[GTX 670] ASUS DCUII TOP or Gigabyte Windforce OC? Which one, and why? - Page 3

For someone complaining about mis-information, you're kinda dishing a lot of it yourself

1) Although there may be a certain maximum for a given model (i.e. 670's in general), the exact max on the power target % is variable depending on the individual card/chip, not the brand.

2) You're extrapolating quite a bit about max clocks and such without a very large sample size. Remember that reviewers are NOT OCN'ers. Personally I think there's no possible way with the amount of data out there to conclude with any certainty that the Asus is universally going to be capable of much higher clocks (core or especially memory ... they're the same memory chips and memory controllers) than the GB as you've suggested. You may be right, but you're stating some of what you say there with a bit more 'certainty' than I think is warranted at this juncture

3) There's no evidence that Asus models have any more 'voltage control' than the other 670's. The max voltage is still 1.175V. Furthermore your link to TPU doesn't say anything there's anything unique about the Asus bios so I dunno where you got this whole 'custom bios with a lot of volt mods' spiel. The description there is generic 670 information.

4) There's no real evidence to the suggest the Asus is any quieter than the GB at equal temps. So your 'likely run a whole lot quieter' and 'loud fans' talk ... that's speculation, pure and simple. Both use custom-designed DC coolers built by Coolermaster with very similar specs.

5) The Asus uses a 680-sized board just like the GB, so I dunno why the small PCB is even part of the discussion. I don't know if it's an actual 680 reference board (put it this way, they cannot both be actual 'reference' 680 boards, due to the location of the power section being different but I'm not sure if one of them is, nor which one is), but it's a full-sized 680 board.

6) Do you have some evidence that the 'min voltage control' wouldn't work on a different 670 card if you used the Asus software with that card?

This all being said, if the TOP had been available for $419 (or even $429) when I bought my GB I'd have bought it cause I like the look of the Backplate and exposed heatpipes, I think it looks better for sure, and that's worth something to me. The 6-stage power vs. 5-stage power is nice as well.

But all this talk of performance differences and cooling efficiency ... that's a lot of speculation on your part

We need a same source, like guru3d, to review both of these cards.

1) That might be true, but it doesn't really change the fact that every gigabyte model we've seen so far is at 111% and every asus that has sufficient information in their reviews is at 122%. Either way, this difference is minimal at best as some people are getting better overclocks on the gigabyte by not maxing out the power-target. This is true of the 680 too. I didn't say one was better than the other, i was just pointing out the difference based on our limited data.

2) Of course I'm extrapolating, how else are we to determine which 670 to buy on release day? I'm taking what limited information we have so far and trying to apply it to the question at hand. It should be assumed that everything said in this thread is just speculation until the Asus is released. From what little information we do have though, the gigabyte seems to have slightly better core clocks but lower memory clocks. The asus TOP also comes factory overclocked at around 1250 boost core and the gigabyte is at around 1200 boost core from the factory. So if you get the Asus TOP you're a least assured to hit 1250, this removes at least some of the luck associated with getting a good clocking card. Based on the thread i linked originally, you can see some people are struggling to hit 1230 boosted core on the gigabyte, though the majority are easily hitting 1250+ and up to 1300. That's just the nature of the silicon lottery though, and i would much rather take my chances with a binned chip (Asus TOP's are binned) over a non-binned chip (gigabyte chip's aren't binned).

3) There is some evidence though. In one review the asus has a 'min voltage' setting that the gigabyte doesn't seem to share. I realize this is just software related, but can the gigabyte use this Asus-based software too? I haven't checked yet (if someone wants to, that would be helpful ). Here's a direct link to the graph i was talking about. Right under that graph TechPowerUp says, "We see a huge range of clocks and voltages which show that ASUS did some extra tweaking to their settings," which doesn't necessarily imply a custom bios like i originally stated, but they do have some form of customized voltages. To compare this graph to some of the other cards just for reference: Palit GeForce GTX 670 JetStream, ZOTAC GeForce GTX 670 AMP, Reference NVIDIA GeForce GTX 670. So you can see there's definitely some differences here and there's a lot of way to interrupt and explain the differences, but i was just going with the explanation TechPowerUp chose to go with (custom asus tweaks).

4) There is no direct evidence that the asus will run quieter than the gigabyte, i specifically said it was speculation on my part, but here is what we do know (and i realize it's not completely applicable here, but it's what we have to go with). If we look at the noise that the Asus 680 produces here, we can see it's 19% quieter than the reference model. If we then look at the Gigabyte 680 here, we can see it's only 3% quieter than the reference model. None of the Gigabyte reviewers measured the noise levels of the 670, so a 680 comparison is as good as it gets for now. But what we do know is the Asus is by far the quietest of all of the non-reference models it's been compared to in other reviews (the gigabyte hasn't been compared to anything either). At Hardwareinfo they compared it to a 2 other factory OC models,, at TechPowerUp they compared it to 2 factory OC models (Zotac at 35dba, Palit at 40db) and lastly at Guru3d it's 10% quieter than reference at a significant factory overclock. Another thing to take away from all these graphs besides the direct comparisons to other 670's is the Asus is one of the quietest cards overall compared to any card, clock for clock it's probably the quietest card in each graph.

