If anybodies interested I can post some benchmark results with the P5B Deluxe vs the P5NSLI at the same clocks next week with my lowly 6800GS's ;)
Printable View
If anybodies interested I can post some benchmark results with the P5B Deluxe vs the P5NSLI at the same clocks next week with my lowly 6800GS's ;)
why not... :D ;)
Anybody else come to the conclusion that the low benchmarks in SLI on this board is because of very poor performance in the CPU tests? Particularly '03 and '05
No it's because the two pci-e slots aren't 16x (the second is 4x) and most importantly, the latter is controlled by the sb.
Well this is what lead me to my conclusion
http://24.138.52.57/05compsli.jpg
Top score is my P5NSLI, bottom is the P5B deluxe, exact same clocks. You will note the biggest differences are in the CPU scores. I've noticed stuttering during the CPU tests in both 05 and 03.
The CPU scores return to normal when not running SLI on the P5B
http://24.138.52.57/05comp.jpg
Both P5B, one SLI, one not. Note the CPU scores
can we get some SLI bridge/ no bridge scores on the P5B Deluxe???:confused: ?
I can run some, but I don't think there will be a huge difference with only 6800GS's.Quote:
Originally Posted by TheGoat Eater
Ben Dover?...for shame!
The SLi Vs non-SLI scores are pretty convincing in terms of final value. What drivers were you using? You seem to be getting a much bigger benefit from SLI than Coolaler didQuote:
Originally Posted by Dr_BenD_over
Thanks for the tests :)
What FSB were you running? after 400FSB, the chipset strap changes, that might be part of it
84.56_Mod2, they seemed to be a bit more stable than the 85.96's
Those are totally stock. I can't say I've been impressed by my runs at 2804 (401x7). Almost thought about putting the P5NSLI back in...
Thanks ;) 401 will kill memory bandwidth.. if you feel like it, please could you try both boards below 400FSB? :) 380 or something :)
As for the tests on the P5B SLI Vs non-SLI, I can only guess at whats happening. SB load is being passed to CPU? CPU needs extra cycles when SLI is enabled, at least it did in the good ol days
-> just tried to use my PCI-E 4x for some bench runs... damn SATA ports get in the way, so nowt happening here tonight except for thinking :(
Haven't had a chance to do any unbridged tests yet. Whatever the glich is it doesn't appear to affect '06 very much with 6800GS's. My 5717 run last night pretty much shows that.
I had 2x7600GT and got around ~5100 '06 marks on P5B Deluxe. then I've added the SLI bridge, and got ~6500. might be a reason to see slow SLI results because no bridge used ?
6800GS SLI rips 7900GTX SLI on this board (comparing Dr Bens results to Coolalers)
possibilities (the way I see it)
*driver differences
*the lower spec of the 6800 is not eating up so much bandwidth, especially on the 4X slot so there is no bottlenecking, although there is a CPU hit
Any input? I`d love to see Coolalers opinion on this :)
@Dr Ben...sorry to ask more of you, but if you get a chance please can you try the other 3DMarks? :)
That was my hope when I got the board. I'd kick myself for doing so, if only the 680i boards weren't so expensive :nono:Quote:
Originally Posted by K404
Anyway, I'll try some other benches this weekend (I hope).
4858 = bridged
4626 = unbridged
http://173.212.117.111/unvssbridge06_2.jpg
3.2Ghz, cards at 425/1000 (stock)
How did you get the 9114 score in `06? (Post 105) Thanks for trying the extra benches :) Numbers look about right for full SLI results.
That was an '05 runQuote:
Originally Posted by K404