Gtx 1080 benchmarked

I don't understand it either, do people actually buy every new generation of card? I just wait until I can't play the games I want at the detail level I want. Let's see, "recently" I've had 8800 GTS -> GTX 460->GTX 460 SLI->Radeon 7950 Crossfire (only sold due to Bitcoin mining making the cards worth MORE on the used market than what I bought them for)->Radeon 280x Crossfire (sold again).

I'm going to be buying either one of these new generation cards or a used 970/980/290x/390x either way, but I'm hopeing for a half decent GTX 1080 or 1070 to buy, if there isn't a big difference I'll buy the 1070 with the idea of possibly picking up another down the road.

I keep my cards for 3 years on average
 
I really don't understand the disappointment, I don't.

If this is how you react to Pascal I can't wait to see the reaction to Polaris

The disappointment come from the relatively small increase in performance between generations. There isn't much weight to the word "upgrade" anymore. Don't get me wrong, I don't mind getting more life out of my hardware due to there just not being a big enough gap between generations, I just want some good value for those upgrades though and its just hasn't been there the last few generations.
 
Sadly I have 28 pci lanes or something like that with my 5820 so adding in a 2nd 980 ti might be tricky.

I bet 2 x 980 ti was be killer performance.
 
If these benchmarks are real then it comes down to price. If they can deliver 980 Ti performance for 970/980 prices then that's a good start. When overclocked you would have something that performs really well in 1440p like the 980 Ti does already.

I don't see moving from my 980 Ti until we have 4K 120 Hz G-Sync/Freesync displays and GPUs that can deliver similar performance as I'm now getting at 1440p.

I've said it before that people expecting double the performance of 980 Ti are way optimistic.
 
Eagerly looking forward to reading HardOCP's review of Pascal/Polaris, but as for actual purchase? Gonna wait until "Big Pascal/Vega" next year. I have a good enough video card for my 4K needs (in sig).

Side note: looks like Wizzard (of TPU) was right on the money, when he claimed late last year that GTX 1080 would probably be around 30% faster than a stock GTX 980 Ti, noting that 980 Ti owners had no real need to upgrade.
 
Eagerly looking forward to reading HardOCP's review of Pascal/Polaris, but as for actual purchase? Gonna wait until "Big Pascal/Vega" next year. I have a good enough video card for my 4K needs (in sig).

Side note: looks like Wizzard (of TPU) was right on the money, when he claimed late last year that GTX 1080 would probably be around 30% faster than a stock GTX 980 Ti, noting that 980 Ti owners had no real need to upgrade.
I know it's too much to ask but I'd love ot know when big pascal might drop. I'm expecting 1Q2017 but so much depends on AMD's releases and how well Nvidia's new cards take off
 
It's 30% faster than a stock 980 Ti. I am pleasantly surprised.

Edit: Quick math, actually about 20-25%. This better be a sub $650 card.
 
Last edited:
Hopefully this pushes the price down some for a 980 or 980ti

Unfortunately, that's not how it works. The retailers and distributors know months in advance when the next product release is going to be. They buy accordingly. At some point before the release of the new products they'll quit ordering new stock of the old hardware and place orders for new products once pricing is determined. Any stock that didn't sell maybe discounted to get rid of it, but these companies (distributors and retailers) won't take a huge loss on them. They are in the business to make money and they can move 90% of the units without deep discounts. Either by people who are unaware of impending releases of newer technology, or people that want them for whatever other reason there might be.

Only on the used market, both the 980 and 980TI are already discontinued and they're just selling off stock now. Most will likely be gone by the time the new cards hit.

Exactly. There may be some small price cuts to give people incentives to buy existing stock, but once it's gone those cards will be replaced by new cards.

I've been really disappointed in the incremental steps in performance with GPUs lately. Going from a 7900gtx to an 8800gtx is still fresh in my mind, so seeing how little GPUs improve with each new release makes it hard to get excited about anything video card related anymore.

The disappointment come from the relatively small increase in performance between generations. There isn't much weight to the word "upgrade" anymore. Don't get me wrong, I don't mind getting more life out of my hardware due to there just not being a big enough gap between generations, I just want some good value for those upgrades though and its just hasn't been there the last few generations.

