A baker’s dozen of GTX 960
When you have this kind of hardware on hand, you might as well overclock it.
Back on the launch day of the GeForce GTX 960, we hosted NVIDIA's Tom Petersen for a live stream. During the event, NVIDIA and its partners provided ten GTX 960 cards for our live viewers to win which we handed out through about an hour and a half. An interesting idea was proposed during the event – what would happen if we tried to overclock all of the product NVIDIA had brought along to see what the distribution of results looked like? After notifying all the winners of their prizes and asking for permission from each, we started the arduous process of testing and overclocking a total of 13 (10 prizes plus our 3 retail units already in the office) different GTX 960 cards.
Hopefully we will be able to provide a solid base of knowledge for buyers of the GTX 960 that we don't normally have the opportunity to offer: what is the range of overclocking you can expect and what is the average or median result. I think you will find the data interesting.
The 13 Contenders
Our collection of thirteen GTX 960 cards includes a handful from ASUS, EVGA and MSI. The ASUS models are all STRIX models, the EVGA cards are of the SSC variety, and the MSI cards include a single Gaming model and three 100ME. (The only difference between the Gaming and 100ME MSI cards is the color of the cooler.)
Jenga!
To be fair to the prize winners, I actually assigned each of them a specific graphics card before opening them up and testing them. I didn't want to be accused of favoritism by giving the best overclockers to the best readers!
The Testing Process
Our overclocking process was identical for each card. We used EVGA's Precision X to overclock each card to avoid any conflicting issues with different software packages, but we were also assured by all parties (NVIDIA, EVGA, ASUS, MSI) that our software selection should have no effect on the results. In other words, ASUS cards don't necessarily overclock better or worse with ASUS software (and so forth).
My step by step methods are pretty simple: install the card, starting turning GPU clock up until instability occurs, starting turning memory clock up until instability occurs. Interestingly, because of the power efficiency of the GM206 GPU there was no need to adjust or modify either the power target or GPU temperature target – we never got close to maximizing those in our overclocking at all. Of course I also maximized the memory voltage offset in Precision X as well. (Note that both the EVGA and MSI cards allowed for +100 mV offset while the ASUS STRIX cards only allowed +50 mV.)
Rather than reporting GPU or memory offset results, our graphs will show the peak frequency that was reached as a result of the offset. Because the three different GTX 960 options have different starting base and Boost clocks, our offsets ranged from +210 MHz down to +140 MHz, though the resulting clock rates are very close. In my testing the GPUs were able to maintain their peak clock rate in perpetuity – they don't seem to waver or bounce around in the way we are used to seeing cards like the GTX 780 do. Hence, we are only reporting a single maximum GPU and memory clock speed for each card.
Our stability testing was done with Metro: Last Light, a demanding PC game that quickly allows us to enter a repeating benchmark mode. To be deemed "stable", the test run had to complete five complete iterations, which is equivalent to about 15 minutes of game time. Obviously, to be 100% sure of the stability of any overclock, you should burn-in each card for a longer period of time, but the task of getting through 13 cards forced us to lower it down to something more reasonable.
Results and Distribution
After many hours of testing, the results are shown in the two graphs below. First up is the GPU clock rate, sorted by lowest to highest peak overclock. There is a secondary blue line that is overlaid, which shows the peak temperature of the GPU during that overclocked period (with a scale on the right side of the graph).
Click to Enlarge
Remember that the reference base clock of the GTX 960 is 1127 MHz and the reference Boost clock is 1178 MHz – that should help put these overclocks in perspective. The worst result in our overclocking testing game with one of the four MSI GTX 960 100ME/Gaming graphics cards, hitting just 1501 MHz. The best result was a 1576 MHz overclock, also from an MSI card, which should show you right away that picking a particular hardware vendor with the hopes of getting the "best" overclocker just isn't the way to think about it.
The 1576 MHz result is also the biggest outlier – it differs from its immediate neighbor by 18 MHz. The next largest gap we see in consecutive results is 15 MHz, going from 1504 MHz to 1519 MHz on the far left. The median result is 1542 MHz.
Temperatures on the overclocked cards ranged from 61C to 75C with the EVGA SSC card the only one to get above the 70C mark. The coolers from ASUS and MSI are definitely doing a better job of keeping the GPU cool, but there is absolutely nothing wrong with the 75C level that EVGA reaches. That temperature is a good margin away from the 80C target that NVIDIA defaults to.
