*PROOF* that a 3930k bottlenecks SLI Titans.

[Guest D1G1T4L1Z3R]

Battlefield 3 is pretty well optimised. I get 70-80 FPS at 1080p on ultra with an overclocked GTX 670. If I had sli 670s, I would be getting around 120 FPS which is close to the OP's results and he has 2 titans! Other people have had much better results than the OP with similar specs, I don't think this is an optimisation or a bottleneck problem. I think there is something wrong with the OP's system.

 

Don't buy a second gtx670 for battlefield 3, it not worth it, Unigine Heaven 4.0 will have 1600+ score instead of 850+ but Battlefield won't be changed at all you will stuck like you were playing with only and only one card, I have two now, it doesn't change anything.

[Jackbe]

Don't buy a second gtx670 for battlefield 3, it not worth it, Unigine Heaven 4.0 will have 1600+ score instead of 850+ but Battlefield won't be changed at all you will stuck like you were playing with only and only one card, I have two now, it doesn't change anything.

Well, you either have a serious bottleneck or another fault because that is definitely wrong. 

[fryman]

64 player bf3 is a huge bottleneck for gpus. My sli 770s hardly get a work out because my 3570k oc to 4.4 can't keep up. Get minimum frames as low as 50 on some maps. Bf3 is all CPU for the larger player counts

[Guest D1G1T4L1Z3R]

Well, you either have a serious bottleneck or another fault because that is definitely wrong. 

 

Sorry, I didn't explained this correctly.

 

I got a gain in Unigine heaven 4.0 (one card= 850 points) (two cards= 1650 points)

 

Battlefield 3= when you are on the gulf of oman and seeing this:

 

 

and having two gtx670 and wonder why you don't get any increase of fps... I doubt that SLI brings you something in this game...

[epkoc]

what would happen if i ran an old quad core i5-760 @4hgz with a single titan? how much bottleneck ?

[Guest D1G1T4L1Z3R]

64 player bf3 is a huge bottleneck for gpus. My sli 770s hardly get a work out because my 3570k oc to 4.4 can't keep up. Get minimum frames as low as 50 on some maps. Bf3 is all CPU for the larger player counts

 

Now, try 4500MHz ... see the result

I actually got a big change when came from 4,3 to 4,5GHz on Noshahr Canals, from 56fps to 68-72fps 

 

when you need to benchmark, watch it here:

 

 

On this view, I came from (4300MHz) 56fps to 68-78fps (4500MHz) with two gtx670 (3570K bottlenecks for sure ! and I thought that overclocking was useless... when you have two cards, it's useful as hell

 

http://valid.canardpc.com/ng0451

[Jackbe]

Sorry, I didn't explained this correctly.

 

I got a gain in Unigine heaven 4.0 (one card= 850 points) (two cards= 1650 points)

 

Battlefield 3= when you are on the gulf of oman and seeing this:

 

 

and having two gtx670 and wonder why you don't get any increase of fps... I doubt that SLI brings you something in this game...

All of the back to karkand maps (BF2) will get worse FPS because they are not as optimised as the modern maps. 

[Jackbe]

Now, try 4500MHz ... see the result

I actually got a big change when came from 4,3 to 4,5GHz on Noshahr Canals, from 56fps to 68-72fps 

 

when you need to benchmark, watch it here:

 

 

On this view, I came from (4300MHz) 56fps to 68-78fps (4500MHz) with two gtx670 (3570K bottlenecks for sure ! and I thought that overclocking was useless... when you have two cards, it's useful as hell

 

http://valid.canardpc.com/ng0451

I don't understand why your frame rates are so low. I get the same FPS as you with a 3770K @4.5GHz and one GTX670! 

 

Or are these specific examples of views that get very poor FPS?

[Kuzma]

Also, is it maybe possible to monitor all the 12 cores (task manager)? I'd be very interested in the results! Thanks in advance

threads* cores are different to threads

[Overl0rd]

threads* cores are different to threads

Threads are also known as logical or virtiual "cores".

[Cacao]

Threads are also known as logical or virtiual "cores".

There's a difference between virtual and physical cores. A 3930k actually has 6 physical cores on the CPU, then windows will see that as 12 because of the hyperthreading feature, but they aren't actual cores.

[Overl0rd]

There's a difference between virtual and physical cores. A 3930k actually has 6 physical cores on the CPU, then windows will see that as 12 because of the hyperthreading feature, but they aren't actual cores.

