Galaxy GTX 780 SLI vs Evga GTX 680 4GB SLI (Synthetic Benchmarks)

Joined
May 27, 2007
Messages
615
Before getting started I'd like to give a big thanks to Kyle and Tiger Direct, as well as Ric, Shane, and Andrew (from Galaxy) for putting on the GTX 780 launching party in Dallas today. As always, it was fun to talk with the reps / fellow gamers, and as a result, I may have become a Galaxy convert in the process. :p

While I was out in Dallas I picked up a pair of Galaxy GTX 780's (for 10% off retail) to test against my 4GB GTX 680's from EVGA. For this thread I'm only going to show the results of a few synthetic benchmarks. This is mainly due to time and the fact that I do not yet have many new titles to test. Since I'm still working through a sizable backlog of games (which is a problem I'm sure most of you can relate to) I haven't bothered to pick up anything new. However, if there is a newer game on sale that you'd like to see, just let me know and I may be able to pick it up / test it for you.

---Anyway, onto the meat of this thread.---

Nvidia Driver Version: 320.18

GPU Specs:
2x EVGA GTX 680 4GB FTW Edition cards (P/N: 04G-P4-3687-KR)
2x Galaxy GTX 780 3GB (Reference) cards (P/N: 78NNH5DN00G0)

PC Specs:
-See 'Desktop--' specs in Signature-

Stock and Overclock Settings
GTX 680's Stock: GPU Boost Clock* - 1176
GTX 680's Stock: Memory - 3005


GTX 780's Stock: GPU Boost Clock* - 993
GTX 780's Stock: Memory - 3005

___________________
GTX 680's OC: GPU Boost Clock* - 1228 (+60)
GTX 680's OC: Memory - 3351 (+350)


GTX 780's OC: GPU Boost Clock* - 1137 (+150)
GTX 780's OC: Memory - 3105 (+100)


* According to my G15's LCD panel via EVGA Precision X (v. 4.20)

Power Target:
GTX 680's (Stock and OC): 131% - (Max)
GTX 780's (Stock and OC): 106% - (Max)

Fan and Temps
For both tests I set the fans to 85%, This is because the 680 FTW's are limited to 85% in Precision X and I wanted to compare the amount of noise generated between both cards at the same fan speed. I can say, "objectively", that the 780's are slightly quieter at 85% than the 680's but they ran much hotter at load (worth it for the performance difference though as you will see);
GTX 680's Overclocked: 72C (GPU1) / 68C (GPU2)
GTX 780's Overclocked: 83C (GPU1) / 78C (GPU2)
Ambient: ~24C
I forgot to capture temps at stock settings so only have OC'd temps listed.


***Graphs were done in Excel because this is my first time to post comparative benchmarks and I didn't feel like looking for a free benchmark graphing utility. Suggestions however are welcome***

Valley Demo (Ultra Quality, 8xAA, 2560x1600):

valleyfps780vs680sli.png


Heaven Demo (Ultra Quality, 8xAA, Medium Tesselation, 2560x1600):
heavenfps780vs680sli.png


3D Mark Vantage (Basic Edition; default settings):
3dmarkvantagescore780vs.png


At this point I haven't actually played any games with the GTX 780 cards (they were just placed in my PC this evening) so I can't yet comment on any actual gameplay experience, but I will provide some feedback very soon. I have a heavily modded Skyrim install, that regularly hits 2.8GB of VRAM, which I plan to test over the next few days and will report my results on as well.


In Conclusion: I'm extremely happy with my GTX 780's and fully plan to sell off my GTX 680's after doing some gaming. From the synthetic benchmarks, my performance differences are are ridiculous (39% Stock / 40% OC increase for Valley, 54% Stock / 57% OC increase for Heaven) but I recognize that these will not translate 100% to in-game performance. When I have some in-game results I will post those in a separate thread.

I hope this review was helpful to at least a few folks. I'd welcome any feedback and will try to post results from some games in the next few days as well.
 
Last edited:
Sweet. Thanks for doing the tests....I'm looking forward to the game ones
 
Excelent test.... Hoping to see soon a review with games.. Regards...
 
Could you possibly do a run of both benchmarks using single and SLI mode, at 1080p? I am curious how my 580 SLI setup is doing these days compared to the top cards.
 
Try turning the AA mode to supersampling. That will really show weather the extra gigabyte of RAM has an advantage.
 
Try turning the AA mode to supersampling. That will really show weather the extra gigabyte of RAM has an advantage.

You do not want weather in your pc. Bad thing. Now WHETHER or not the extra memory works or not is certainly important for future proofing.
 
Been a busy weekend and haven't had much time to test. Tuesday will probably be my next chance to do the game benches.

Could you do a 680 sli vs a single 780?

Yeah I can probably throw in a few single 780 tests for comparative purposes.

I won a third GTX 780 at the [H] event in Dallas (well, my wife won it for me really) and I will be trying tri-SLI as soon as I get everything switched out into a new case (not enough expansion slots in current case). Unfortunately it will be a slightly gimped tri-SLI as my mobo can only run the 3rd card in 4x PCI-e 2.0 but it should still provide a noticeable increase.
 
Congrats on your SLIed 780s! Specially at 10% off.

Very Jealous.

This is good info. I am waiting for some moneis to come in.

After I am going 3x1440p.

So from 680 I cannot make up my mind which way to go. So many options.
 
If it wouldnt be too much trouble could you please see what avg FPS you get on furmark 1080p burn in test? I currently own two of these http://www.newegg.com/Product/Product.aspx?Item=N82E16814162118 and i am strongly debating if i should just keep them or upgrade and sell. Keep in mind i dont run 1080p res in my setup i have three asus vg248qe 144hz monitors however i always look for furmark avg fps scores in order to see if its worth the upgrade. If i could i would tri sli and get another one of my cards but i cant due to my sabertooth z77 only supporting 2 way sli. Anyways if u can get back to me it would be much appreciated.

Thanks! ;)
 
Last edited:
Back
Top