Is my i7 930 going to be a bottle neck for these new Nvidia cards in SLI?

E

ebeattie

Limp Gawd
Joined
Jan 29, 2007
Messages
441
Hi all. Well about 3 months ago I finally got my second GTX 580 for SLI and thought my system was the cats ass. finally a high end SLI system!

...... Then I read about the GTX 680.... then the rumors about the GK110..... :eek::eek::eek:

So here is my Question:

Is my i7 930 system in sig (@3.8 Ghtz) going to be a bottle neck for the new generation Nvidia cards?
 
a 3.6ghz i7 was a bottleneck for GTX 580s in SLI, so a good chance that the 680s might need a bit more power.
 
Drat.... lol well I guess Ill be looking up parts for a 2700K build.... damn technological advancements!
 
My 920 at 3.8 is still doing just fine. Waiting to see what IB brings before considering upgrading.
 
It really depends on the game and resolution.

You want to get 680s assuming because you do not have enough performance at the moment.

Well, load up the games in which you want more performance in and check your GPU usage.

If your GPU usage is below ~95% then your CPU is what's holding you back. Getting 680s wont increase performance if you GPU is not already maxed out now.
 
MavericK96 said:
My 920 at 3.8 is still doing just fine. Waiting to see what IB brings before considering upgrading.
Click to expand...

Quagmire LXIX said:
Where do you get that?
Click to expand...

Original article run with a 3.6ghz i7 920
http://hardocp.com/article/2011/04/28/nvidia_geforce_3way_sli_radeon_trifire_review

Redone with a 4.8ghz 2600k
http://www.hardocp.com/article/2011/05/03/nvidia_3way_sli_amd_trifire_redux/

sorry that was tri-sli vs tri-fire, I'm not sure if it will translate to 2 card SLI but I imagine 2 GTX 680s will be about as fast as 3 GTX 580s.
the original article did include GTX 580 2-way sli
 
Last edited:
I'd say yes a 3.8Ghz 930 will be a BIT of a bottleneck, but honestly not that bad.
 
Digital Viper-X- said:
Original article run with a 3.6ghz i7 920
http://hardocp.com/article/2011/04/28/nvidia_geforce_3way_sli_radeon_trifire_review

Redone with a 4.8ghz 2600k
http://www.hardocp.com/article/2011/05/03/nvidia_3way_sli_amd_trifire_redux/

sorry that was tri-sli vs tri-fire, I'm not sure if it will translate to 2 card SLI but I imagine 2 GTX 680s will be about as fast as 3 GTX 580s.
the original article did include GTX 580 2-way sli
Click to expand...

I'm not exactly sure what detail in there is supposed to make your point, but to call a 3.6 quad a "bottleneck" because a 4.8 quad can open up the system performance more is diserving, of course 4.8Q will shine over 3.6Q. You're insinuating that the experience would be so hampered, that someone has to go get a new SB system and oc the donkey doo out of it to get good use out of a couple 680s.

In my testing of the 920/SLi over its years, 3.8 did make a good diff over 3.6 and 4 was of course better, a 930 should be able to do 4 on air rather easily, but 3.8 is very good for running SLi and I don't see the OP needing to upgrade the X58 just for 680s.

This could all just be about how severe ppl view "bottleneck". Can a 4.8SB do better, yes, but is a 3.8 i7 a loser in the case, no.
 
Looks like the 2GB on each card is an SLI bottleneck instead of the proc, wait for H review first before saying the processor is the bottleneck.
 
Quagmire LXIX said:
I'm not exactly sure what detail in there is supposed to make your point, but to call a 3.6 quad a "bottleneck" because a 4.8 quad can open up the system performance more is diserving, of course 4.8Q will shine over 3.6Q. You're insinuating that the experience would be so hampered, that someone has to go get a new SB system and oc the donkey doo out of it to get good use out of a couple 680s.

In my testing of the 920/SLi over its years, 3.8 did make a good diff over 3.6 and 4 was of course better, a 930 should be able to do 4 on air rather easily, but 3.8 is very good for running SLi and I don't see the OP needing to upgrade the X58 just for 680s.

This could all just be about how severe ppl view "bottleneck". Can a 4.8SB do better, yes, but is a 3.8 i7 a loser in the case, no.
Click to expand...

I'm going to go ahead and assume you didn't read much of either link.