I know, but I couldn't care to write "could you maybe monitor all the 6 hyperthreaded cores (which will actually make it look as if the cpu has twelve cores, but which isn't really the case since hyperthreaded cores need to share the cache memory and thus do not perform as well as 2 individual physicial cores)"

[Kuzma]

I know, but I couldn't care to write "could you maybe monitor all the 6 hyperthreaded cores (which will actually make it look as if the cpu has twelve cores, but which isn't really the case since hyperthreaded cores need to share the cache memory and thus do not perform as well as 2 individual physicial cores)"

Orrrr you could have just said "could you maybe monitor all the 12 threads" and been done with it

[Overl0rd]

Orrrr you could have just said "could you maybe monitor all the 12 threads" and been done with it

You could have also gotten my point and not started complaining

[Guest D1G1T4L1Z3R]

I don't understand why your frame rates are so low. I get the same FPS as you with a 3770K @4.5GHz and one GTX670! 

 

Or are these specific examples of views that get very poor FPS?

 

These angles of view are poor in FPS, but, these angles are the most important if you are a CRANE CAMPER with a sniper without watching in the scope seeking an enemy to eliminate...

This killed all the gameplay !

And... I don't talk you about back to karkand maps... it's unplayable !

 

Don't get a second GTX670, it doesn't worth it for this game, or games in general, there's just Unigine Heaven 4.0 that are nicely optimized for SLI/CFX

[Xelithium]

First of all, turn off HT. http://chipreviews.com/main-feature/main-news/frostbite-2s-limit-6-core-performance-in-battlefield-3/3/

 

 

Second of all, benchmark Crysis multiplayer. GPU usage and CPU usage there is great. BF3 just relies so heavily on the CPU (on 64 player) that it isn't even funny. Hopefully FB3 improves the GPU usage and actually makes use of the available power.

 

Third of all, FB2 is capped @ 200 FPS.

 

Doesn't BF3 use the HT threads if you have it enabled? Instead of the actual cores? I think it was the FB3 benchmark from alpha that allocated everything to the extra HT threads when HT was enabled.

 

Edit: yea...

[Kuzma]

You could have also gotten my point and not started complaining

I'm just trying to make sure that anyone that reads this thread that doesn't know the difference or does but might get confused by your post knows that cores and threads are different.

[Overl0rd]

I'm just trying to make sure that anyone that reads this thread that doesn't know the difference or does but might get confused by your post knows that cores and threads are different.

If they came here they probably would know that already

[Kuzma]

If they came here they probably would know that already

a lot of people come here to learn

[Sotti]

Hi there,

 

Hey, I'm very interesting in the result of this thread (not logical either physical but forum thread).

 

Maybe we haven't seen that diference between GTX780 ant Titans in SLI because of CPU bottleneck (I know drivers is an issue here too).

 

Keep going.

[Analog]

Well honestly anyone could say 'similar performance', I wish I could see some actual test results from people with the same setup as me.

 

Actually there is video on Youtube, where Tiny Tom Logan from Time to Live Customs was testing an i7 3960X with GTX 780 SLI and after that with GTX TITAN SLI. In some cases the GTX 780 SLI outperfomed the dual TITANS. Go ahead and search for this video if you are interested.

[Cacao]

Actually there is video on Youtube, where Tiny Tom Logan from Time to Live Customs was testing an i7 3960X with GTX 780 SLI and after that with GTX TITAN SLI. In some cases the GTX 780 SLI outperfomed the dual TITANS. Go ahead and search for this video if you are interested.

I know I've seen that video, but once again he doesn't address the bottleneck issue and that's really just the results from 1 guy. Also I think that all goes down the drain once you get to 5760x1080p or 4k, since you use over 3gb of vram in most current gen games.

[Analog]

I know I've seen that video, but once again he doesn't address the bottleneck issue and that's really just the results from 1 guy. Also I think that all goes down the drain once you get to 5760x1080p or 4k, since you use over 3gb of vram in most current gen games.

 

I now and I totally agree with you. What I meant is that most probably nVidia drivers are more optimized for GTX 780SLI, because it's more common (at 1080p or 1440p). Once we reach 3x or 4x 1080p I don't think that GTX 780 will be able to match the performance of the TITAN mainly because of it's 3GB vs 6GB of memory.

[LOST TALE]

 

While running at stock speeds my GPU usage never went above 60%, staying mostly at around 40-50%. When I overclocked my GPU usage stayed between 60-70% and didn't really go below 50%.

you mean when you overclocked your CPU

[Cacao]

you mean when you overclocked your CPU

Yes, when I overclock my CPU.