• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Bad mobo choice for CrossFire?

Status
Not open for further replies.
G

Gen.Viper

Limp Gawd
Joined
Mar 20, 2002
Messages
160
Hey all. I'm in the process of building a new comp, and have already ordered everything but the graphics cards. After doing research on them, however, I think I might have made a bad choice for my motherboard. What I have is this:

Asus P5K Deluxe
Corsair Dominator DDR2 1066 - 4GB (2x2GB)
OCZ EvoStream 720W PSU
Intel E8400 CPU

I'm using a 24" wide screen Dell monitor @ 1920x1200. I was planning on getting an HD 4850 to upgrade the cooling on and OC a bit, then decide whether I want to go the extra bit and get a second to CrossFire. With the res. I run, though, it sounds like the 1GB RAM version might work better (or even get a single 1GB HD 4870). Also, if you look at the mobo specs, it says:

Expansion Slots
PCI Express x16 2 x PCI-E x16 (blue @ x16 mode, black @ x4 or x1 mode)

To clarify, does this indeed mean that the second card (in CrossFire) would only run at x4 max (presumably with the other at 12x?); will this prove to be a significant performance change from a x8/x8 setup, like the 2 below?

If that's the case, I might sell the mobo I bought and get another one, perhaps one of these (which has PCIE 2.0 [Does the 1.0 -> 2.0 jump really improve that much?] and CrossFireX for x8 slot):
ASUS P5Q Pro
MSI P45 Platinum

or anything else you recommend. Thanks in advance!
 
hmm? I'd probably keep that mobo and shoot for a single 4870...should be good enough for that res.....JMO of course, do some more research :D
 
or you could get a x38 mobo which will also solve that issue as all p35 run like that...
 
Anything less than full bandwidth shows a big reduction in Crossfire performance. If you're planning on going Crossfire in the future, you'll want an X38 or X48 board to get the full performance out of it.
 
I have P5K and when I first bought it I thought it was 2@16x too because of how they word it but when you run two video cards it is only 1@16x and 1@4x. I don't plan to CF anyway so it doesn't matter to me.
 
You guys should actually read those comparisons, they don't exactly say what you claim they do. The Tweaktown link just posted actually shows that in games, the difference between P45 and X48 is insignificant unless you're running at 2560x1600, and even then it's not a lot, not what I would consider deal-breaking anyway. Crossfire's going to give you gains larger than the ~10FPS max you'll be missing out on by running it on P45.

Delta between P45 and X48 average FPS based off TT's data:
PT Boats
2560x1600: 9.6
1920x1200: 0.9
1280x1024: 1.2

World in Conflict
2560x1600: 9
1920x1200: 5
1280x1024: -2

Crysis - note that the lower numbers at higher resolutions are because you're getting into single-digit unplayable framerates anyway.
2560x1600: 2.81
1920x1200: 5.91
1280x1024: 10.79

Unreal 3 - note here that the FPS numbers are in the 100's, any amount of extra FPS from a different chipset doesn't even really apply because the game is not even a challenge to a single card on either mobo.
2560x1600: 39.01
1920x1200: 9.26
1280x1024: -4.64

They also do tests with AA/AF on - at 2560x1600 - and the results follow the pattern - at that particular uberhigh resolution there is a difference. I saw another comparison elsewhere - Bit-Tech I think it was - that showed the same overall thing here: Unless you're on a 30-incher the difference between P45 and X48 Crossfire is overblown. You're still getting a massive performance boost by adding a 2nd GPU, I don't consider the board holding it back by 5 FPS to "negate Crossfire" or make it a "waste of money" like people keep insinuating.

EDIT: Also, couple this data with the fact that X48 mobos are extremely expensive...I can see it being arguable that the money is better spent towards a 2nd GPU in terms of gaming performance, especially if you're looking at Crossfiring an inexpensive midrange card like the 4850.

 
people need to stop using tweak town as any kind of metric on new hardware. these guys pop up whenever new stuff appears on the market, throw a crap review together, then sink back into their festering peat bog until the next new thing comes around. wait to hear from a more established site before passing judgment. just because it on the web doesn't make it true. i wish i could post some 4870 cf tests between the p45 and x38 but...i'm....too lazy to get a new psu.
 