#1) with a 3.6ghz i7 920, the Tri-Fire system is FASTER than the tri-sli 580s.
#2) With a 4.8ghz i7 2600k, the Tri-SLI System is considerably faster then the 3.6ghz.

What more proof do you need that a 3.6ghz i7 QUAD was bottlenecking the TriSLI setup? by bottleneck ,it means the CPU is holding back the GPU, which is true in this case.


As we have seen here today, we have fully uncovered some real world gaming scenarios that we were using that were CPU-limited in terms of the NVIDIA system, and we have seen some results that are fully confusing from Tri-Fire that need to be looked into further. It is one thing we have found from using real world game testing over all these years is that nothing is ever as simple as a canned benchmark would make you think it. We have learned some things here, and now have even more questions, and we have our readers to thank for that. Thank you.
Click to expand...
 
So a stock clocked c2q 9550 is really in need of a upgrade.. ugg.. the wife is going to kill me :D
 
Digital Viper-X- said:
I'm going to go ahead and assume you didn't read much of either link.

#1) with a 3.6ghz i7 920, the Tri-Fire system is FASTER than the tri-sli 580s.
#2) With a 4.8ghz i7 2600k, the Tri-SLI System is considerably faster then the 3.6ghz.

What more proof do you need that a 3.6ghz i7 QUAD was bottlenecking the TriSLI setup? by bottleneck ,it means the CPU is holding back the GPU, which is true in this case.
Click to expand...

I'm going to assume you didn't read my post and you're not paying attention to your own admission that the articles are TRi-SLi (as you well know it takes more cpu for that than SLi).
I don't have to read much to know your position is simply a 4.8SB is better for TRi-SLi than a 3.6i7 and then you assume similar results for SLi. Well duh, everyone just throw out their oc i7s..
Like I said in my post, depends on the perception of the word "bottlenecking" and does it mean that a system can't run it, so throw it away for a whole new mobo and cpu.

@ebeattie You will have less performance out of them than with a higher clocked SB, but that isn't rocket science, however, you won't need to throw away that 3.8 930 to get good use from 2xGK110 imo.

loxety said:
So a stock clocked c2q 9550 is really in need of a upgrade.. ugg.. the wife is going to kill me :D
Click to expand...

rofl
 
Quagmire LXIX said:
I'm going to assume you didn't read my post and you're not paying attention to your own admission that the articles are TRi-SLi (as you well know it takes more cpu for that than SLi).
I don't have to read much to know your position is simply a 4.8SB is better for TRi-SLi than a 3.6i7 and then you assume similar results for SLi. Well duh, everyone just throw out their oc i7s..
Like I said in my post, depends on the perception of the word "bottlenecking" and does it mean that a system can't run it, so throw it away for a whole new mobo and cpu.

@ebeattie You will have less performance out of them than with a higher clocked SB, but that isn't rocket science, however, you won't need to throw away that 3.8 930 to get good use from 2xGK110 imo.



rofl
Click to expand...

I never said it couldn't run it, an athlon x2 can run them, doesn't mean he'll get the full potential of the cards,

if he is upgrading from a GTX 580 sli -> GTX 680 SLI, or GTX 580sli to GK110 SLI, there is a good chance that he won't see the benefit since his CPU will hit a wall before his GPUs do.

Lets see dropping $1000 on gpus, and worried about a $450 mobo/cpu upgrade?
 
loxety said:
So a stock clocked c2q 9550 is really in need of a upgrade.. ugg.. the wife is going to kill me :D
Click to expand...

Get it over 3.3 ghz and it won't be a huge bottleneck. A 45nm C2Q at 3.0 is about equivalent to a SB I3. At 3.3, its at the heals of the dual core I5s. Then at 3.8ghz its similar to the lower end I5 quads. The 775 platform isn't dead yet. Its definately gasping because of memory bandwidth, but it still uses most AMDs as toilet tissue.
 
Digital Viper-X- said:
I never said it couldn't run it, an athlon x2 can run them, doesn't mean he'll get the full potential of the cards,

if he is upgrading from a GTX 580 sli -> GTX 680 SLI, or GTX 580sli to GK110 SLI, there is a good chance that he won't see the benefit since his CPU will hit a wall before his GPUs do.

Lets see dropping $1000 on gpus, and worried about a $450 mobo/cpu upgrade?
Click to expand...

