One thing annoys me about GTX1060 and that is the lack of SLI support. I was thinking about selling my 2x EVGA GTX780Ti SC cards for two GTX1060 simply coz of twise memory amount, that is the only thing thats not enogh with 2xGTX780ti SC on higher res, but Sli isn`t supported So i might go with 2x RX480 instead, but not unless there are 3d party cards with better coolers, if not i might hold out untill 1080Ti and AMDs cards are released.
Well thats a thing from supply and demand. Doesnt have anything to do with introduction price. You just have to be quick with ordering
I've never had issues with the card working to be honest. On either brand. How much differentcan the drivers be for 380 vs 290x. To me if 480 drivers are as good as my experience has been that difference is big.
@Ryu5uzaku/sunnyp_343 depends. if someone is looking just at the price, sure, but once i care about things like power/efficiency/heat/noise and things like shadowplay, i dont have a problem to spend 50-100 more on a card that's silent while running like a 980. plus, for me at least, i never got an ati card to run as expected in the past (my rig, or friends; 2010 was last time) might be better now, as drivers have changed and things like DDU helps getting rid of nv "leftovers" (which was most likely part of the problem), but im just reluctant to "try" again, since the last 5 times it didnt work out.
Not just that. If we take a moment and imagine that we all live in the dystopia that is the USA. A horrific place where a RX480 can be found for their advertised prices of $199.99 and $239.99, and a 1080FE for $699 ( Not including taxes of course). No partner has yet, to my knowledge, released a card at that $599 point. I expect the 1060 will have the same treatment.
Well said. I expect 1060 to hold that 300$ price point even when non-FE cards arrive. As for 3GB ver i dont even wanna talk about about becouse no one should consider it.
If screen goes to power saving mode, it might not come back on (backlight comes on, but screen stays black) without turning it off and back on, some games are not playable as those crash to desktop too often. Those are the main problems.
It can't really compete with the AMD 480 since it doesn't have SLI so I don't see Nvidia selling many 1060s. You might as well go for 2x gtx970's
Pascal have some improvements in DX12 space over Maxwell (one of that improvements should be Async Shaders in hardware) . I assume if in DX11 will be in the same class as 980 (maybe a little slower) , in DX12 can be a bit faster than 980
Pascal doesn't have Async Shaders or any type of hardware based asynchronous support. All it has is faster pre-emption which they use to do dynamic load balancing between graphics/compute. There is a slight gain in DX12 with Pascal but everything I've seen is like 2%-3%. The reason why AMD gets such a big boost in DX12 is because portions of their hardware essentially do nothing in DX11 and it alleviates driver overhead issues they have.
My friend had such issues with some display actually he replaced that acer with asus and no more that issue dunno why. None here for whatever reason I guess it's down to the display also. Which games for you ctd a lot? Out of interest if I have any of em.
The 4GB AMD and the 3GB NVIDIA should be disregarded for now, as they will be meaningful once the market normalizes, which is at least 3 months from now. Maybe then they will matter, for now it's the 8GB RX480 vs the 6GB GTX 1060. A few points for both: Power consumption: NVIDIA {This should be an easy one for the 1060, although it won't be so graceful when overclocked for sure. Either because of GloFo, or the design itself, the AMD card will probably consume 30-80% more power. For my case, that won't matter at all, for others it might}. Overclocking: NVIDIA {Unless the AMD card is severely power starved, and that is the reason it won't really go above 1.4-1.5GHz, then this is an easy one for NVIDIA. We have no actual hardware yet, and I believe that sustained 2GHz clocks will be much harder to get than what most people believe, still it seems that at least 200MHz over the default 1.7GHz boost clock should be doable for almost everyone}. Pixel Throughput/Tessellation: NVIDIA {It's rumored that the 1060 will have 48 ROPs. NVIDIA has had the efficiency in ROPs for a few generations now, the extra 50% number of units, along with the higher clocks should give this one a handy advantage over the 480. It doesn't seem to be as clear cut as in the past though. The 480 seems to have pipeline improvements that keep it on par on pixel-restrained titles, with cards with the double number of ROPs. Tessellation performance seems markedly improved too on the hardware level. An NVIDIA win, but probably not as great as the power consumption one}. Compute: AMD {This is an easy one. The 480 has 1026 more shader processors than the 1060. It clocks them lower, but still a low-clocked reference design (at 1266MHz, instead of the 1328 a lot of them come) gives you 5.83Tflop(6.12Tflop with the 1.328MHz versions) of performance, vs the 4.35Tflop (4.86Tflop at 1.9GHz OC) of the 1060. AMD's pipeline also seems to be more capable in general with much more sophisticated hardware-level scheduling than the NVIDIA one. This time they have decreased the number of ACEs to something more reasonable (four), and they include two new hardware scheduling units that I HAVE NO IDEA WHAT THEY DO (pls, someone an architecture review). The difference is not as great as the pixel one but it seems that these days cards tend to get bottlenecked more on compute rather than pixel throughput, unless the resolutions