Click to Enlarge
Our memory results aren't sorted independently but, instead, follow the exact same ordering as the GPU results above. This allows you to see if there is any correlation between maximum GPU clock and maximum memory clock – and the answer is clearly no. Our highest memory overclock is 2073 MHz, or actually 8.292 GHz effective rate! That's well above the 7.0 GHz reference rate but the card that achieved that was only able to hit the 1542 MHz GPU clock. On the other hand, the top GPU clock card, MSI #1, only hit 1976 MHz for memory speeds.
Performance Advantage?
How does this kind of overclocking change the gaming experience of using the GeForce GTX 960? To find out we compared our top overclocked card at 1576 MHz peak GPU clock and 1976 MHz peak memory clock to the retail ASUS GTX 960 card that is already pre-overclocked out of the box. Without changing settings, the ASUS STRIX card hits a peak GPU clock of 1404 MHz and a peak memory clock of 1800 MHz.
In Metro: Last Light at Very High quality settings and 2560×1440 resolution, the average frame rate jumps from 35.5 to 40.1 with the additional overclocking we were able to achieve. That is a 12.9% increase in average frame rate which matches up nicely with the 12.2% GPU clock speed increase and 9.7% memory clock speed increase.
Final Thoughts
Though our testing group is only 13 different GTX 960 cards, it is likely a bigger sample size than just about any other outside source, excluding companies like NVIDIA, ASUS, EVGA and MSI. My findings indicate that there is as pretty wide margin of overclocking capability in the GM206 GPU, ranging from 1501 MHz to 1576 MHz sustained Boost clock. That gap of 75 MHz is approximately 5% of the median clock speed of 1542 MHz. How that compares to other graphics cards, like the GTX 980 or GTX 970, is hard to say as we didn't have nearly the quantity of those products to test.
Woot monkey supervises from the chair
As is always the case, overclocking capability is never guaranteed by the card vendor. Just because you get a product that is only hitting around 1500 MHz does not make it a warranty concern. But at least now you know where your own GTX 960 card sits based on our days of testing.
There are also a couple of interesting closing facts for you to consider. First, the 1576 MHz maximum Boost clock is 33.7% faster than the "typical" Boost clock of 1178 MHz on the reference GTX 960 specifications. That is really impressive but, at the same time, makes me wonder why NVIDIA would consistently leave that much performance on the table. Ostensibly, it is to allow board partners to differentiate in coolers and performance, but it might have benefited NVIDIA in the initial comparison to AMD's Radeon R9 285.
Second, I did notice an interesting phenomenon around the GPU clock offset and chip to chip clock results. Even from the exact same graphics card models, setting a specific GPU clock offset does not result in similar frequencies. For example, with our five different EVGA GTX 960 SSC cards, with a GPU offset of +90 MHz, the resulting peak Boost clock speeds were 1544 MHz, 1558 MHz, 1519 MHz, 1533 MHz and 1507 MHz. But that did not indicate where the final stable overclocked speed would sit. One card would reach as high as 1547 MHz (with a +130 MHz offset) while others would find the +90 mark to be their limit. Essentially, the GPU offset is a rather useless metric for comparing card clock rates, even among identical products.
I hope you find these GTX 960 overclocking results interesting. Maybe, we'll be able to use this story to show manufacturers the benefits of getting a large collection of identical products going forward. This might be the first time I have been able to show overclocking output I can be confident about – single CPU or GPU results are far less useful.
Wonder which of the MSI’s was
Wonder which of the MSI’s was the one I happened to nab. 😛
I was actually surprised at the EVGA SSC cooling and overclocking, you would imagine that would be a better binned chip.
I doubt the SSC is binned at
I doubt the SSC is binned at all. Now time to choose one for the HTPC.
While the sample size is only
While the sample size is only 13 cards, so this is to be taken with a grain of salt, the EVGA card binning appears to place them closer to the median of maximum overclocked speeds.
Thanks for the review of
Thanks for the review of GM206 clock rates and relative potential. Your conclusion regarding the performance on the table was appreciated. I began to think the same as I read through the article.