A "good chance" now. Then a little more enlightenment to your original post would have been warranted, don't you think? Possibly suggesting what he might need to get the full potential and how much performance he might lose with only a 3.8i7 instead of your suggestion. Who's to say a 4.8SB will give full potential from SLi GK110, 5GB will do better.

$450 is 45% of $1000 and it is significant, where in the world are you buying a substantially improved cpu over a 3.8i7, and a performance mobo worthy of a couple new GKs for $450 anyway, unless you're trying to calculate some return on selling his old stuff.

Anyway, the OP has heard enough I'm sure. Take care DV-X.
 
Quagmire LXIX said:
A "good chance" now. Then a little more enlightenment to your original post would have been warranted, don't you think? Possibly suggesting what he might need to get the full potential and how much performance he might lose with only a 3.8i7 instead of your suggestion. Who's to say a 4.8SB will give full potential from SLi GK110, 5GB will do better.

$450 is 45% of $1000 and it is significant, where in the world are you buying a substantially improved cpu over a 3.8i7, and a performance mobo worthy of a couple new GKs for $450 anyway, unless you're trying to calculate some return on selling his old stuff.

Anyway, the OP has heard enough I'm sure. Take care DV-X.
Click to expand...

2600k~ $300
Z68 Motherboard ~$150

I'll test it for you tonight with a 3.8ghz i7 and a 4.8ghz i7 with GTX 680 SLI :) good?
 
I have a pair of GTX 680's coming in tomorrow. I'm going to pair them with an [email protected] for the time being. Its performance should be very close to the OP's 930. I also own a [email protected] that I can use for comparison. Depending on what I experience, an upgrade to Z77+IB may be in my immediate future.
 
Digital Viper-X- said:
#1) with a 3.6ghz i7 920, the Tri-Fire system is FASTER than the tri-sli 580s.
#2) With a 4.8ghz i7 2600k, the Tri-SLI System is considerably faster then the 3.6ghz.
.
Click to expand...

Something is weird with those benchmarks, the i7 920 is faster in battlefield 3 with the amd cards than the 2600k was. Dragon age 2 is almost identical between the two of them as well. Almost looks like they've updated the nvidia drivers they used more than anything imo. The F1 benchmark also shows the 920 as being faster with xfire, though the 2600k is much faster for the sli (which apparently only got 6 more FPS for the third card with the 920?).

Just seems like some weird benchmarks to me.

Edit: They explicitly state in the opening that they're using the same drivers =/

Second edit: According to newegg that motherboard that they used for the first test was 16x 16x 4x for pci-express. It might be changeable in the bios somewhere though? Mine can switch between 16x / 1x for the second and third slots and 8x / 8x. Otherwise You're using two cards (6990 and a 6970) vs. 3 on a motherboard that's going to be neutered by pci express bandwidth for the third card.
 
Last edited:
I have a 930 system. Might upgrade, might not, when Ivy comes out. Just ordered two 680s.

I figure at 4.4GHz though my CPU isn't going to be much of a bottleneck.
 
Digital Viper-X- said:
2600k~ $300
Z68 Motherboard ~$150

I'll test it for you tonight with a 3.8ghz i7 and a 4.8ghz i7 with GTX 680 SLI :) good?
Click to expand...

Little cheap on the board when speaking of a high end system, but a $150 Z68 will get your cpu to mid/upper 4's and keep SLi (just 2 though) working at PCIe2.0x16. Any more than that however, I call the board as a "bottleneck" :)

I won't turn down results :cool:, I'm sure the OP would also like to know how much more performance he'll get out of his SLi680 for $450 + new memory over his already owned X58. Thanks :)

Rizen said:
I have a 930 system. Might upgrade, might not, when Ivy comes out. Just ordered two 680s.

I figure at 4.4GHz though my CPU isn't going to be much of a bottleneck.
Click to expand...


There it is, X58 to ivy!
 
Quagmire LXIX said:
Little cheap on the board when speaking of a high end system, but a $150 Z68 will get your cpu to mid/upper 4's and keep SLi (just 2 though) working at PCIe2.0x16. Any more than that however, I call the board as a "bottleneck" :)

I won't turn down results :cool:, I'm sure the OP would also like to know how much more performance he'll get out of his SLi680 for $450 + new memory over his already owned X58. Thanks :)




There it is, X58 to ivy!
Click to expand...

you don't need new memory going from x58 to p67/z68, they both require 1.5V DDR3
 
Here you go! Now this was done with an empty map, I only imagine a full 64man server would be even heavier on the cpu. Big enough difference for ya?

