Originally Posted by Pedros
When reviewing a GTX580 Direct CU2 in SLi, we were using an Xeon W3520 at 4.2Ghz and we were scoring 44k on the GPU department ... after we changed the bench table to a SB 2600k at 4.7Ghz ... the gpu scores ramped to 47k! ...
Yeah, in 3D Mark Vantage the cpu plays a big factor, because even sandy bridge will bottleneck in the gpu parts of the testing, the difference then becomes how much of a bottleneck it creates.
Originally Posted by Catscratch
5217 MHz vs 4304 MHz
9943 vs 9474
I wonder what score you'd get if you also were at 4304 mhz. And what about at 1080p ?
4.3GHz is pretty close to Phenom II max on water, I see no problem with benching SB at 5.2Ghz when I could have gone higher. I also have a problem with my motherboard, I can only run my ram in single channel mode as the first two dimms simply do not work. That causes a 50% loss in my memory speed, but I didn't think it was important enough when comparing two platforms. The Phenom II has 6 cores, SB has 4, nobody thinks it would be a good idea to disable two cores to compare chips do they? One chips limitations shouldn't bring another chip down to its level.
My board also only runs x8 x8 x4 his was running x16 x16, again nobody would ask him to lower his bandwidth to his gpus to make the testing more fair would they?
Originally Posted by LastBucsfan
Balla, is it possible that all the other variables in the systems make this comparison invalid? Just like an old post of yours I was just reading comparing your own 1090T to your SB... Maybe the problem was that you ran DDR2 with the 1090T and a really high end set of DDR3 with the SB?
I already know that even in more fair comparisons with identical systems that the i5-2500k will outperform the 1090T in benchmarks. Don't forget this is Intels latest architecture vs 3-year old AMD architecture so it should be expected. I don't think that there's enough information here to call the AMD chip a "bottleneck" for any particular GPU though. There are just too many other variables involved.
As far as price, there's more to it then just initial cost. Did you hear that I'm going to be able to drop Bulldozer into my current motherboard? =P
My ram ran just slightly below 10k speeds which put it on slightly below high end DDR3 on the Phenom II northbridge, the DDR2 to DDR3 for AMD at this time isn't a huge upgrade, going from 14k speeds on DDR3 for AMD to near 30k on SB would be a much larger impact than DDR2 to DDR3. If you remember right I ran it 5-5-5-15 2T @ 1200 MHz. I also ran my NB around 3.2GHz.
People are still buying cpus, this is just an info post. I've heard about AM3 boards supporting AM3+ chips, I've also heard AMD directly state several times that they would not support this. We'll have to see the results of such a match up before we consider it a feature or a failure.
Originally Posted by sleepergsr
@ balla. when you compare both cards..make sure you put the cards value at retail...just because you bought your card used doesnt mean you can price it at how much you got it. you should price it as how much they usually go for online or at a store front value. its like saying i got my card for 180 bucks and they usually go for 300 bucks. =) those gtx 470 are pretty much similar price to the 6950. clearly tells me that the 6950 are faster then the gtx 470 as well too. =) from my experience with the 470...i notice some benchmark favors phyx. for some reason..it does boost cpu score on benchmark due to phyx but in actually game play..my 5870 beats the 470.
I thought it was pretty close, I bought my ASUS card new for $185. But to be fair newegg currently has 470s for $210 so $420 (lol) for my cards.
Phsyx was turned off, or the feature was in 3D Mark Vantage, my 9800GT is dedicated phsyx, if you want to see what I get on my cpu with the 9800GT or second gtx 470 I can run that for you.
To be fair though I love getting deals. I got my cpu for $185 and my board for $180. I got my new asus for $185 and the evga card for $180 (backplate /w lifetime warranty). I figured in ALL that I could drop the total price on SOMETHING
BFBC2 isn't a cpu intensive title, play a title like World of Warcraft and let me know what kind of frames you get in Stormwind and Orgrimmar on a high pop sever.
I won't ever say you can't run games on an AMD system, or that for many of us all you need is a min of 40+ fps, but if you're spending $700+ on half the parts it's probably best to ignore colors and brands and get the products that go well with each other.Edited by BallaTheFeared - 3/30/11 at 6:27pm