source software is utterly awful.
The community asked them for a driver to support the popular media player Xine. Instead they created their own version against an old branch of Xine and artificially crippled it. I had to write a patch to enable features that they had literally just commented out. In the end the OSS community overtook them by creating proper drivers and plugins and now no-one (AFAIK) uses VIA's cripled rubbish.
Frankly I dont trust them.
|
|
|
|
|
Logged
|
|
|
|
|
MrbLOB9000
Ace
Posts: 12,896
Join Date: Apr, 2002
|
|
« Reply #3 on: April 10, 2008, 12:59:08 AM » |
|
I just checked while talking to reflex about it, the currently released Via CPUs are in order, but the soon to be released Isaiah CPU will be out of order, check out info here and I link to a few places with more info from my post as well http://aberrantech.com/?p=461 |
|
|
|
|
Logged
|
|
|
|
|
MrbLOB9000
Ace
Posts: 12,896
Join Date: Apr, 2002
|
|
« Reply #4 on: April 10, 2008, 01:00:35 AM » |
|
VIA open source software is utterly awful.
The community asked them for a driver to support the popular media player Xine. Instead they created their own version against an old branch of Xine and artificially crippled it. I had to write a patch to enable features that they had literally just commented out. In the end the OSS community overtook them by creating proper drivers and plugins and now no-one (AFAIK) uses VIA's cripled rubbish.
Frankly I dont trust them. didn't all that happen like 4 or 5 years ago? that's a lot of time for a genuine change of heart to happen and Linux is way more mainstream now than it was 4 or 5 years ago (though the word mainstream may be overstating it). |
|
|
|
|
Logged
|
|
|
|
|
Reflex
Ace
Posts: 11,331
Join Date: Dec, 2001
|
|
« Reply #5 on: April 10, 2008, 01:08:08 AM » |
|
Suprised me, although it goes a great distance towards explaining the poor performance of Via offerings(not that it matters much in their target market).
|
|
|
|
|
Logged
|
|
|
|
|
|
|
Hlafordlaes
Veteran
Posts: 402
Join Date: May, 2006
|
|
« Reply #7 on: April 27, 2008, 12:01:56 AM » |
|
As a VIA chipset junkie, I have found it challenging and entertaining to discover how to get VIA boards to perform as well as they can. Eg, apart from the well-known memory tweaks on the old Apollo Pro chipsets, using chipset data from VIA I was able to uncover a very conservative setting in the southbridge that, once tweaked, seriously improved IDE performance. My Maxtor 160GB drives were faster in my VP6 than they are on the newer PT880 chipset.
But VIA waits centuries before releasing chipset data, and it does not seem that the open source effort has changed this much (granted, it is directed more toward the mini-itx boards). I am convinced that VIA would see a much greater appreciation of their products if they were more forthcoming. Bios settings on PT880 boards include a slew of advanced memory tweaking options... that no one anywhere has been able to crack since we have no datasheets (I have tried Fiona Gatt directly, as well as getting more reputable people from Anandtech, VR-Zone and Madshrimps to try as well, all to no avail so far).
So until VIA really understands partnering with the user and open source communities, efforts such as open source drivers are likely to be less successful than they could be.
|
|
|
|
|
Logged
|
|
|
|
|
MrbLOB9000
Ace
Posts: 12,896
Join Date: Apr, 2002
|
|
« Reply #8 on: April 27, 2008, 01:19:47 AM » |
|
does anyone release that info though?
|
|
|
|
|
Logged
|
|
|
|
|
Connor
Ace
Posts: 3,192
Join Date: Nov, 2002
|
|
« Reply #9 on: April 28, 2008, 12:24:10 AM » |
|
But VIA waits centuries before releasing chipset data, and it does not seem that the open source effort has changed this much (granted, it is directed more toward the mini-itx boards). I am convinced that VIA would see a much greater appreciation of their products if they were more forthcoming. Bios settings on PT880 boards include a slew of advanced memory tweaking options... that no one anywhere has been able to crack since we have no datasheets (I have tried Fiona Gatt directly, as well as getting more reputable people from Anandtech, VR-Zone and Madshrimps to try as well, all to no avail so far).
So until VIA really understands partnering with the user and open source communities, efforts such as open source drivers are likely to be less successful than they could be. I couldn't agree more, on the original epia board I found a bug in the Video Drivers. It only affected one Game Engine that I know of but I'm not sure how many actual games were affected. The same problem had been reported for another video chip: the Voodoo3 and 3DFX had been forced to rewrite their drivers to fix it. I tested with the Trident reference drivers for the graphics chip and was successful but lost out on other video features that were included with the VIA driver (particularly TV-Out) Like you I contacted John and Fiona Gatt to see if a driver fix could be released - without success the chipset was being depreciated. I asked them for Driver Source Code no response - there were open source drivers (not made by VIA) for the chipset but were not directly applicable to repairing the windows drivers and again did not support all video features. This from a company whose idea of a support process was to set up a forum for unpaid customers to solve the problems of other customers - but does not give the experienced users the tools or the incentive to do so. All the while John and Fiona lurk in the background running competitions and giving out management platitudes - but rarely helping anyone. Fed up - especially with the way that certain trolling posters were being given far to much leeway by john gatt because he knew them personally- I took action that saw me permanantly banned for far less than some of the trolls were comitting. I agree with Zut that VIA cannot be trusted when they say they are eager to work with the open source community. |
|
|
|
|
Logged
|
|
|
|
|
Hlafordlaes
Veteran
Posts: 402
Join Date: May, 2006
|
|
« Reply #10 on: April 28, 2008, 02:58:56 AM » |
|
does anyone release that info though? Intel does, including for chipsets as recent as the 975X. That's pretty good, and no need to send in any requests to get it, either. |
|
|
|
|
Logged
|
|
|
|
|
JQPublic
Member
Posts: 530
Join Date: Mar, 2004
|
|
« Reply #11 on: May 12, 2008, 06:35:42 PM » |
|
|
|
|
|
|
Logged
|
|
|
|
|
CrystalCowboy
Ace
Posts: 1,648
Join Date: May, 2004
|
|
« Reply #12 on: May 14, 2008, 05:09:12 PM » |
|
From your link:
VIA has released over 16,000 lines of code that provides a frame-buffer driver in the Linux kernel. This code is licensed under the GNU GPLv2 and appears to be crafted by VIA's Joseph Chan. Supported by this driver is VIA's Unichrome CLE266, K400, K800, PM800, CN700, CX700, K8M890, P4M890, P4M900, and VX800 IGPs. We're still pouring over the code, but it seems to be in pretty good shape and does support digital connections (and does seem to support HDMI already) -- in other words it appears to be further along than when the RadeonHD driver started out.
That's nice to hear. |
|
|
|
|
Logged
|
|
|
|
|