Typod on the Min/Max/Avg, labeled both as 4.8, the one on the right is 3.8.


3848.png
 
Digital Viper-X- said:
you don't need new memory going from x58 to p67/z68, they both require 1.5V DDR3
Click to expand...

240 pin, true enough. The OP shows "12GB Corsair XMS DDR3 1600", idk if their 1.5 or up to 1.65, if their 6x2 or 3x4, so he may be able to use 4x2 and have an extra 2 sticks or........you get where I'm going, right.

may not need new mem, but it should be addressed. Mem is so cheap now that it isn't going to dent the $450 much anyway.
 
Quagmire LXIX said:
240 pin, true enough. The OP shows "12GB Corsair XMS DDR3 1600", idk if their 1.5 or up to 1.65, if their 6x2 or 3x4, so he may be able to use 4x2 and have an extra 2 sticks or........you get where I'm going, right.

may not need new mem, but it should be addressed. Mem is so cheap now that it isn't going to dent the $450 much anyway.
Click to expand...

that's fine, toss in $35 for 8GB or $70 for 16GB of new memory. Sell the old sstup for $300-$350 and now ur looking at $150~ to upgrade :p

I added the result in case u didn't see, Min FPS now drops below 60, and average goes from 104 to 77
 
i think its plenty, not worth the upgrade to achive that extra 7-10fps.
 
TroyX said:
i think its plenty, not worth the upgrade to achive that extra 7-10fps.
Click to expand...

huh :S it's 28fps, and i'm playing @ 3600x1920, like i said, it won't be crippled to the point of non-playable, but it's a significant difference, to the point where it might not be worth getting the 680s and sticking with the 580s
 
Digital Viper-X- said:
Here you go! Now this was done with an empty map, I only imagine a full 64man server would be even heavier on the cpu. Big enough difference for ya?

Typod on the Min/Max/Avg, labeled both as 4.8, the one on the right is 3.8.


3848.png
Click to expand...

All you're showing is 4.8 is considerably better than 3.8, nobody would question this. Even picking that test, what I see with 3.8 is very playable fps, I'd bet Kyle would throw up a patented no discernable difference in game play. So you're just not hearing or don't care about my point and continuing would be irrelevant, thanks for the test though.

ebeattie, throw away your 930, it's garbage :rolleyes:

Digital Viper-X- said:
that's fine, toss in $35 for 8GB or $70 for 16GB of new memory. Sell the old sstup for $300-$350 and now ur looking at $150~ to upgrade :p

I added the result in case u didn't see, Min FPS now drops below 60, and average goes from 104 to 77
Click to expand...

Wow, who is buying this old system for that much when they could spend another 100-150 for SB goodness.

Digital Viper-X- said:
huh :S it's 28fps, and i'm playing @ 3600x1920, like i said, it won't be crippled to the point of non-playable, but it's a significant difference, to the point where it might not be worth getting the 680s and sticking with the 580s
Click to expand...

At 3600x1920, is the OP even thinking about high res.
 
Last edited:
Digital Viper-X- said:
that's fine, toss in $35 for 8GB or $70 for 16GB of new memory. Sell the old sstup for $300-$350 and now ur looking at $150~ to upgrade :p

I added the result in case u didn't see, Min FPS now drops below 60, and average goes from 104 to 77
Click to expand...

Not to be pedantic, but you're inappropriately rounding one of them up and one of them down. Maybe a bit more fair to say 104 and 78. But hot damn that's quite the difference, almost 30fps just out of 1ghz in the CPU.
 
Quagmire LXIX said:
All you're showing is 4.8 is considerably better than 3.8, nobody would question this. Even picking that test, what I see with 3.8 is very playable fps, I'd bet Kyle would throw up a patented no discernable difference in game play. So you're just not hearing or don't care about my point and continuing would be irrelevant, thanks for the test though.

ebeattie, throw away your 930, it's garbage :rolleyes:



Wow, who is buying this old system for that much when they could spend another 100-150 for SB goodness.


At 3600x1920, is the OP even thinking about high res.
Click to expand...

Again, Why bother getting 2 GTX 680s when your 580s are already being held back? I'm willing to bet it wouldn't make much of a difference other then power / heat

Also, since I am playing at 3600x1920, I have more GPU demand vs playing at say 1920x1200 which will show more CPU demand.