go higher than this segment is serving anyways}. Memory: AMD {This is an easy one. The 960 and the 380 proved that although tricks to compress memory transfers do work nicely, they are very much a per-case thing that can give no guaranteed performance. The 128bit 960 chocked vs the 256bit 380, and the 380 was, in a lot of cases, even losing to the 280x and its 384bit bus. NVIDIA recognized this by equipping the 1080 with GDDR5x. People seem to believe that this time NVIDIA saw their "mistake" since the card doesn't have a 128bit bus, and hence it doesn't follow the 960 paradigm, which was literally a card with half the hardware of the 980. It's not exactly like that though. The 1080 has much faster memory this time (2Gbps faster), taking us back again to the 960/980 paradigm. The 8GB vs 6GB difference won't matter initially, but my feeling is that it will eventually. 8GB seems to be the industry standard now days for both consoles and new GPUs. And no, as any developer can tell you, what the GPU memory contains doesn't have to scale with resolution. The bus itself will also matter, since it will have lower than console raw memory bandwidth, and the consoles themselves will be using whatever bandwidth tricks Polaris is using}. Games: DX11/Older Games: NVIDIA DX12/Newer Engines: AMD {I'm not even sure that this needs more explanation. Even by design, the 1060 is a much more serialized one than the 480. It has fewer "lanes", but things on them run faster. Most newer game engines are tweaked towards the AMD paradigm (because of the hardware they have to optimize for, ie the GCN-based consoles). I expect the 1060 to be much faster than the 480 in most DX11/older stuff, and I also expect it to either be not as fast, or to be outright slower with newer engine iterations, even under DX11. Under DX12 and the general propensity of newer APIs to use more memory and do more stuff in parallel (in completely layman's terms), I expect the 480 to be much faster than the 1060, and a generally better long-term investment}. Drivers: DX11: NVIDIA DX12: AMD Short term: NVIDIA Long term: AMD {Also quite clear cut. The AMD DX12 ecosystem seems to be much better than the NVIDIA one at this moment. This is probably a combination of the hardware architecture and AMD's own investment at lower level APIs. Unless NVIDIA's architecture changes dramatically, I can't see this changing either. Furthermore, the majority of big publishers seems to be on the AMD bandwagon now. EA, Ubisoft, Sega, Square Enix, are all AMD partners and have announced that all their new AAA titles will be using DX12. On the other hand, with NVIDIA's DX11 driver you will get 100% out of your hardware close to launch day, which is, frankly, amazing. AMD has improved their DX11 driver tremendously the last 18 months and they keep adding features, but essential things like enforced double vsync, vsync and triple buffering are still missing, which is inexcusable. Due to AMD's commitment to a single architecture, expect more long term gains than the 1060. The 1060 seems to be a much more streamlined design overall, getting 100% of its hardware capability out almost immediately. GCN seems more like the closet you find after your grandpa is dead. You won't believe how many things it has inside, but it usually takes AMD more time to get them all out}. I wrote all this as a guide to myself also, since one of these two is probably what will replace my thrice-baked 7970. I really wish NVIDIA went with 1500+ shaders and a 256bit bus. Even at $279. Then there would be no question. Now, I'm not so sure. The new consoles having Polaris/Vega iterations, doesn't make it sound like a sound investment for more than a year. Close to the run that the 960 had, more or less. AMD and NVIDIA will price these cards according to their launch day performance, meaning that the AMD card will probably have a much higher performance/dollar if you plan on keeping it for more than a year.
Dynamic load balancing and improved pre-emption both improve the performance of async compute code considerably on Pascal compared to Maxwell. Although principally this is not exactly the same as Asynchronous Shading or Computing. Because Pascal still can’t execute async code concurrently without pre-emption. This is quite different from AMD’s GCN architecture which has Asynchronous Compute engines that enable the execution of multiple kernels concurrently without pre-emption. What Nvidia is doing with preemption and dynamic load balancing right now, while not exactly async compute, can be used to accomplish similar goals. So Nvidia can execute async codes but WITH preemption , AMD can do without .Anyway preemption is active by default on Pascal and more of that ... is on pixel level . In the end depends of developer or the game if takes the advantage of this features . If a developer will make the game engine to be aware of preemption then , async code will run faster on Pascal than GCN . Read more: www.*************/nvidia-geforce-gtx-1080-dx12-benchmarks/
259$ pricing look nice, cant wait to see reviews, then for MSI/EVGA to show there cards, Really want to know how it compared to a 660gtx, I Got my evga 660gtx ftw for 259$ and if really is same price then might be great pick up
That black screen problem made me go nuts after a while when I had R9 290. Also memory didn't go power saving mode when I ran my display at 75 Hz which caused high idle temps and even watching youtube caused the fan to speed up. Those were actually the reasons why I sold it.
$249 for the 6gb card is really decent. Nobody should be looking at saving some $30 on the 3gb version.