This is simply not true from my POV. I run 4850cf on an intel bad axe 2 975x chipset board which has PCIE 1.1 x16 and x8 slots and both slots run at x8 in crossfire. PCIE 1.1 x8 is the same as PCIE 2.0 x4 which would be similar to what a P35 pr P45 board would be in crossfire mode and I get textbook scaling in crossfire. Sure it's probably somewhat faster on a PCIE 2.0 x16 dual slot board but probably less than 10% would be my guess. Until I see hard numbers where the same hardware is used on 2 or 3 different motherboards then I gotta call BS on this one. Faster? Yeah. Night and day faster? No. Probably not even noticeable would be my guess. I would still buy an x48 if I were buying a new board and hadn't already purchased some other older board, but I wouldn't replace a perfectly good board in the hopes of eeking out an unnoticeable 10% improvement in frame rate and regardless of what anyone tells you, crossfire is certainly not wasted on older boards. I get kickass frame rate improvements with it even on my old board.
 
I'd either get a 4870 or wait for the 4870 X2 if you really feel you need the horsepower... The board is well regarded from what I can tell. If you HAVE to do crossfire get an X38 or X48 based board.
 
p35/965 has one 16x and one 4x pcie 1.1 slot. this setup is junk for cf. trust me, i dropped p35 for p45. p45 has two 8x pcie 2.0 slots. the assertion that p45 cf is drastically inferior to x38 or x48 is highly debatable largely because of the dearth of information. a shitty article from tweak town does not a whole story make.
 
Bit-Tech sorta-kinda compared P35/P45/X48 motherboard Crossfire performance in their Gigabyte GA-EP45-DQ6 review. They used older Radeon 3870's though:

Bit-Tech GA-EP45-DQ6 Review

Sure looked like the P35's lower PCI-E bandwidth hurt it in Crossfire...
 
Aha, that is the Bit-Tech article I was referring to. Testing at 1680x1050 and 1280x1024, they are also showing little to no difference between P45 and X48 for Crossfire performace. P35 is obviously crippled though.

 
Avrithor said:
Aha, that is the Bit-Tech article I was referring to. Testing at 1680x1050 and 1280x1024, they are also showing little to no difference between P45 and X48 for Crossfire performace. P35 is obviously crippled though.

Click to expand...

It would seem to make sense that if 16x + 4x Crossfire is a limitation on the P35 even with the older 3870's, then the 8x + 8x Crossfire of the P45 should limit the newer/higher bandwidth 4850/4870 cards in Crossfire as well.

I'd personally like to see additional testing from reputable sites before purchasing a new X48 or P45 motherboard.
 
rennyf77 said:
p35/965 has one 16x and one 4x pcie 1.1 slot. this setup is junk for cf. trust me, i dropped p35 for p45. p45 has two 8x pcie 2.0 slots. the assertion that p45 cf is drastically inferior to x38 or x48 is highly debatable largely because of the dearth of information. a shitty article from tweak town does not a whole story make.
Click to expand...

Go research it then because other articles besides tweaktown say the same thing. And I see nothing wrong with teaktown reviews. You have some evidence that they cheat on the benchmarks or something? They set up the hardware and benchmark it and post the results. How is that any different from what any other review site does?
 
Well its bottlenecked thats for sure, it might not be significant in games we have now but it shows the problem exists and that just not a good set up to have. Grab a cheapish x38 board like the Asus P5E with a 16x/16x PCI-e 2.0 set up
 
I don't see why you would want to get a p35 for crossfire. If you are in the market for a new board and want to CF then get a good motherboard for it that will maximize the potential power that CF can obtain. Even if the delta for the difference between the CF performance for each motherboard is minimal I would still get the mobo that is 16 in both lanes. It won't cost you that much more if anything.... If you are willing to drop cash on two video cards you should be willing to shell out the extra cash for a motherboard that will support those two cards.
 
You've basically got it right. The P35 boards are going to give you (for the most part) only an option of 16x4 for Crossfire. P35 boards are also only PCI-Express 1.0/1.0a. This is less than ideal. The newer P45 boards are going to do 8x8 and they are PCI-Express 2.0. Clearly for Crossfire the P45 is the way to go. However X38 or X48 are an even better choice as those will do Crossfire in a 16x16 configuration which is the best possible way to go at present.
 
milkweg said:
Go research it then because other articles besides tweaktown say the same thing. And I see nothing wrong with teaktown reviews. You have some evidence that they cheat on the benchmarks or something? They set up the hardware and benchmark it and post the results. How is that any different from what any other review site does?
Click to expand...

where did i say they cheat on their reviews. their reviews suck because of the lack of information they provide. the bit tech p45 review is much better. fell free to link any other reputable sites.

i googled 'crossfire on p45' and aside from the tweak town and bit tech reviews, every entry for two pages of search results links back to the tweak town review. thats what i meant by a dearth of information. get a fucking clue.
 
Flaming will get you no where. Tweaktown ran enough benches to show me that 2 16x lanes makes enough of a difference. What more do you want? A description of the outside of the video card box?
 
I also added this to the other pcie 2.0 discussion here.

Here's some motherboard testing from Legion Hardware.

