Are you interested in how well the latest nVidia card scales in SLI? Thanks to Zotac we can find out.
Continue Reading
Zotac GTX680 SLI Review
- Thread starter tinytomlogan
- Start date
Are you interested in how well the latest nVidia card scales in SLI? Thanks to Zotac we can find out.
Continue Reading
Excalabur50
Well-known member
Nice review as always and it's still nice to see the brands evenly matched It's been awhile and as you have said many times in your reviews it don't matter which camp you want they are both as good as each other if you want Nvidia go that way if you want AMD go that way it great to see.
Rastalovich
New member
I don't understand the CPU limited quote, and I'll be glad when those oc-bench-score-being-slower-than-the-stock-card disappear off the comparison charts. Both the bios numbers the same on these ? (should be easy to see on the stickers)
yassarikhan786
New member
Really nice review as always Tom. Not too impressed with the temps tbh. 77 C on an open test bench seems too hot imo, but since the fans don't really spin up making the cards uber quiet, it's a reasonable compromise.
dugdiamond
New member
just imagine the fan power needed to cool these in a closed case ????
YogevSkill
New member
AMD, Nvidia, comes down to specs at the end of the day, 3Dmark favor's the strong, and so do I.
SWeet review TTL. Thx.
SWeet review TTL. Thx.
VonBlade
New member
We report the findings, for good, bad or indifferent. As long as cards continue to produce non-linear overclock results, we'll continue to have them in the graphs.
yassarikhan786
New member
Grim_Reaper
New member
Great review! As usual
Yep. 7970 seems a bit cooler, but a bit noisy (stock). The only difference between 680 is that AMD has no any useful/useless features like 680 has. Still I see no reasons for fanboy pee-war.
Or- you can spin up the 680 fan, to make it as noisy and cool as 7970 does
Yep. 7970 seems a bit cooler, but a bit noisy (stock). The only difference between 680 is that AMD has no any useful/useless features like 680 has. Still I see no reasons for fanboy pee-war.
.
Or- you can spin up the 680 fan, to make it as noisy and cool as 7970 does
The reason the 680 preforms worse overclocked, is the overclock is to high for that load.
Just like Intel Speed Step
if the 680 has more headroom it overclocks higher(single thread apps), if it is loaded down (all cores active) it doesn't boost .
Somewhere there will be a line where a slight OC will let it boost to the max TDP, but pushing the OC just a few MHz more would cause the 680 to not boost at all, cutting performance.
Overclocking video cards just got harder, now you have to actually find a balance over your entire spectrum of applications and workloads instead of just find the highest stable MHz, because you could be cheating yourself out a bit of performance.
From Toms SLI videos the pair of 680s were never really at 100%, so they boost all they can, whatever Tom raises the base frequency, they still boost higher.
With a single card, it's workload is higher, so it doesn't boost as much, if at all.
Just like Intel Speed Step
if the 680 has more headroom it overclocks higher(single thread apps), if it is loaded down (all cores active) it doesn't boost .
Somewhere there will be a line where a slight OC will let it boost to the max TDP, but pushing the OC just a few MHz more would cause the 680 to not boost at all, cutting performance.
Overclocking video cards just got harder, now you have to actually find a balance over your entire spectrum of applications and workloads instead of just find the highest stable MHz, because you could be cheating yourself out a bit of performance.
From Toms SLI videos the pair of 680s were never really at 100%, so they boost all they can, whatever Tom raises the base frequency, they still boost higher.
With a single card, it's workload is higher, so it doesn't boost as much, if at all.
feathers632
New member
Having owned a Zotac 680 for 2 weeks, first week air cooled, second week liquid cooled, overclocking it brings a big performance boost. I get 500mhz+ on the vram and at least 188mhz on the core. Running kombustor DX11 immediately shows the performance boost with the minimum fps, average and max increasing significantly.
The figures I wrote down from my OCing on air:
Kombustor DX11 default clocks = 216min, 229 average.
+175core/+500vram = 247min, 266 average.
That difference is enough to make gameplay smoother on BF3 ultra on the most demanding 64 player map: Damavand Peak.
Yes, agree, if the clocks are pushed too far it can actually reduce performance but it's so easy overclocking and you quickly establish what too far is using Kombustor Dx11 mode (gives you immediate feedback of the min, max, average).
It's also important to use the EVGA OC scanner to check for artifacts. Although I had the vram at 650mhz, artifacts show past 500mhz.
Have been as high as 247+ on the gpu core.
The figures I wrote down from my OCing on air:
Kombustor DX11 default clocks = 216min, 229 average.
+175core/+500vram = 247min, 266 average.
That difference is enough to make gameplay smoother on BF3 ultra on the most demanding 64 player map: Damavand Peak.
Yes, agree, if the clocks are pushed too far it can actually reduce performance but it's so easy overclocking and you quickly establish what too far is using Kombustor Dx11 mode (gives you immediate feedback of the min, max, average).
It's also important to use the EVGA OC scanner to check for artifacts. Although I had the vram at 650mhz, artifacts show past 500mhz.
Have been as high as 247+ on the gpu core.
