The NVIDIA GeForce GTX 1080 Preview: A Look at What's to Come
by Ryan Smith on May 17, 2016 9:00 AM ESTFirst Thoughts
Wrapping up our preview of the GeForce GTX 1080, I think it’s safe to say that NVIDIA intends to start off the 16nm/14nm generation with a bang. As the first high-end card of this generation the GTX 1080 sets new marks for overall performance and for power efficiency, thanks to the combination of TSMC’s 16nm FinFET process and NVIDIA’s Pascal architecture. Translating this into numbers, at 4K we’re looking at 30% performance gain versus the GTX 980 Ti and a 70% performance gain over the GTX 980, amounting to a very significant jump in efficiency and performance over the Maxwell generation.
Looking at the bigger picture, as the first vendor to launch their 16nm/14nm flagship card, NVIDIA will get to enjoy the first mover’s advantage both with respect to setting performance expectations and with pricing. The GeForce GTX 1080 will keep the performance crown solidly in NVIDIA’s hands, and with it control of the high-end video card market for some time to come. NVIDIA’s loyal opposition, AMD’s Radeon Technologies Group, has strongly hinted that they’re not going to be releasing comparable high-performance video cards in the near future. Rather the company is looking to make a run at the much larger mainstream market for desktops and laptops with their Polaris architecture, something that GP104 isn’t meant to address.
The lack of competition at the high-end means that for the time being NVIDIA can price the GTX 1080 at what the market will bear, and this is more or less what we’re looking at for NVIDIA’s new card. While the formal MSRP on the GTX 1080 is $599 – $50 over what the GTX 980 launched at – that price is the starting price for custom cards from NVIDIA’s partners. The reference card as we’ve previewed it today – what NVIDIA is calling the Founders Edition card – carries a $100 premium over that, pushing it to $699.
| GeForce GTX 1080 Configurations | ||||
| Base | Founders Edition | |||
| Core Clock | 1607MHz | 1607MHz | ||
| Boost Clock | 1733MHz | 1733MHz | ||
| Memory Clock | 10Gbps GDDR5X | 10Gbps GDDR5X | ||
| Cooler | Manufacturer Custom (Typical: 2 or 3 Fan Open Air) |
NVIDIA Reference (Blower w/Vapor Chamber) |
||
| Availability Date | June 2016? | 05/27/2016 | ||
| Price | Starting at $599 | $699 | ||
While the differences between the reference and custom cards will be a longer subject for our full review, the more immediate ramification is going to be that only the Founders Edition cards are guaranteed to be available at launch. NVIDIA can’t speak definitively for their board partners, but at this point I am not seriously expecting custom cards until June. And this means that if you want one of the first GTX 1080s, then you’re going to have to pay $699 for the Founders Edition card. Which is not to say that it’s a bad card – far from it, it’s probably NVIDIA’s finest reference card to date – however it pushes the card’s price north of 980 Ti territory, some $150 higher than where the GTX 980 launched in 2014. For those who can afford such a card they will not be disappointed, but it’s definitely less affordable than past NVIDIA x80 cards.
Anyhow, we’ll be back later this week with our full review of the GeForce GTX 1080, so be sure to stay tuned.
| Spring 2016 GPU Pricing Comparison | |||||
| AMD | Price | NVIDIA | |||
| $699 | GeForce GTX 1080 FE | ||||
| Radeon R9 Fury X | $609 | ||||
| $589 | GeForce GTX 980 Ti | ||||
| $429 | GeForce GTX 980 | ||||
| Radeon R9 390X | $399 | ||||
| Radeon R9 390 | $289 | GeForce GTX 970 | |||
Facebook
Twitter
RSS
262 Comments
View All Comments
Ryan Smith - Tuesday, May 17, 2016 - link
I can already tell you right now that at an architectural level, per-clock per-core performance between GP104 and GM204 is virtually identical. Throughput of the ALUs, texture units, ROPs, etc has not changed. What makes GP104 faster than GM204 is the larger number of SMs, the higher clockspeeds, and a memory subsystem fast enough to feed the former.(Which is not to discount the Pascal architecture. 16nm FinFET alone won't let you ramp up the clockspeeds like this. NVIDIA had to specifically engineer the architecture to hit those clockspeeds without driving up the power consumption)
modeless - Tuesday, May 17, 2016 - link
There's only one thing I want to know about this card. Does it support the special instructions that the Tesla P100 has for half precision float (FP16), which double throughput? This is very important for deep learning and nobody has confirmed yet.vladx - Tuesday, May 17, 2016 - link
It most surely doesn't support those like P100 as that is the whole point of selling Tesla at such high price.dragonsqrrl - Tuesday, May 17, 2016 - link
So has that been confirmed by Nvidia or a review, or is that your assumption?vladx - Tuesday, May 17, 2016 - link
I'm 100% sure since GTX 1080 which is based on GP104 has a different architecture than P100 and almost identical to Maxwell.Yojimbo - Wednesday, May 18, 2016 - link
I'm 100% sure you're wrong, because the GP106, or something like it, will be used in the Drive PX 2 and will have double throughput half precision support since it's going to be used as a machine learning inference engine. If the GTX 1080 doesn't support double throughput half-precision support it's probably because they purposefully disabled it to prevent the cards from being used in high quantities for compute workloads. They will probably, at some point, come out with a Tesla based on the GP104 and/or the GP106 that does support double throughput half-precision compute, to replace the M40 and M4 cards. Pascal does everything better than Maxwell so it would be starving a growing industry to leave that segment to Maxwell for too long.vladx - Wednesday, May 18, 2016 - link
Huh? We know next to nothing about GP106 architecture unlike with GP100 and GP104 chips which like I said have different architectures with GP104 (GTX 1080) being almost identical to Maxwell on a hardware level.vladx - Wednesday, May 18, 2016 - link
Anyways, see Ryan Smith's answer below. I was 100% correct.dragonsqrrl - Wednesday, May 18, 2016 - link
Not necessarily. In fact I would be very surprised if GP104 doesn't support the double FP16 throughput. Like yojimbo said, the more likely scenario is that half precision performance is capped in some way on GeForce cards (likely through firmware).vladx - Thursday, May 19, 2016 - link
When we're talking about an architecture, we're speaking about an instrustion set. Since GP104 lacks certain compute instructions compared to GP100 in P100 then we can accurately say they have different architectures. Yes they are both Pascal and they have the same featurees on software-level but hardware-wise they are different. Doesn't matter how Nvidia enforces those differences, what matters is that they're different at an architecture-level(instruction set).