NVIDIA GeForce GTX 1060 Founders Edition Review @ [H]

Cheapest one I can find in stock at Overclockers here in the UK is a Palit card at £269.99

However, the XFX 4gb RX480 is in stock (not sure if this is the one that can be unlocked to 8gb) @ £215.99 or the cheapest 8gb version in stock is a PowerColor one at £224.99 - both are the reference coolers, of course.

These all have free delivery, and quite frankly is the 1060 worth the extra £45 (presuming you're getting the 8gb card)......no, I don't think so.

Interestingly the Devil card is showing as a pre-order, but its £249.95 - as soon as this comes into stock, that is almost an essential purchase!
 
Couldn't agree more. Multi-GPU (be it on two separate boards or two GPU's on one board) has a place.

That place is "A single fastest GPU on the market is not fast enough so I need two"

Using two slower GPU's instead of a single faster GPU in order to get the same performance is - however - just plain stupidity considering the many drawbacks of multi-GPU implementations.

Some people just have to learn the hard way I guess.

I completely and 100% agree with you on the point that buying 2x cards right away to match 1 high end card makes no sense and is something only for the foolish or the most dense fanboys (looking at everyone who bought 2 RX 480's off the bat). Where I disagree is its value to people in this segment. When you are budget constrained to the point where you can only buy 1 card at a time, sometimes going CF/SLI can still be a good bang for the buck purchase if you wind up adding a second card down the road when they get cheaper. Adding that second card definitely adds headaches, but can also be a much bigger boost than an equivalent outlay on a newer card. In my personal past I went from 1x gtx 460 1GB to 2x rather than going to a 560Ti, and it was a better bang for my buck at that time. I think removing that option does detract from the potential value of a card in this segment.
 
I completely and 100% agree with you on the point that buying 2x cards right away to match 1 high end card makes no sense and is something only for the foolish or the most dense fanboys (looking at everyone who bought 2 RX 480's off the bat). Where I disagree is its value to people in this segment. When you are budget constrained to the point where you can only buy 1 card at a time, sometimes going CF/SLI can still be a good bang for the buck purchase if you wind up adding a second card down the road when they get cheaper. Adding that second card definitely adds headaches, but can also be a much bigger boost than an equivalent outlay on a newer card. In my personal past I went from 1x gtx 460 1GB to 2x rather than going to a 560, and it was a better bang for my buck at that time. I think removing that option does detract from the potential value of a card in this segment.

I'd argue that in this case, selling the first and and using the proceeds towards a faster single card would probably be wiser.

Or, maybe even putting the faster card on a credit card and paying it off by the time you would have bought the second card might be an option.

(The paying it off portion is important though. Don't run up debt kids! :p )
 
I'd argue that in this case, selling the first and and using the proceeds towards a faster single card would probably be wiser.

Or, maybe even putting the faster card on a credit card and paying it off by the time you would have bought the second card might be an option.

(The paying it off portion is important though. Don't run up debt kids! :p )

Well in my case, I probably could have sold my 460 for $100, then had to pay $250 for a 560 Ti, rather than spending $150 for a second 460 that gave me better performance in virtually every scenario, even with poor scaling (as long as there was an SLI profile at all). Granted, that was a few years ago and things have changed.
 
I completely and 100% agree with you on the point that buying 2x cards right away to match 1 high end card makes no sense and is something only for the foolish or the most dense fanboys (looking at everyone who bought 2 RX 480's off the bat). Where I disagree is its value to people in this segment. When you are budget constrained to the point where you can only buy 1 card at a time, sometimes going CF/SLI can still be a good bang for the buck purchase if you wind up adding a second card down the road when they get cheaper. Adding that second card definitely adds headaches, but can also be a much bigger boost than an equivalent outlay on a newer card. In my personal past I went from 1x gtx 460 1GB to 2x rather than going to a 560Ti, and it was a better bang for my buck at that time. I think removing that option does detract from the potential value of a card in this segment.