Link

By simply firing up 3Dmark Vantage we found that the new X48 chipset was able to provide 17% more performance than the older P35 chipset when using two Radeon HD 4850 graphics cards in Crossfire mode. More over, the X48 was less than a percent faster than the P35, which was just a percent faster than the X38 chipset. These numbers would not exactly have X38 owners rushing out to pick up an Intel P45 or X48 motherboard.

However, as we suspected this did not turn out to be the full story, as we began testing with some real games. When testing with games such as Crysis, we quickly discovered that the X48 chipset was able to deliver 10% more performance when compared to the X38, while it was slightly faster than the P45. While this was impressive, Crossfire is not all that well suited to Crysis, so we continued on and hit Devil May Cry 4, which works a treat.

Now when testing with the new hit title Devil May Cry 4 we found that a typical P45 setup is going to be 36% faster than a P35 system at 1920x1200 when running Crossfire Radeon HD 4850 graphics cards for example. The X48 was also roughly 6% faster than the P45 chipset, and I would suggest that this is a best case scenario for the high-end X48 chipset. That said, this did make the X48 around 15% faster than the older X38 chipset.

For the most part the X48 is just 5% faster than the cheaper P45 chipset, while this chipset held roughly the same margin over the older X38 chipset. Therefore as we see it, X38 owners really would need to purchase an X48 motherboard to see any real performance gains, while the P45 is a massive step up from the P35 chipset when talking Crossfire performance.

Right now it is possible to purchase a decent ASUS P45 motherboard, such as the P5Q Pro for example, at just $150 US. Pairing this board with Crossfire Radeon HD 4850 graphics cards would cost around $550 US, making for an extremely lethal combo! The cheapest ASUS X48 board on the other hand, is the P5E Deluxe at $220 US, and we are not sure that at this price it is worth the 5% Crossfire performance gain. Still, if you are looking to get the most out of your Crossfire setup, then the X48 is the weapon of choice.
Click to expand...
 
milkweg said:
Flaming will get you no where. Tweaktown ran enough benches to show me that 2 16x lanes makes enough of a difference. What more do you want? A description of the outside of the video card box?
Click to expand...

that's utterly ridiculous. how about minimum frame rates or testing methodology? do they tell whether their using custom benchmarks, recoded real time game play, in game benchmarks? if they're using recorded game play or custom benches, what stages are they using. for their 'high quality' aa and af test they use 3d mark. are you fucking kidding me? speaking of aa. what kind of aa do they use in their tests? multi sampled or super sampled?
box filter, narrow tend, or wide tent. these are things that matter. these jokers throw up a bunch of arbitrary numbers and you just nod i approval? this is the kind of 'more' i was looking for. you should respect yourself more and demand better from these quacks as a patron.
 
I believe the major selling point of the of the P45 would be a single card solution with an ageia card.
 
you are also forgetting price points, I'm running a P5Q motheroard right now with a 4870 CF set up, my results mimic alot of what [H] got, and every other site as well, infact the numbers were close enough to the actual results that I don't care about the 5% difference since I'm running these ganks cranked at 1920x1080. So even though I spent $600 on video cards I only spent $120 on the mother board, allowing me to go that extra mile for my video card solution with out having to cut other corners.

To the OP if you want to save some cash and not see a difference go with the P5Q, bad ass motherboard and also saves on power, if you want to go all out get the P5E, if you don't like ASUS I don't know what to suggest to you since I've pretty much only purchased intel and ASUS motherboards :cool:
 
Whew! You guys are pretty awesome at setting everything out; got some more research to do, but yea, looks like I'll probably be upgrading the mobo, an x38/x48, possibly either a
- Gigabyte GA-X48-DS4
- Asus P5E WS Professional
They support the 1066 RAM I already have, and have dual LAN (though the ASUS has a healthy price jump from the GB board), The Asus P5E and P5E deluxe don't look bad either, even though they don't have the dual LAN. If you have any suggestions as to better boards, or reasons why not to pick one of those, I'd love to hear them.