Now I shall rumor-monger, donning my tin-foil hat and bringing my fingers to a steeple, perhaps this overhead is left available for a 960Ti.
I kid, it is most likely left for board partners to do crazy things.
The only thing missing is
The only thing missing is power consumption. That’s the real reason for the stock underclocking, to keep up NVIDIA’s image of being the most power efficient. Every 970 can hit 1400MHz, most can do 1500, so why not make that the stock clock? Answer: they lose their marketing edge to AMD on power efficiency.
Well, if you look at power
Well, if you look at power draw when overclocked maxwell doesn’t take that much more power in most cases.
Wouldn’t the GPU vendors send
Wouldn’t the GPU vendors send you their best overclockers out of what they have to make you think this card is very overclockable? Although these results seem reasonable since we know how well maxwell overclocks..
The key for this story is
The key for this story is that they did not know I was going to do this with these graphics cards. They assumed they were prizes going to winners only.
So in theory it should be pretty accurate!
Also, it’s worth pointing out
Also, it's worth pointing out (and I'll add this in the story above) that the samples that were sent to me for review from MSI, ASUS and EVGA were actually NOT in the top half of our results above. So either these companies are not cherry picking parts for the media or they are pretty bad at it.
That’s quite a revelation. I
That’s quite a revelation. I had always assumed that companies would send their better samples.
if I were to control the
if I were to control the media I would make sure that the what is considered the most hardcore reviewers got the best samples and then regular reviewers who do not pursue high overclocks as much got better than average samples. this would cater to what each group cares about and also does not arise suspicion.
they will always be one step ahead of you. mind control…
What I would like to see as
What I would like to see as an addition for future overclocking articles on PCPer is a look at the increase of power consumption.
Clocking the GTX 960 above 1450 MHz boost does not make much sense, as the performance gain decreases, while the power consumption increases.
Good Job Boyz! I hope the
Good Job Boyz! I hope the prize winners all appreciate you hard work:) DId you include the signed “PCPER certified OC results” in the box?
BTW…I don’t believe Ryan did all this testing!
You would be wrong SIR, I did
You would be wrong SIR, I did ALL of this testing. 🙂
Wow! Did the OC rating notes
Wow! Did the OC rating notes all get signed with a kiss 🙂
No kisses, but definitely
No kisses, but definitely hand written notes.
It’s not like you only have a
It’s not like you only have a single test bed…oh wait.
Not sure if cards were
Not sure if cards were stacked or GPU orgy…
>CTRL+F
>No Zotac’s AMP
>CTRL+F
>No Zotac’s AMP Edition (which is THE best 960).
LOL, INSTANT FAIL.
P.S. Even though 960 as itself is a completely dead-on-arrival piece of garbage, still, if you’re so stupid as to actually spend money on one, might as well get the most best one out of them all, and that one is Zotac’s AMP Edition and no other.
and what’s your source of
and what’s your source of this statement?
1. I’ve personally got to try
1. I’ve personally got to try all of the 960 models which were released so far. Including Colorful’s absolutely unnecessary monstrosities. That’s what I do for my work.
2. 960 is a card that only looks somewhat decently in an ITX/HTPC setup. If you’re using it for anything besides just ITX/HTPC build – you’re doing it absolutely wrong.
3. Zotac’s AMP Edition IS the most best 960, simply because it’s THE most best 960 for ITX/HTPC (and as I’ve already said above, if you’re using 960 for anything else besides just ITX/HTPC – you’re doing it absolutely wrong and you obviously don’t know jack about building of the quality hardware setups).
So a card that performs just
So a card that performs just slightly slower than last year’s GTX770 is now only good for HTPC. Right.
Hey Ryan,
Maxwell has the
Hey Ryan,
Maxwell has the same +/- 13 core clock bin as Kepler. So getting two cards with actual clocks not within a multiple of 13 of each other isn’t possible as far as I know (Jacob confirmed this also, as did skyn3t). Did you check the actual clocks in the GPU-Z sensors tab? PrecisioX settings applied aren’t necessarily always the ones effected.
Yup, I used GPU-Z for all the
Yup, I used GPU-Z for all the clock speed monitoring.
So talking to a few more
So talking to a few more guys, it does seem like a Boost enabled BIOS allows for <13 bins in core clocks with Maxwell. Very cool indeed, thanks for the article. So far all the Boost disabled BIOS have meant no one really noticed this or cared much for it.This info will help tweak 3rd party BIOS more for anything Maxwell.