But if you can get that performance for roughly $200 less (if we go on the basis a RX480 is $250 & the 1080 is $699) then it starts to make a bit more sense (well, it does to me as I have a family to sustain, and not a lot of disposable income!)

IF the cost is roughly the same, then of course it makes total sense to get the single card.
 
But if you can get that performance for roughly $200 less (if we go on the basis a RX480 is $250 & the 1080 is $699) then it starts to make a bit more sense (well, it does to me as I have a family to sustain, and not a lot of disposable income!)

IF the cost is roughly the same, then of course it makes total sense to get the single card.
When dual card has visual issues like strange textures, unfixable uneven frame times, lack of support etc, you are demoted to 1 card.
Dual card is often less smooth unless you boost the framerate by another 20 to 50% above the target framerate. If the dual cards cant do this you may have a stuttery experience.

As long as you are happy to game on one card when issues occur, no problem.
But its not uncommon.
 
But if you can get that performance for roughly $200 less (if we go on the basis a RX480 is $250 & the 1080 is $699) then it starts to make a bit more sense (well, it does to me as I have a family to sustain, and not a lot of disposable income!)

IF the cost is roughly the same, then of course it makes total sense to get the single card.

The $200 savings may make sense...

...until you've tried SLI/Crossfire and realized how many headaches, how much stutter, and how low occasional minimum framerates it introduces.

Then you wind up wishing you'd just spent the extra $200 to begin with.
 
I'd go with the 1060 strictly because AMD drivers have always sucked for me. Nothing but problems with their drivers everytime I have a system with one of their cards.


I'd argue AMD's drivers suck more than Nvidia's when comparing Crossfire to SLI, but that they are fine in single card configurations. Certainly Nvidia's drivers come across as more polished, better laid out and easier to tweak, but AMD's work just fine for single cards in my experience. If they came out with something that rivaled or beat the 1080 (and had HDMI 2.0, as I need it for 4k) I would have no concerns buying it.
 
When dual card has visual issues like strange textures, unfixable uneven frame times, lack of support etc, you are demoted to 1 card.
Dual card is often less smooth unless you boost the framerate by another 20 to 50% above the target framerate. If the dual cards cant do this you may have a stuttery experience.

As long as you are happy to game on one card when issues occur, no problem.
But its not uncommon.

The $200 savings may make sense...

...until you've tried SLI/Crossfire and realized how many headaches, how much stutter, and how low occasional minimum framerates it introduces.

Then you wind up wishing you'd just spent the extra $200 to begin with.

Fair points well made.
 
I'd go with the 1060 strictly because AMD drivers have always sucked for me. Nothing but problems with their drivers everytime I have a system with one of their cards.

Seems like the tide has begun to change, AMD started sorting their shit majorly about 2 years back, now Nvidia is dropping the ball, quite a few unresolved issues, win10 driver was not 'game ready' on launch, DSR took months to fix etc etc..
 
Just wait till **insert future tech not heavily utilized** kicks in. Then the **insert AMD product** will REALLY shine.
Wasn't everyone using the same damn argument for the 390? - DX12 ZOMG SO MUCH FASTER THAN NVIDIA! - How'd that work out for you guys? That 390 kicking the crap out of nVidia cards yet? How about we start focusing on delivered performance and not potential performance? The GTX 1060 is a bit lackluster IMO, but if nvidia DOES improve their drivers I'm sure it will siphon sales from the 480.
Really? So tell us, between the 780, Titan and 780Ti vs the 290X, which one has been shown to "shine" brighter and longer?
 
oh wow. I didn't realize it was that unpopular.

And some number of those "players" were probably review sites running benchmarks.
Maybe most of them?

And you guys have been wondering why I have called AotS a benchmark rather than a game....since hardly anyone actually plays it?
 
Really? So tell us, between the 780, Titan and 780Ti vs the 290X, which one has been shown to "shine" brighter and longer?