5) You're indeed correct, they're both using 680 sized PCB's . That was a definite oversight on my part. Either way, i wasn't saying one was better than the other, just that if someone already had a 680 backplate or waterblock, there was a possibility they could use it on the 670 too. Now that we know they're both using 680-sized PCB's then this point is moot. However, the asus does come with a nice looking black backplate pre-installed which is a definite plus.

I was just presenting the hypotheses i drew based on the limited amount of data we have so far. It's definitely appreciated you took the time to analyze and respond to what i originally presented, as this healthy debate will help us get all of the information sorted and analyzed without bias. i just want to decide which card to buy just like the original poster
Edited by SeanPoe - 5/12/12 at 12:36pm

It's my first time purchasing a Gigabyte videocard. So far it's been great (though the accessories and all seem a bit....plain) running games/synthetic benchmarks. I was able to push the core clock to 1354Mhz with boost.

I've also applied my own thermal paste vs the original one that came with it. The temperatures have slightly changed, and fans are very quiet (just like the DC versions of ASUS).

Overall, many people dislike the PCB color itself (blue). To be honest, both cards are great, but it would seem that ASUS offers more accessories/softwares with the card.

Thats good then. Far better. The triple slot is an incredible cooling solution but really inconvenient for expansion.

Ugh, think my eyes hurt after the few huge walls of text
We've done a little testing at work with both cards and they both are quiet. Overclocking really depends on silicone lottery, but if the TOP cards are binned, it might give better chances for decend overclock. Yeah, the blue pcb on gigabyte card is kinda hideous if you havent got a blue system, but it ain't a dealbreaker in my opinion.
The ASUS's new 2-slot cooler looks to be an improvement from their past coolers, on GTX500-series I would have gone windforce or twin frozr over asus's cooler any day, as their 2-slot design was inferior and 3-slot is just plain stupid imo, but now it looks like ASUS's card might be a winner here. We don't have a sound pressure meter at work, but the ASUS was a bit quieter (2 fans vs. 3 fans, no surprise there).
Don't think I have anything else to say, the facts have already been stated earlier (more or less).

You could check and see what the power target is on each card. Is the asus at 122% and the gigabyte at 111%?

@ Sean

Yeah, so I was just doing some research on teh 111% vs 122% thing ... thought it might be Asus Software vs. Prec.X related but I'm seeing Ref cards at 122% max in PrecX as well ... seems really odd the GB would have a lower power target ... although of course 111 may be plenty given the voltage cap being at 1.175 ... time will tell on that. It may also be modify-able through a bios flash as well like you mentioned. I'd like to know if the Asus software works on the GB card, and how the power limit looks on each, as well as whether the min voltage meter is enabled (and actually works) in both cases..

Otherwise, this later version of your post is much more 'accurate' and detailed. I'd quibble about the 'cooling' suppositions a bit ... in reality comparing coolers and quietness accurately is incredibly hard unless someone with all the right gear does so in exactly the right way ... it's so hard to get apples/apples comps, but you're aware of your extrapolations here, so ...

Also, not sure I 'buy' TPU's talk about voltage customizations ... given that the most complex/varied graph of volt/clocks is the Reference card ... go figure.

Honestly, afa voltage goes, the only reason I'd care about any of it is if one card or the other could do >1.175V. Aside from that, none of this stuff really matters. Not even the extra power phases on either of these cards probably matters at all

I generally don't disagree on your suppositions, and like I said earlier I think the Asus DCII are the best ... DOOO EEET!

Wow, 1354 core? I think we have a new record!



Now ... can you show us some of that? Have you thoroughly tested stability at that clock? Could we see a 3dMark11 score at that clock?

Not that I don't believe ya ... I really WANT to believe you in fact

And what's the mem clock at when you get that boost?

No reason to take noise levels from separate reviews. Guru3d.com has reviewed both Gigabyte and ASUS models of GTX 680, and their noise levels are exactly the same.

I would honestly take the one which is cheaper.

I didn't see that Asus 680 review before. Thanks for pointing that out. t is interesting though what a huge difference the two review sites got for the same card. At TPU the 680 was 19% quieter than the reference 680 and at Guru3D it was only 3% quieter. This entire industry really needs to become more scientific because right now, the methodology all of these reviewers use is appalling and completely inconsistent.