People need to understand that the massive jumps in computer technology that we enjoyed for over two decades were because the technology was in its infancy. Lots of R&D was being done and lots of people had creative ideas for moving forward as the industry grew. All industries eventually reach a point of product development where they hit a wall and progress slows or even stagnates. The frantic pace of virtually doubling performance every generation or so can't be maintained indefinitely. We've also reached a point where we are nearing the upper limits of what current materials and manufacturing technologies can do. It will take a significant breakthrough that could take years or even decades to achieve for things to improve at the pace we are accustomed to.

Decades ago people asked about the future predicted flying cars. While our current vehicles are no doubt way more advanced, they aren't that advanced. We still have the same general engine / drive train layouts, fuel source, and operating principals in place that we've had for the last 100 years. Technological advancement can be stalled for a lot of reasons, not the least of which is cost, economies of scale, or dependencies on other technologies. I think its important to realistically temper ones expectations. Necessity is the mother of invention and frankly, we don't have the software needs to justify the rapid progress we enjoyed in the 1990's.
 
Last edited:
It's 30% faster than a stock 980 Ti. I am pleasantly surprised.

Edit: Quick math, actually about 20-25%. This better be a sub $650 card.

I think with ready-to-go drivers, the 1080 has a decent chance of hitting the 25-30% mark when released.

Given that the GTX 780 came out for $649 (I was an early adopter), sadly I don't think that's going to be the case (admittedly the 980 was $549 at release 1.5 years ago, but I think NVIDIA will be using the "new features" on the 1080 to up the price).
 
Just remember this is only a Midcard of the new generation.

Just like the 680 GTX, and 980 GTX.

I expected a lil more performance to be honest with you, but a boost of 1800mhz? LOL Awesome!

I will wait for real benchmarks. And why the fuck is NVidia doing the SAME STUPID SHIT AMD DOES. Really a press release about a product that wont be released yet. I better see people from [H] and users bitch about it since everyone bitched (including me) about the Nano/Fury Duo Pro launch.
 
Honestly both companies have done that numerous times but it seems AMD gets the flak because people have some higher standard for them. I enjoy a press release. Yes you should take it with a grain of salt, but its so much better than a million fud articles about rumored specs.
 
Honestly both companies have done that numerous times but it seems AMD gets the flak because people have some higher standard for them. I enjoy a press release. Yes you should take it with a grain of salt, but its so much better than a million fud articles about rumored specs.

LOL AMD held to a higher standard. ...Both fucking companies should be help to the same standard.
 
LOL AMD held to a higher standard. ...Both fucking companies should be help to the same standard.

For paper launches I have seen both companies engage in it but I see more vitriol AMDs way when it happens. I saw less outrage over the wooden screws than I saw over some AMD press releases.
 
  • Like
Reactions: Creig
like this
It's 30% faster than a stock 980 Ti. I am pleasantly surprised.

Edit: Quick math, actually about 20-25%. This better be a sub $650 card.

For a midrange card its about as expected. I am more interested in the overclocking potential!
 
I don't understand it either, do people actually buy every new generation of card? I just wait until I can't play the games I want at the detail level I want. Let's see, "recently" I've had 8800 GTS -> GTX 460->GTX 460 SLI->Radeon 7950 Crossfire (only sold due to Bitcoin mining making the cards worth MORE on the used market than what I bought them for)->Radeon 280x Crossfire (sold again).

I'm going to be buying either one of these new generation cards or a used 970/980/290x/390x either way, but I'm hopeing for a half decent GTX 1080 or 1070 to buy, and if there isn't a big difference I'll buy the 1070 with the idea of possibly picking up another down the road.

I usually buy at least one card each generation. If you sell cards before they get "old" you can recoup most of the cost. Then I can buy a shiny new toy...
 
For paper launches I have seen both companies engage in it but I see more vitriol AMDs way when it happens. I saw less outrage over the wooden screws than I saw over some AMD press releases.

Yea 1 thing Nvidia has over AMD is a WAY better marketing department. I mean they spun that 3.5/0.5g 970 GTX well and basically ignored it and delete posted on the geforce forums.

Either way cant wait to see real benchmarks
 
25-30% faster for less than $600 and I'm on board.
Anything else I'll wait for the big chips next year.
 
So when can I order (pre-Order?) one?

Why you would pre-order an unreleased card with no benchmarks or anything? What happens if you pre-order a $650 card, only to find out AMD's $350 card is just as fast and $300 cheaper? (speculation of course I have no proof and no idea what to expect)

I would hold out until both camps have released the cards. This is all new tech on a new node, the best thing to do is to wait it out.

P.S. Do not forget we need to see DX12 benchmarks too!
 