I agree that the 290x has better performance today. Now, how many original 780/780ti/190x owners still have their card? Performance improvements 2-3 years later is fabulous, if you keep your card that long. Many enthusiasts don't.
 
  • Like
Reactions: noko
like this
I agree that the 290x has better performance today. Now, how many original 780/780ti/190x owners still have their card? Performance improvements 2-3 years later is fabulous, if you keep your card that long. Many enthusiasts don't.

Enthusiasts on a budget (or wannabe enthusiasts on a budget *cough*like me*cough*) do... :)
 
After reading this article - despite me usually being an Nvidia fan, I came away with the impression that for $199, most people should just buy the RX480. Its good compared to the 1060, and cheaper!

And I'm glad this is the case, as AMD really needs to be able to sell some of these cards. If the 1060 had crushed the RX480, AMD's financial position would have looked dismal.
First off, there were no $199 cards in that review. A ~$240 RX 480, a ~$300 1060 and a ~$280 980.

While the 1060 did in fact show up a bit slower in terms of avg framerate than the 980, if you are playing at 1080p or are willing to take the IQ hit at 1440p, it is still hard to tell you to buy a 1060 over a 980 just in terms of "future proofing" your purchase. From what we saw looking forward somewhat in terms of game, the 480 is looking damn good for the price. We gave the 480 an award and we did not give one to the 1060. That should be a bit telling.
 
Cheapest one I can find in stock at Overclockers here in the UK is a Palit card at £269.99

However, the XFX 4gb RX480 is in stock (not sure if this is the one that can be unlocked to 8gb) @ £215.99 or the cheapest 8gb version in stock is a PowerColor one at £224.99 - both are the reference coolers, of course.

These all have free delivery, and quite frankly is the 1060 worth the extra £45 (presuming you're getting the 8gb card)......no, I don't think so.

My apologies if this rant has already been posted in this thread. Didn't see it.

In Sweden the cheapest RX480 8GB (no 4GB variant here) is SEK 2735. The cheapest GTX1060 8GB is a Palit one, SEK 2799. The cheapest 1070 is SEK4500. Sweclockers (Sweden's biggest computer hw site) labelled the GTX1060 a "given choice" over the RX480 because as cheap, lower power draw, slightly higher performance (they did not test Doom with the vulkan patch).

Maybe Swedish retailers are simply great fans of leather jackets? I don't know, but the deviation from the recommended price in AMD's case and not Nvidia's is annoying. This country sometimes..
 
Ah, fair enough. I could have sworn I remembered the RX 480 launching at $199... Was that one of those theoretical press release prices that never materialized, or am I just misremembering things?

I'm still curious if you have any thoughts regarding the differences in minimum frame rates, and what that did to the overall game experience, as that has the potential of turning the results in many of the benchmarks on their head. Because when I look at minimum frame rates (which is what I usually do when GPU shopping) in this set of benchmarks (like the one I included above) the RX480 actually comes up way on top, like 3 times faster than the 980... Was that actually felt in game in the form of stutter or moments of low framerate, or was it just an anomaly in testing?

Those are for 4GB 480's, which are few and far between...Think like it was when trying to buy a 1080 or 1070 the first week or three. :LOL:
 
Ah, fair enough. I could have sworn I remembered the RX 480 launching at $199... Was that one of those theoretical press release prices that never materialized, or am I just misremembering things?

4GB is $199
8GB is $239+

They tested the 8GB version.
 
Ah, fair enough. I could have sworn I remembered the RX 480 launching at $199... Was that one of those theoretical press release prices that never materialized, or am I just misremembering things?