Things became a bit trickier with the Nvidia uber-price drop, but I'm still liking the ATi solutions better; I think I'm gonna get a 4850CF setup. Now if only the stuff from Newegg would work when it arrives....I had to return the PSU because it would power up just long enough to make the fans shudder, and still did it when it was pulled out and connected with the 'PSU test' config (green + black wire). Just got my replacement yesterday, and it's still doing it, which makes me wonder if it could be something else; I dont see how though, because it isn't working when it's tested by itself. To the guy who said it's a bad PSU choice, I must ask why? It's by a good company, and it's in one of the top tiers for PSU performance/reliability. my main concern was whether it would have enough juice, but from what I've read it seems like it should. Anyway, thanks for all the feedback, you guys rock! :D
 
If you are going to cheap out on the motherboard at least buy a p45. The p35 is still only pci-e 1.1 so trying to use the Tweaktown numbers is just stupid. They tested 8x/8x pci-e 2.0 no 16x/4x pci1.1, HUGE difference. That being said, i spent the extra money and bought a x48 because i want the max performance out of my 4870 crossfire set up. If i have to skip a few nights eating out to pay for it, so be it.
 
veedubfreak said:
If you are going to cheap out on the motherboard at least buy a p45. The p35 is still only pci-e 1.1 so trying to use the Tweaktown numbers is just stupid. They tested 8x/8x pci-e 2.0 no 16x/4x pci1.1, HUGE difference. That being said, i spent the extra money and bought a x48 because i want the max performance out of my 4870 crossfire set up. If i have to skip a few nights eating out to pay for it, so be it.
Click to expand...

QFT. If you have to wait a bit to get enough money I suggest you do. my x48 Asus Rampage Formula is quite a kickass board. I've had absolutely no issues with it... even flashed the BIOS without a hitch. Well worth the money, IMO.
 
The Rampage Formula is the one I'll use if I like the retail reviews of the HD4870x2s.
 
rennyf77 said:
that's utterly ridiculous. how about minimum frame rates or testing methodology?
Click to expand...

I believe I already said their result are confirmed by other website articles so the only thing ridiculous here is your hostile attitude.
 
milkweg said:
I believe I already said their result are confirmed by other website articles so the only thing ridiculous here is your hostile attitude.
Click to expand...

and when i asked you to link other reviews that corroborate tweaktown's findings you gave me nothing. their numbers with no explanation or exposition were all you needed dismiss p45 as a viable cf option. if not for the legion hardware review you would be under the assumption that x48 and x38 were the only way to go. the editors at legion assert that p45 cf performance is even with x38 and sometimes better and still very close to x48. while interesting, one cannot take this information at face value without more corroborating results. minimum frame rates are always more important than average, and until i see that kind of data the proposition that 16x 2.0 pcie cf is markedly better than 8x 2.0 pcie cf remains in limbo. furthermore, there were no hostilities in my posts until you decided to put words into my mouth about tweak town and cheating their results. again, i never said tweak town made up their numbers. i only noted my dislike of them, chiefly their haphazardly thrown together reviews that pop up on the enthusiast radar only when new hardware is about to appear on the market. the fact that you would defend them in spite of the utter lack of detail in their article is what is wrong with the enthusiast community. you could have responded with a, 'what don't you like about tweaktown', but that's not what you did. so here's a some more hostility for you: fuck off.
 
milkweg said:
I was given a 1 week vacation for saying less than you. Where is the mod now to give you some vacation time too?

Oh, and gee golly-whiz, Tweaktown does show minimum and average FPS. I doubt you even read their CF reviews of 4850 and 4870 seeing as you claim they don't show minimum FPS in their benchmarks.

http://www.tweaktown.com/reviews/1481/6/sapphire_radeon_hd_4870_in_crossfire/index.html
Click to expand...

where did they show it in the p45 vs x48 shit-for-brains?

and it still doesn't answer my question regarding the different anti aliasing methods. i've pinned down my objections to the crossfire comparison very specifically, and you keep skirting around the issues. anything else?
 
milkweg said:
Right here.

http://www.tweaktown.com/articles/1472/4/intel_p45_vs_x48_crossfire_performance/index.html

and here

http://www.tweaktown.com/articles/1472/6/intel_p45_vs_x48_crossfire_performance/index.html

BTW, I will no longer tolerate your attitude and have reported you to a mod.
Click to expand...

that's fucking hilarious. you can't tolerate my attitude yet you keep responding to my posts. now back to the matter at hand, i stand corrected about tweak town not posting minimum frame rates. you're right. they show mins in some pt boat game that i've never heard of and world in conflict, but why didn't they do the same for crysis and unreal 3? by the way, what were the settings they used for those games? you can't tell me right? that's because tweak town doesn't mention them. don't you think this would be kind of useful information to have? all of this acrimony really comes down to the fact that you think tweaks town's work is adequate in portraying the effectiveness of one crossfire implementation over another. i think their work is far from adequate. i think the blurb they write in the summary at the top of the article is laughable: 'we take the time to analyze the difference between 8/8 cf on p45 and 16/16 cf on x48.' the amount of work they put into this article can be completed over one's lunch hour. hell, a chimp prodded with enough fruit could probably accomplish the same job. bitch about me to the mods all you want. in the end, i still think you're a fool and you still think i'm an ass, and that's not going to change any time soon.
 
Status
Not open for further replies.
Back
Top