[Forum] Articia: Still DMA-problems? | ANN.lu |
Posted on 20-May-2003 11:16 GMT by Senex | 122 comments View flat View list |
From a news-item at VGR.com it appears that the current Articia-northbridges might still have DMA-related problems:
"Mr. S.W. who sends over screen shots of newer AmigaOne 800mhz G4 hardware, has sent over 2 screen shots of the new A1 with G4 running 'amigaone xe/g4 7451 800mhz, excalibur radeon 9000 pro 64mb ddr ram, valueram kvr133x72rc3/256 kingston'. He wants to know '.. if other aone users have that problem as well..', 'it never outputs the same checksum with dma enabled when testing an iso file which is larger than the swap space used'
You can see the Radeon 9000 Pro screen shots here: DMA On & DMA Off."
|
|
List of all comments to this article |
Articia: Still DMA-problems? : Comment 112 of 122 | ANN.lu |
Posted by Mayonaise on 22-May-2003 20:09 GMT | In reply to Comment 107 (Nate Downes): "
The same bug exists on both the pre-April'd Pegasos and the current
AmigaONE's. These two motherboards run different southbridges
(admittedly, from the same manufacturer, VIA). These two southbridges
have different bugs, which are well documented and are not alike in
mannerism. "
VIA IIRC use a multi core production process(I forget the name, but its used in a bunch of 8051 network/secure mcu's I've got lying around), anyways they basically implant various interchangeable modules into the package or design to create a chip for that situation.. Thus its quite possible the chip-set used in the pegasos shares faulty components from the 686B.
Its also interesting that the same problems here were believed to be in the north bridges(KT133 and Apollo chips took the brunt of the blame) on various PC motherboards, but were later traced to the VIA south bridge. I've lost a site detailing the VIA problems in great depth (Logic probes,Direct tinkering with the north/south bridges etc :)) but VIA is reported to have released a data sheet detailing the problems in the 82c* south bridges and the required software fixes some time ago and I believe this is what Alan Cox referred to in the Linux changelogs as "The offical VIA fix".
Turning the VIA option off in the linux kernel config here(x86) produces bad MD5Sums and even worse, some nice random crashing :| |
|
List of all comments to this article (continued) |
|
- User Menu
-
- About ANN archives
- The ANN archives is powered by #AmigaZeux. It was updated daily (news last: 22-Oct-2004; comments last: 18-May-2005).
ANN.lu was created, previously owned and maintained by Christian Kemp, www.ckemp.com.
- Contribute
- Not possible at this time!
- Search ANN archives
- Advanced search
- Hosting
- ANN.lu was hosted by Dreamhost. Sign up through this link, mention "ckemp" as referrer and he will get a 10% commission on any account you purchase.
Please show your appreciation for any past, present and future work on ANN.lu by making a contribution via PayPal.
|