I'm still curious if you have any thoughts regarding the differences in minimum frame rates, and what that did to the overall game experience, as that has the potential of turning the results in many of the benchmarks on their head. Because when I look at minimum frame rates (which is what I usually do when GPU shopping) in this set of benchmarks (like the one I included above) the RX480 actually comes up way on top, like 3 times faster than the 980... Was that actually felt in game in the form of stutter or moments of low framerate, or was it just an anomaly in testing?
Well, you have to remember how the test was conducted. We found best playable on the 1060, then matched apples to apples with other cards. That does not mean we were at best settings for the 980.
 
Really? So tell us, between the 780, Titan and 780Ti vs the 290X, which one has been shown to "shine" brighter and longer?
What difference does that make? The 290X offered good performance out of the gate, the CF support was pretty crap IMO, but that launch wasn't "sure the performance isn't amazing right now but just you wait for this API to take hold" - and the same goes for the 780 / 780ti. Sure Mantle came later, but people weren't buying 290s in the hope that Mantle would make a mediocre card amazing. How long has the 390 shined? Did it ever shine? That's the first card I remember people getting giddy at the promise of future performance gains via API, how did that turn out? Buying a lackluster card today because it *might* be faster than the competitor's lackluster card tomorrow seems like pretty crap logic to base a purchase on. Unless you're getting paid interest or the price of the card will jump when these performance gains are realized, why take the gamble? Best case scenario you've put up with a lackluster card for some time, and now it performs better than it used to. Worst case you're stuck with a lackluster card that is quickly obsolete. Not much of an upside if you ask me. If your'e really that concerned about longevity what's a month or two? Wait and see if vulkan adoption takes off. Maybe it does, amd continues to optimize drivers while nvidia does nothing, and the 480 is the card to have. Or maybe amd spends most of their money working on the 490 rather than drivers, nvidia improves their drivers, and the 480 turns out to be a dog. Either way you've spent no money, you're no worse off than you are today, and can make an informed decision.
 
Ah, fair enough. I could have sworn I remembered the RX 480 launching at $199... Was that one of those theoretical press release prices that never materialized, or am I just misremembering things?

I'm still curious if you have any thoughts regarding the differences in minimum frame rates, and what that did to the overall game experience, as that has the potential of turning the results in many of the benchmarks on their head. Because when I look at minimum frame rates (which is what I usually do when GPU shopping) in this set of benchmarks (like the one I included above) the RX480 actually comes up way on top, like 3 times faster than the 980... Was that actually felt in game in the form of stutter or moments of low framerate, or was it just an anomaly in testing?

Which games specifically are you referring to? In general I did not feel stutter or lag or choppiness or slow responsiveness out of any of the video cards tested. What was playable on the 1060 was more than playable on the 980, and what was playable on the 1060 also matched up pretty well with what was also playable on the 480.

/edit I read back a bit at your previous posts, that dip on the 980 in Tomb Raider was very brief, it might have been in a spot where player control wasn't affected, a scripted sequence or such, I didn't feel it negatively impact my gaming experience when I played it. I didn't even know it dipped down that much when I was playing it till I graphed out the data.
 
Last edited:
So, my ASUS ROG Strix GTX 1060 OC arrived today. I took the old GTX 970 out, slotted the 1060 in, connected the four auxiliary monitors to the 1060. I then ran DDU in safe mode to remove the old drivers, which didn't recognize the 1060. Loaded latest WHQL drivers, everything's good.

Non-gaming temperature of the primary video card, an ASUS ROG Strix GTX 1080 that drives a 4K G-sync monitor, dropped about 8 degrees C according to GPU-Z. Mission accomplished. :)
 
This one looks like the sweet spot:

MSI GeForce GTX 1060 Armor 6G OC: $259

I would pick up the Strix, but the $329 price tag is ridiculous.
 
Which games specifically are you referring to? In general I did not feel stutter or lag or choppiness or slow responsiveness out of any of the video cards tested. What was playable on the 1060 was more than playable on the 980, and what was playable on the 1060 also matched up pretty well with what was also playable on the 480.

