670 sli, 120hz lcd and 3570k cpu.

[setter]

Was the bottlenecking bad gregster, iirc you were running a 2500k? I know that the 3930k has ht, (not sure if 2500k had as i skipped that gen). Does ht make much of a difference in newer games such as bf3? I remember testing with ht on/off on my 920 and i couldnt notice any difference tbh, albeit at 1920x1200 with 400 series cards that may have been a big factor.

 

[james_2k]

there will always be a bottleneck somewhere. the chances of anything being so well synced are slim to none. and on the next game it would be different.

it just depends where the bottleneck is. just need to make sure the limit of a given component is above your fps target. but there will always be a bottleneck

 

[Gregster]

Was the bottlenecking bad gregster, iirc you were running a 2500k? I know that the 3930k has ht, (not sure if 2500k had as i skipped that gen). Does ht make much of a difference in newer games such as bf3? I remember testing with ht on/off on my 920 and i couldnt notice any difference tbh, albeit at 1920x1200 with 400 series cards that may have been a big factor.

Not at all, If I didn't check on GPU-Z or AB, I would never have known. The 2500K has got HT and it is still a great chip but modern GPU's are pushing it. Knowing how close the 670 is to the 680, I would say you would be close to 'not' bottlenecking. Maybe hitting 95% GPU usage in SLI (pure guess though). Also, I run both my 680's with a core clock of 1324Mhz, which is pushing them extremely hard. Maybe on stock, I wouldn't have got it but I didn't do any extensive testing.

Edit: If I had water cooled my I5, I could have pushed it to maybe 4.9/5GHZ and taken away any bottleneck.

 

[setter]

I know what you mean james, at the moment bf3 plays fantastically on my setup, (ran of an ssd as well). 60fps and the game plays extremely well, just that id love a new monitor, and be able to play games at a good frame rate with a good amount of details. I loved moh 2010, max settings on my old setup, a beautiful looking fps, gameplay got a bit rubbish though after a while.

 

[setter]

Not at all, If I didn't check on GPU-Z or AB, I would never have known. The 2500K has got HT and it is still a great chip but modern GPU's are pushing it. Knowing how close the 670 is to the 680, I would say you would be close to 'not' bottlenecking. Maybe hitting 95% GPU usage in SLI (pure guess though). Also, I run both my 680's with a core clock of 1324Mhz, which is pushing them extremely hard. Maybe on stock, I wouldn't have got it but I didn't do any extensive testing.

Edit: If I had water cooled my I5, I could have pushed it to maybe 4.9/5GHZ and taken away any bottleneck.

Very good clocks on the 680's, and the cpu too. Are you considering putting the cards under water cooling at some stage? Gotta admit, nvidia got it right with the 670 series, very impressive cards and cheaper than the 680 which they can match when oc'd, of course but you can clock the 680, the non ref cards such as the windforce are a bit expensive, but take the lower models and put them under water or a good after market cooler and theyre a pretty good card. Shame voltage control is locked though.

 

[Nickolp1974]

The 2500K has got HT and it is still a great chip .

Erm no it doesn't, 2600 does

Must be getting late!

 

[Gregster]

Very good clocks on the 680's, and the cpu too. Are you considering putting the cards under water cooling at some stage? Gotta admit, nvidia got it right with the 670 series, very impressive cards and cheaper than the 680 which they can match when oc'd, of course but you can clock the 680, the non ref cards such as the windforce are a bit expensive, but take the lower models and put them under water or a good after market cooler and theyre a pretty good card. Shame voltage control is locked though.

I have 2 water blocks ordered for my 680s (should have been delivered on the 12th) and can't wait to get them under water. I got lucky with my 3930K chip as it clocks to 5Ghz under 1.5V (which is very good).

If they had released the 670 at the same time as the 680, it would have been a no brainer purchase for me by getting the 670. I needed a new GPU and the 7970 drivers were not doing that great card justice, so I went for the 680. £100 cheaper than a 680 and 5-10% performance difference makes it a fantastic card.

My main disappointment is with the voltage locks but there is always a way round it and I will be playing when I get the cards properly cooled. Artmoney SE allows for hacking the volts

 

[Gregster]

Erm no it doesn't, 2600 does

Must be getting late!

