AOA Forums AOA Forums AOA Forums Folding For Team 45 AOA Files Home Front Page Become an AOA Subscriber! UserCP Calendar Memberlist FAQ Search Forum Home


Go Back   AOA Forums > Hardware > AMD Motherboards & CPUs

AMD Motherboards & CPUs Questions or comments on AMD products?


Reply
 
LinkBack Thread Tools Rate Thread
  #41 (permalink)  
Old 6th March, 2006, 05:47 AM
Gizmo's Avatar
Chief BBS Administrator
BassTeroids Champion, Global Player Champion, Aim & Fire Champion, Puzzle Maniax Champion, Othello Champion, Canyon Glider Champion, Unicycle Challenge Champion, YetiSports 9: Final Spit Champion, Zed Champion
 
Join Date: May 2003
Location: Webb City, Mo
Posts: 16,178
Send a message via ICQ to Gizmo Send a message via AIM to Gizmo Send a message via MSN to Gizmo Send a message via Yahoo to Gizmo Send a message via Skype™ to Gizmo

Quote:
Originally Posted by cadaveca
I also notice, Aedan and Gizmo, that neither of you have had the chance to play with a dualcore cpu, or to play games on one, without the newest version of the AMD driver, which led VERY SPECIFICALLY, to a racing condition in games, more often than not because audio was been offlaoded to the second cpu, and caused a race condition.
The only reference I've been able to find to a race condition which required a change to the AMD driver was related to a HARDWARE BUG in the CPU, which had NOTHING to do with games specifically, nor SLi, nor CrossFire, nor PCI-e. It had to do with the Timestamp Counter (TSC) running at irregular rates, thus causing anything that depended on the timing between CPU cores being accurate to not work properly. If there is something else that the driver fixed with respect to race conditions that was targeted specifically at video games, I cannot find it, and would appreciate a pointer to it.

Quote:
Originally Posted by cadaveca
ANd gizmo, Mr electronics master...why don't you scope the lanes, and see waht going on? No other way can deny or confirm this better.
Somewhere along the line here, I have somehow peed in your Post Toasties. I haven't a clue how or when that happened. All I asked you was if you could provide me with the sources for your information, and I seem to be getting a bunch of attitude about it. I'm sorry, but I'm a little lost at your reaction.
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #42 (permalink)  
Old 6th March, 2006, 09:04 PM
Chief Systems Administrator
 
Join Date: September 2001
Location: Europe
Posts: 13,075

Quote:
Originally Posted by cadaveca
one GPU will get me 97xx in 3dmark for example, while @ the same speeds, 2 gets me 13495. what happened to that rendering power? Gpu's are doing less work for each frame, not more, right?
That *is* the effect of having to synchronise multiple processors of any kind. There is deadtime whilst the processor with the most load completes it's task.

Quote:
Originally Posted by cadaveca
Aedan...when you unconvince yourself that there is only one way to render on SLi and crossfire, we can continue, as what you describe only describes part of one rendering mode, and not all.
No, I did not claim that there was only one way. As you were specifically discussing performance increases and the sending of identical data to multiple GPUs, I think it's fair to assume you were discussing what some call split frame rendering.

Quote:
Originally Posted by cadaveca
In the mode you mention, the driver tag's frame one for GPU0, and frame two for GPU1.
That was not the mode I mentioned. I was specifically discussing SFR rather than alternate frame rendering (AFR), which is what you appear to be discussing in the post you made.

Quote:
Originally Posted by cadaveca
If the cards are not ready for the info, it gets "cached" into what nvidia like to call the "pushbuffer", up to two frames.
I think you'll find that nvidia didn't make up the term pushbuffer.

Quote:
Originally Posted by cadaveca
The chipset notices this flag,takes the data from the pushbuffer(all frames hit the push buffer first, so the chipset can identify them) and sends the data destined for the second card down the second set of lanes...bypasssing the first card, which, by the way, is still electrically connected to those lanes, and onto the second GPU.
If you replaced the word "chipset" with "driver", I might have got half way to agreeing with you. Otherwise your assumption here is blown apart by the fact that SLI works on non-nVidia chipsets. However, at 2.5GHz the PCI-Express lanes can only handle a point to point connection. Putting a 'T' on a PCI-Express lane causes reflections and corrupts data.

Quote:
Originally Posted by cadaveca
Is that doubt?
If you have the time, you can complete the disassembly of the driver. I don't have the time to do such things if you wish to see AOA continue running.

Quote:
Originally Posted by cadaveca
ANd gizmo, Mr electronics master...why don't you scope the lanes, and see waht going on? No other way can deny or confirm this better.
Perhaps you should examine the issues with "scoping" lanes to find out why Gizmo is unlikely to do this before suggesting it?
__________________
Any views, thoughts and opinions are entirely my own. They don't necessarily represent those of my employer (BlackBerry).
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
  #43 (permalink)  
Old 7th March, 2006, 10:23 AM
Chief Systems Administrator
 
Join Date: September 2001
Location: Europe
Posts: 13,075

Quote:
Originally Posted by cadaveca
I'm done. I not sure why i bothered. Don't think i will in the future either.
You're not wrong. This is just stupid. This has been rolling around in my head overnight, so I need to do something about it. So why don't I offer you an apology?

As is obvious from this thread so far, I have been a complete jerk in places. For someone who's supposedly helping run this forum, I would expect better from myself. In some ways, I've been highly hypocritical. Whilst I can't make up for this in retrospect, I'd like to offer an olive branch so we can move forward.

Even though it's not apparent from this thread, I do value your contributions to AOA and overclocking in general. I know that you and I don't always see eye to eye on a range of things. Going forwards, I'd like to be able to use this division as a strength, but I know that this will be something difficult for me personally to acheive.
__________________
Any views, thoughts and opinions are entirely my own. They don't necessarily represent those of my employer (BlackBerry).
Digg this Post!Add Post to del.icio.usBookmark Post in TechnoratiFurl this Post!
Reply With Quote
Reply



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
OCZ launches 1.8 GHz DDR3 danrok General Hardware Discussion 4 20th July, 2007 10:21 PM
AMD launches chipset for Intel chips Gizmo Intel Motherboards & CPUs 16 26th February, 2007 04:43 AM
Who said no 2x16 pci-e? cadaveca AMD Motherboards & CPUs 5 16th June, 2005 08:40 PM
Epox Nvidia Crush 18D Chipset Boards Clocker EPoX MotherBoards 9 31st March, 2003 08:00 PM
New Epox boards based on GE/PE chipset RoadhawK EPoX MotherBoards 1 8th October, 2002 01:05 PM


All times are GMT +1. The time now is 03:28 PM.


Copyright ©2001 - 2010, AOA Forums
Don't Click Here Don't Click Here Either

Search Engine Friendly URLs by vBSEO 3.3.0