/edit I read back a bit at your previous posts, that dip on the 980 in Tomb Raider was very brief, it might have been in a spot where player control wasn't affected, a scripted sequence or such, I didn't feel it negatively impact my gaming experience when I played it. I didn't even know it dipped down that much when I was playing it till I graphed out the data.


Well, I saw this trend of minimum frame rates falling in the opposite order of the average frame rates and associated blurb in many of the tests, but most notably in this one:


image.png


Where the RX 480 has almost three times the minimum frame rate of the 980. I would have thought that ~10fps minimum frame rate on the 980 would have been very noticeably bad, compared to the 28fps minimum on the RX480, yet the blurb beneath it crowns the 980 as the winner.

Looking at the chart though, it looks like something happens in the last quarter of the test causing the 980 to fall off the deep end. Maybe it was just an anomaly in the testing?
 
This one looks like the sweet spot:

MSI GeForce GTX 1060 Armor 6G OC: $259

I would pick up the Strix, but the $329 price tag is ridiculous.

Opinions differ. I'd go with the EVGA SC for $259 ($249 after MIR). Decent factory OC for a reference-based design, single-fan cooler for smaller cases (and still performs admirably), and when you factor in the MIR it adheres to the MSRP. Probably the best value day-1 card.
 
Looking at the chart though, it looks like something happens in the last quarter of the test causing the 980 to fall off the deep end. Maybe it was just an anomaly in the testing?
You do realize that the framerate is still over 30 on that graphing, right? It is not uncommon to see these tiny downspikes that never show to be an issue while gaming. Now if you have a lot of those, the experience can be very different. And this is PrensentMon data too, and it takes multiple framerates per second.
 
Thanks for the review. The founders looks to be a lot slower than the EVGA SC model.

Bottom line. GTX 1060 is better than the RX 480. Like we all knew it would be. It performs better & does it much more efficiently and quietly.

The only reason can think of getting any RX 480 would be for higher resolutions than 1440p to take advantage of the extrea 2GB of ram. Looking forward to AMD's AIB versions of the 480. Hopefully they can cool it off.
 
Thanks for the review. The founders looks to be a lot slower than the EVGA SC model.

Bottom line. GTX 1060 is better than the RX 480. Like we all knew it would be. It performs better & does it much more efficiently and quietly.

The only reason can think of getting any RX 480 would be for higher resolutions than 1440p to take advantage of the extrea 2GB of ram. Looking forward to AMD's AIB versions of the 480. Hopefully they can cool it off.

Did we read the same review?

I came away with mostly tied but slight advantage 1060 at 1440p, but the RX 480 runs away with it at 1080p...
 
Thanks for the review. The founders looks to be a lot slower than the EVGA SC model.

Bottom line. GTX 1060 is better than the RX 480. Like we all knew it would be. It performs better & does it much more efficiently and quietly.

The only reason can think of getting any RX 480 would be for higher resolutions than 1440p to take advantage of the extrea 2GB of ram. Looking forward to AMD's AIB versions of the 480. Hopefully they can cool it off.
sorry but did you actually read the review? Because that wasn't the conclusion at all.
 
Here are some OC vs OC comparisons for the GTX 1060 vs RX480.
gtx-1060-oc-mordor-1080.png
gtx-1060-oc-mec-1080.png
gtx-1060-oc-gta-v-1080.png

Note how the FE and the MSI gaming both hit ~2090mhz ? Pretty much any 1060 will be hitting those clocks.

Look at AotS:

gtx-1060-bench-ashes-4kfps.png


The 5.9 TFlop/s RX480 is ON PAR with a reference 1060; ( 4.5tflop @ 1750mhz). Like I've been saying, the RX480 underperforms here, where GCN has traditionally shined because of it's compute throughput. DOOM with shader intrinsics is now the go to benchmark when people want to argue that there is something inherent in DX12/Vulkan that favors AMD. There's no such thing.
 