What if the OP is playing @ 120hz 1080p? are you going to tell him having an average for 120fps+ is useless? I made my point, @ that speed the CPU will hold the GPUs back, and upgrading to 680s will be a waste in most cases.

Look at any buy /sell forums and see what i7s are going for, mobo and cpu easily sell for $300, Hell I just sold my 920 / p6x58 premium for $350 with 12GB of ram.
 
what resolution is OP using?

ViperX, i think its better if you show him what he's missing at the resolution that he plays with.
 
Thanks for the test DV-X. You have definitely shown that the 3.8 is bottlenecking the capability of two 680's when compared to the 4.8.

Based on the OP's needs (resolution, desired FPS, etc.) it could make a huge difference on whether it's worth the upgrade or not.

Nice job showing up in the thread with some results and let the OP distinguish what is relevant or not based on that. Worth a heck of a lot more than opinion of what a bottleneck really is.
 
Digital Viper-X- said:
Again, Why bother getting 2 GTX 680s when your 580s are already being held back? I'm willing to bet it wouldn't make much of a difference other then power / heat.
Click to expand...

1) New features. (3+1 being a biggie in my book)
2) Because he would get more performance out of the 680s than the 580s on the same system.
3) Have new gpus ready when he does upgrade platform, then gets to enjoy a whole new level of performance when he can crank up a SB (cough IB) to 5GHz.

Are you saying that there will not be a performance increase with the 680s and he would be stuck at the same fps as the 580s because he can only clock at 3.8? I'll be testing this myself, but at 4GHz, I know I will see a good performance increase.

Digital Viper-X- said:
What if the OP is playing @ 120hz 1080p? are you going to tell him having an average for 120fps+ is useless? I made my point, @ that speed the CPU will hold the GPUs back, and upgrading to 680s will be a waste in most cases.
Click to expand...

Why would I, are you just trying to insult me with that question. So you made a point that is commonly known, slower cpu speeds will hold back on gpu performance than higher cpu speeds. You didn't make the case that he wouldn't enjoy a couple 680s and their advantages because his system is simply too slow to use them.

And I have no idea why the hell I'm still here, the OP is nowhere in his thread.


Digital Viper-X- said:
Look at any buy /sell forums and see what i7s are going for, mobo and cpu easily sell for $300, Hell I just sold my 920 / p6x58 premium for $350 with 12GB of ram.
Click to expand...

To someone with no interest in a SB I guess, or someone who realizes that quality SB components comparable to the quality of those X58 components won't be $450-500.
Good for you on the buyer.


ELox said:
Thanks for the test DV-X. You have definitely shown that the 3.8 is bottlenecking the capability of two 680's when compared to the 4.8.

Based on the OP's needs (resolution, desired FPS, etc.) it could make a huge difference on whether it's worth the upgrade or not.

Nice job showing up in the thread with some results and let the OP distinguish what is relevant or not based on that. Worth a heck of a lot more than opinion of what a bottleneck really is.
Click to expand...

Ahh the peanut gallery :p

Yah Quagmire showing up in a thread to make Viper feel the need to show back up in this thread with some results instead of just his opinion that 3.8 would be a bottleneck based on 3.6 implying that the OP wouldn't see worthy improvements
 
Quagmire LXIX said:
1) New features. (3+1 being a biggie in my book)
2) Because he would get more performance out of the 680s than the 580s on the same system.
3) Have new gpus ready when he does upgrade platform, then gets to enjoy a whole new level of performance when he can crank up a SB (cough IB) to 5GHz.

Are you saying that there will not be a performance increase with the 680s and he would be stuck at the same fps as the 580s because he can only clock at 3.8? I'll be testing this myself, but at 4GHz, I know I will see a good performance increase.



Why would I, are you just trying to insult me with that question. So you made a point that is commonly known, slower cpu speeds will hold back on gpu performance than higher cpu speeds. You didn't make the case that he wouldn't enjoy a couple 680s and their advantages because his system is simply too slow to use them.

And I have no idea why the hell I'm still here, the OP is nowhere in his thread.




To someone with no interest in a SB I guess, or someone who realizes that quality SB components comparable to the quality of those X58 components won't be $450-500.
Good for you on the buyer.