Quite correct sorry, thanks for the correction Had it in my head that speed step was Hyper Threading

 

[MichaelM]

From what I remember with 670 SLI,
2500k @ 4.7GHz
670 SLI @ 1246MHz
1080 @ 120Hz
BF3 - Max quality apart from motion blur - Caspien Border 64 man server
~70-90% GPU usage
~90fps min

Same setup but with single 670, ~90-99% GPU usage and ~70fps min.
Your bank manager will feel the difference more than you will.

 

[chipachap]

I wouldn't have thought so, Gregster hit a slight bottleneck with a 2500K (was at 4.4 or 4.6, can't remember) and a pair of 680's in BF3, GPU usage sitting around 90% on each card.

The 3570K should hold up much better if you can get it towards 4.5ghz.

There's only about 5% max between a 2500k and 3570k.

 

[james_2k]

From what I remember with 670 SLI,
2500k @ 4.7GHz
670 SLI @ 1246MHz
1080 @ 120Hz
BF3 - Max quality apart from motion blur - Caspien Border 64 man server
~70-90% GPU usage
~90fps min

Same setup but with single 670, ~90-99% GPU usage and ~70fps min.
Your bank manager will feel the difference more than you will.

Hmm. Not much of a bottleneck there. Not an impactful one anyway

 

[chipachap]

From what I remember with 670 SLI,
2500k @ 4.7GHz
670 SLI @ 1246MHz
1080 @ 120Hz
BF3 - Max quality apart from motion blur - Caspien Border 64 man server
~70-90% GPU usage
~90fps min

Same setup but with single 670, ~90-99% GPU usage and ~70fps min.
Your bank manager will feel the difference more than you will.

No way you get a 70fps min on a single card?!

 

[MichaelM]

No way you get a 70fps min on a single card?!

Depends on gameplay style obviously because the fps does tend to "tank" when you are in a tank or do extremely far sniping. That's just what I saw with my average play style, it's not a comment on extremes.

 

[setter]

2500k seems to clock higher than an ib such as the 3570k, but i believe/read on here that the difference is negligible between an sb at 4.7ghz or an i5 3570 at 4.3ghz. Factor in that ib is a pretty hot running cpu once you up the voltage, (1.244 here at 4.2), clocked via asus turbo so it could go lower on vcore for that speed.

@MichaelM, did you find the b2k maps to be a bit more demanding than the release maps and other dlc releases such as cq and ak?

 

[MichaelM]

@MichaelM, did you find the b2k maps to be a bit more demanding than the release maps and other dlc releases such as cq and ak?

I don't have any dlc's for it and haven't played it in about 3 months, sorry.

 

[setter]

B2k maps, particularly karkand and gulf of oman will give me drops into the mid 50's, especially in vehicles. I think those two maps are a bit badly coded.

 

[hyperst]

it's not worth it, my 3570k is supposedly the bottleneck with bf3 and sli480, but after some time in game and trying different things, I have come to the conclusion DICE has messed up somehow, if server has just 2 players in, usage on both my cards are 99%, if server fills up usage goes down and I have seen it as low as 50% on both cards.

Yet my cpu is only being used to around 70%, to me it seems DICE has messed up the optimization in bf3.

 

[Gregster]

^^ quite possible. I did some testing but nothing extensive.

 

[Rusty0611]

it's not worth it, my 3570k is supposedly the bottleneck with bf3 and sli480, but after some time in game and trying different things, I have come to the conclusion DICE has messed up somehow, if server has just 2 players in, usage on both my cards are 99%, if server fills up usage goes down and I have seen it as low as 50% on both cards.

Yet my cpu is only being used to around 70%, to me it seems DICE has messed up the optimization in bf3.

Yes. More players = more CPU work.

Your CPU doesn't need to be running at 100% for it to be bottlenecking.

Whether an i5 bottlenecks 2 high end cards isn't debatable. It bottlenecks them end of. It's whether the releasing of the bottleneck by moving from
i5 to i7 is worth it or not where the debate comes in.

Settler: speaking from experience too, average FPS up around 3-5% by swapping and minimum FPS up a little more than that. I didn't bench before and after so don't have FPS readings to be more accurate.

Is it worth it? Only if you want both your cards running at maximum as the difference is only noticeable with monitoring software. Additionally as I said before on a single 1080 screen your usage is going to fluctuate anyway as the game engine tends to reduce load where FPS is very high.

 

[ltnmzh]

I read soemwhere that there is driver issues for 120HZ and Nvidia locked it to < 120Hz for higher res in driver version later than 301.