What really pisses me off about the RX480 launch is the way people are back to ignoring the reality of today and making wishful claims about the future. The first thing people need to stop repeating is "AMD has better DX12/Vulkan performance", unless you follow it up with "compared to their DX11/OGL performance". There's nothing about DX12 or VK that guarantees AMD will have an advantage.

Having said that, AMD promised much improved DX11 performance, and much improved geometry performance. I don't see how they delivered frankly. DX11 performance still suffers, despite their having claimed the command processor was the issue and their updating it with a new revised version on Polaris. The geometry performance of this card is still subpar, as you can see from Rise of the Tomb Raider for example. Which is just as valid a DX12 benchmark as any other really, GCN sees large gains from async because rasterizer bottleneck was severe.

Fact that RX480 barely performs on par with 390X in tessellation heavy games, and DX11 games where GCN was under-performing means they didn't improve shit

Hey Kyle have you heard any news on when the custom RX 480 are coming out ?? I hope it in next few weeks as I ready for upgrade.

OCUK mentioned Sapphire Nitro being ready for early August
 
Well the fact that both camps are claiming victory and nobody can really conclusively prove one way or another is good for this price point (~$250, I'm not counting FE's), shame we can't have some competition at the high end. Remember the good ole days where the best card was the one that was the most recently released by whoever...
 
Well the fact that both camps are claiming victory and nobody can really conclusively prove one way or another is good for this price point (~$250, I'm not counting FE's), shame we can't have some competition at the high end. Remember the good ole days where the best card was the one that was the most recently released by whoever...

If I had to choose one card, today (assuming all currently released cards were in stock), it would be a 1060. Reference cards usually aren't my thing and AMD's partners have not yet delivered on that front, regardless of the reasoning. My opinion likely will change once those AIB cards are out.
 
Bone stock you guys, from what I've gathered the REFERENCE FE 1060s boost to about ~ 1840mhz with power throttling in extended sessions. So if we assume the average is actually 1800mhz; there's 15% OC headroom on this bitch
http://i.imgur.com/JmrJ5cD.jpg
JmrJ5cD.jpg


Even if we assume the Vulkan version of DOOM will never get improvements (and it will very likely, same can't be same for AMD whose shader intrinsics are already being used) the reference 1060 is performing at 85% of an RX480 in Hitman and DOOM VK.

15% OC @ 0.85x =~ 0.97x. An overclocked 1060 will perform roughly 3% slower than a (stock) RX480 in these two titles. In AotS it's just going to be faster. I love how AotS has now been abandoned as the definitive LL API benchmark lol
 
Ieldra you said it would be decently faster and you were wrong, not even much in it in dx11 and the oc head room is negligible or largely equivalent to the 480.
Why should people listen to your latest future predictions, when the data available says you are likely greenwashing again?
Nvidia wasn't downgraded a few days earlier for no reason. I wonder if it's to do with the 1060 and possibly gp100 if interposer issues are true. They are still not available outside of their own supercomputer box. Its been nearly a year now?

Kyle or Brent, did you look for or notice the possible texture streaming kerfuffle reported recently with doom during testing?
 
Ieldra you said it would be decently faster and you were wrong, not even much in it in dx11 and the oc head room is negligible or largely equivalent to the 480.

Not really rushing to defend, but if he said the 1060 would be decently faster than the 480, then he was correct, at least as far as DX11 performance is concerned. The 1060 IS faster than the 480 in DX11 titles. The 480 leads in DX12 titles that heavily utilize async.

The problem for the prospective 480 buyer is that "DX12 titles that heavily utilize async" is not really a thing.
 
Sure I did. And I've read a lot of others too. Check out Jayztwocents Youtube review on the EVGA SC model.




Jays 2 cents videos are great for learning about water cooling, but I wouldn't trust his GPU reviews particularly much. He just doesn't come across as very knowledgeable there.

That and, why would you ever watch a video review, when a written article is so much better? Must be a millennial. :rolleyes:
 
Back
Top