Ahh the peanut gallery :p

Yah Quagmire showing up in a thread to make Viper feel the need to show back up in this thread with some results instead of just his opinion that 3.8 would be a bottleneck based on 3.6 implying that the OP wouldn't see worthy improvements
Click to expand...


so, if the GTX 580s are already being bottlenecked by the CPU, Why spend $1000 on gpus that will be even more bottlenecked? Yea I don't think he will get much more performance, Why not just upgrade the CPU and stick with his 580s? He will probably get a similar boost for half the price.
 
What exactly did you do in the empty map? Just curious, because these benchmarks are completely the opposite of what I've seen on the few websites that have done benchmarks with CPUs for battlefield 3. If you were in the middle of a giant battle with physics exploding all over the map I could understand some difference, but if it's just an empty map it's almost the same as single player.

http://www.techspot.com/review/458-battlefield-3-performance/page7.html

Shows the I3 as powerful as anything else. Even the A6 is keeping up with everything else. I guess they could be GPU limited, but it's just 1680x1050 with no AA or anything.
 
Hi all, thanks for such a great debate! I currently game on 1920x1200 resolution @ 60htz. I typically have VSYNC engaged and for alot of the maps, its does absolutely fine. but I notice on Strike at Karkand TDM, when the buildings start getting destroyed, I see my frames dip below 50fps and it gets very stuttery when i run near and look at the wrecked buildings. I currently have my GPUs set to about 830Mhtz (over an already factory overclock @797Mhtz). That being the case I think it was more CPU limited.

I did OC my 930 to 4Ghtz yesterday and noticed a decent improvement during those situations. At this point, I do think that my 930 is the limiting factor in performance. Ill probably wait until Ivy Bridge hits to start thinking of an upgrade over my current CPU and mobo. Definitely going with new RAM too as my 6x2GB will be staying with this 930 rig.
 
Pylor said:
What exactly did you do in the empty map? Just curious, because these benchmarks are completely the opposite of what I've seen on the few websites that have done benchmarks with CPUs for battlefield 3. If you were in the middle of a giant battle with physics exploding all over the map I could understand some difference, but if it's just an empty map it's almost the same as single player.

http://www.techspot.com/review/458-battlefield-3-performance/page7.html

Shows the I3 as powerful as anything else. Even the A6 is keeping up with everything else. I guess they could be GPU limited, but it's just 1680x1050 with no AA or anything.
Click to expand...

I run through the map, blowing up some cars, tank in the enemy camp, jumping off the bridge, and shooting out certain windows etc. Unfortunately there is no other way to properly test multiplayer without something changing. The only way I can see that happening, is getting 32people into a private server, and having them do the same thing for every run. Maybe there is some bot program to use?

I did run through Gulf of Oman with both AMD and NV cards with 8 players, I can put the results up later, the guys were from my clan on BF3 so they sort of cooperated and helped me out :)
 
Quagmire LXIX said:
Ahh the peanut gallery :p

Yah Quagmire showing up in a thread to make Viper feel the need to show back up in this thread with some results instead of just his opinion that 3.8 would be a bottleneck based on 3.6 implying that the OP wouldn't see worthy improvements
Click to expand...

Well I'm just saying instead of your opinion which is all well and good you could have posted some benchmarks to back it up. Since I'm not a hypocrite I will put my money where my mouth is. (Not saying you are at all, just saying I would be for asking you to if I didn't contribute more than opinion)

For instance here's a review of 680's in SLI done on a i7 920 at 3.8Ghz.

http://www.techpowerup.com/reviews/ASUS/GeForce_GTX_680_SLI/1.html

There's a few reasons I would find this information useful if I was the OP.

I could compare the results of my (OP's) 580 SLI setup with those of the 680 SLI in the review with what sounds like a very similar setup and see if there is a big performance jump.

I could also see how for example in BF3 at 1920x1200 (op's res) that one 680 gives 70FPS and two 680's gives 129 FPS. In this example at the same res and with a similar setup as the OP is really looks like two 680's are still achieving a very good performance increase over 1 680 (85%) which doesn't seem like a significant bottleneck to me.

The OP can then decide if he feel that two 680's are being bottlenecked significantly by the 920 @ 3.8 based on these results.

Happy now?
 
Last edited:
I know Im happy now! From what I know about the 920 and 930, it seems that base clock is the only difference. I have my clock speed the same as the tech power up review.

My gawd! GTX 680 SLI is darn near 2x faster than my 580s...... and theres an even MORE powerful card coming out? Simply incredible!
 
You must log in or register to reply here.
Back
Top