Very interesting! Please
Very interesting! Please update me (email maybe?) with what comes of that!
960 Pizza tower
960 Pizza tower
I read this article solely
I read this article solely off the uniqueness of the methodology…..
Be careful Ryan don’t look a gift horse in the mouth…at least not on a 960…if ya know what I mean.
MarkT out
Ummm….I do…not…know
Ummm….I do…not…know what you mean. lol
lol
lol
idk maybe it is a little long
idk maybe it is a little long in the tooth.
I think people just wanted
I think people just wanted more from the GTX 960, but the way I see it, it’s a great intro card, yes it would be nice if they had 256 bus memory, yes it would have been nice to have 4GB DDR5 memory, but that’s why they have the GTX 970 and 980, I bought a GTX650Ti Boost in Sep 2013 for $179 which was a highly recommended at the time, the 960 cost about the same but looks to be a good over clocker. To me it looks to be a very good card for 1080P and maybe 1440 also, just may thoughts, happy computing to everyone 🙂
Hello PCper team!
I am
Hello PCper team!
I am interested in knowing whether one vendor’s cooling solution is superior to another. If it’s truly a random silicon lottery (which I admit I don’t fully buy), then it comes down to the cooling right? So about the temperature testing; was the temperature taken at the same overclocks for all cards? A card could be running a higher OC and have higher temperatures which leads us to think that card has inferior cooling, right?
On a random note, I stumbled across a forumer who said he asked if you would check out Skyrim on a load of mods with ENB for testing components and I would like to add that I approve of this idea. 🙂
This is a fun and unique test
This is a fun and unique test that seems to have beaten other sites to the punch. Good video and good initial sample size for a 3 week old product!
Certain points in the benchmark show a 5+ FPS advantage to the stock card – if only for a few tenths of a second?
I’m stoked for my EVGA 02G-P4-2962-KR – eVGA’s “ITX 960” with a factory boost of 1279. Clearly not for long!
Interested to see what temp will my shortie card max out at.
GM 206 GPU is the “First discrete GPU” with full H.265 = Huge value for light 1080 gamers / eager 4k video users like me
Also happens to be a very capable CAD and extremely good casual render card, as would be expected.
Will DX12 low-level APIs make this card and the Ti variants even better values?
-ZCC
i am not sure if this result
i am not sure if this result consider “not bad” or not
only able to get GPU clock from 1253 mhz to 1403 mhz
and memory clock from 1753 mhz to 1803 mhz …
that is the stable clock speed i can get without adding voltage.
PNY 960 with stock cooler
PNY 960 with stock cooler here, 1501mhz Overclock out of the box and rock stable since with temps around 48ºc-60ºc
i did ramp up the fan delta so that its always at 50%+ and i can barely hear it and it sits 3 feet from me on the desk (big ventilated case and not so quiet environment).
gave me good speed bumps in games like stalker on heavy areas from 45ish to 53ish fps (around 10-15%+fps overall in most games)
i wont overclock the ram since i know its not cooled at all, and i already have a nice boost for free 😉
Hi,
I recently bought the
Hi,
I recently bought the evga ssc acx2.0 version was looking up the internet how to do these things as it is also my first time with the
new turbo boost technique.(was using 560’s for so long)
I must say this review is very usefull and efficient , perfect for what I was looking for.
I got mine while testing at peak bios mode 2 @1594 – inspector estimation turbo boost at 166x near 1700 but there started to be artifacts at heaven benchmark to bad :(.
so now @1583 I think it is stable needs more testing.
hi man. hope your day is
hi man. hope your day is going well . i have a question cause i’m so newbie i just wanted to know what is a good stable overclocking for EVGA GTX 960SSC 4GB . i’m doing like that +50Mhz GPU clock offset and +500 memory clock offset and 110% power and max temp to 94 . is it anyhow good or should i change anything . i would be thankfull for an answer . greetings
my memory clocks at 3803 Mhz,
my memory clocks at 3803 Mhz, thats is this because I have a 4gb card?
my memory clocks at 3803 Mhz,
my memory clocks at 3803 Mhz, thats is this because I have a 4gb card?