I really don't understand the disappointment, I don't.

If this is how you react to Pascal I can't wait to see the reaction to Polaris

Many more Nvidia fans here than AMD fans. I think this will be the bigger salty tear fest if it holds true.


Did people even read the article properly, or just skip over in anger?

In this scenario GTX 1080 is faster than typical overclocked GTX 980 Ti (~8700 points).

However if we compare it to a GTX 980 Ti running at almost the same frequency (1.8GHz with LN2 cooling), GTX 1080 is actually much slower clock to clock

TLDR: IPC/clock efficiency is not as good. Very weird. Bus differences? Or underlying architecture being more orientated for different tasks from the get go? Tomorrow US time will be interesting.
 
Many more Nvidia fans here than AMD fans. I think this will be the bigger salty tear fest if it holds true.


Did people even read the article properly, or just skip over in anger?

In this scenario GTX 1080 is faster than typical overclocked GTX 980 Ti (~8700 points).

However if we compare it to a GTX 980 Ti running at almost the same frequency (1.8GHz with LN2 cooling), GTX 1080 is actually much slower clock to clock

TLDR: IPC/clock efficiency is not as good. Very weird. Bus differences? Or underlying architecture being more orientated for different tasks from the get go? Tomorrow US time will be interesting.

Well doesn't the 1080 have fewer cuda cores? Also it is only a midrange chip. So Its expected performance of 25-30% seems about right.

I would want to compare a 980ti with a pascal card with the same cuda cards at the same frequency.
 
980ti has a wider memory bus and very likely more cores, that would do it.
 
This will be a very nice upgrade going from a 760 to a 1070. holding out for the HCP review first tho.
 
Bah, not enough of a performance jump. We need more competition in video card space, sadly. :arghh:
 
Honestly, this takes a lot of pressure off me to upgrade. So this is good news.

Back in 2006 I built a $3,600 dollar beast, huge $300 Lian-Li case, all the fans, duel AMD cards, Intel Extreme CPU that cost me around $1000 etc etc. The following year, not even kidding, I built a new system using all the new stuff and let the system from 2006 just sit. Total waste of money. I've done that many times over, accept until recently.

I get the upgrade itch all the time but I keep it in check now days.
 
Bah, not enough of a performance jump. We need more competition in video card space, sadly. :arghh:

Uhh, it's more like the massive improvements we used to get from die shrinks are not going to happen anymore.

Also, competition breeds nothing different. The HD 7970 was the first 28nm card to launch by several months, and it clocked-in at $550, just like the GTX 980. It was 20% faster than the GTX 580 it was competing with., and just 35% faster than the 6970 it was replacing. It was another year and a half before both companies released top-end silicon in the mainstream, due to issues with the early process.

Early cards on new process nodes are just going to be smaller from here on out. Neither company wants a repeat of what happened with the 40nm process node: HD 5870 (poor availability for months after release) and the GTX 480 (cut, high power, late release).

I expect the same thing here, except further complicated by HBM2 shortages. You can't make the 4k monster when all you have to feed it is GDDR5.
 
Uhh, it's more like the massive improvements we used to get from die shrinks are not going to happen anymore.

Also, competition breeds nothing different. The HD 7970 was the first 28nm card to launch by several months, and it clocked-in at $550, just like the GTX 980. It was 20% faster than the GTX 580 it was competing with., and just 35% faster than the 6970 it was replacing. It was another year and a half before both companies released top-end silicon in the mainstream, due to issues with the early process.

Early cards on new process nodes are just going to be smaller from here on out. Neither company wants a repeat of what happened with the 40nm process node: HD 5870 (poor availability for months after release) and the GTX 480 (cut, high power, late release).

I expect the same thing here, except further complicated by HBM2 shortages. You can't make the 4k monster when all you have to feed it is GDDR5.
It seems silly they can't make a monster w/o HBM2, both GDDR5X and HBM are faster than GDDR5. Looking forward to the 'great unveiling'
 
It seems silly they can't make a monster w/o HBM2, both GDDR5X and HBM are faster than GDDR5. Looking forward to the 'great unveiling'

I didn't just highlight the lack of HBM2, there's the little problem of early large dies being tough to bring to market in quantity. But you ddidn't bother to read past HBM2, did you?

You can bitch and whine all you want about this shit, but it doesn't change the fact that making large semiconductors on cutting-edge process nodes is hard. If you complain, you're only wasting your breath.
 
Back
Top