What speeds we looking at the new 1180 50% faster then the 1080 i have no idea ?
30% max.What speeds we looking at the new 1180 50% faster then the 1080 i have no idea ?
30% max.
There is no actual die-shrink but only an architectural revision like happened with Maxwell... most time die-shrink helps with big jumps in performance.
30% max.
There is no actual die-shrink but only an architectural revision like happened with Maxwell... most time die-shrink helps with big jumps in performance.
30% max.
There is no actual die-shrink but only an architectural revision like happened with Maxwell... most time die-shrink helps with big jumps in performance.
30% max.
There is no actual die-shrink but only an architectural revision like happened with Maxwell... most time die-shrink helps with big jumps in performance.
12nm is not really a "new" die-shrink... it is a enhanced 16nm... it is about the same size.I thought Volta was shrinking down to 12nm.
No... 30% is not good? I think it is a good jump for 1080 to 1180.So you're saying wait for Einstein?![]()
Around that but I was talking more about high-end... in you graph the 980 to 780 is about 30% increase in performance.So 970 was 30% max to 770?
You're going to be waiting for a long time to upgrade then!Still rocking a 970 here and prolly gonna skip this series too. Not planning to upgrade my monitor, until GPUs can handle 4K on 144Hz. 😅
Boss★Moogle;234252979 said:There's no Volta this year, that's for sure. There's no need for it since AMD still isn't being competitive. the RX 500 cards are just rebrands of the weak RX 400 and I doubt RX Vega will beat the regular 1080 and even if it does it won't be by much and nVidia can just drop the price some more.
The 1170 should perform similarly to the 1080 Ti for a much lower price so there's no need for them to release it now and kill their 1080 Ti sales, that card only just came out.
Well there is no way nVidia won't launch something this yearSo wait, are you guys still really thinking that this is expect in May, or is this an ongoing joke?
There's no way an 1180 is coming in May.
Around that but I was talking more about high-end... in you graph the 980 to 780 is about 30% increase in performance.
If it was 50% there would be so much outrage on people who bought one last year, or this past fall. Cards are coming out to quickly if you ask me.
It's one thing if it's a refresh with some nicer clocks and such. But if the performance difference is 30%+ then you have a problem with not really considering what people pay up front for your over priced cards.
The fact that 1070's can be had for 350-379 now is crazy since when they launched they were 459.99.
I really miss the days of gpu's having much larger cycles and having more legs than just a year being the top card for their range.
So wait, are you guys still really thinking that this is expect in May, or is this an ongoing joke?
There's no way an 1180 is coming in May.
Silly mindset to have. You buy the card with the performance you need right now, what will be available a year from now is irrelevant.
Very few people need to upgrade to the latest GPU at release and the ones who do just sell their old card and pay a much smaller amount of money for that new card. People who always buy the latest would fucking love if they released new cards yearly that improved performance by 50% if prices stayed the same. A 50% improvement for around 250-300$ is definitely not a bad deal when money isn't much of a problem.
Holding back performance to please consumers is dumb.
It's not silly if your from a time where video cards had longer cycles. And it use to be like that where the card you bought was the card for at least 2 years until something else came out or was not made irrelevant so quickly as it is now.
To the bold: I guess those people are what Nvidia look for in selling you a mid range card for $300, or selling you over priced reference cards. Those type of people who have that indispensable income are not a big percentage in the consumers who buy PC hardware. Go look at steam stats, most resolutions are 1080p or below. How many of those users are buying $400+ cards every year to play at 1080p?
Well there is no way nVidia won't launch something this year![]()
Volta will most certainly be shown at GTC which will happen in 20 days from now. This however will be GV100 HPC chip, not a gaming GPU and certainly not a new GeForce card - this was always the expectation.
And when was these times exactly?
There is a very interesting rumor currently making the rounds: that NVIDIA has pushed forward its upcoming Volta lineup (Geforce 20 Series Graphics) for a Q3 launch. The rumor cites a source we havent used in a long while, and is none other than MyDrivers.com: a chinese publication that was fairly active in the leak scene till a few years back.
So what does NVIDIA have in store for Volta architecture? Well fortunately, we arent completely in the dark on this one. Apart from what official sources have told us, a very well known leaker from Baidu revealed some details about its upcoming lineup that proved to be pretty enlightening. NVIDIA will be doing some basic optimizations of the SM in the GeForce 20 series, but dont expect a complete overhaul like the one from Kepler to Maxwell. The primary Volta lineup will be divided into four main GPUs:
* Volta GV100 GPU (Enterprise/HPC Only)
* Volta GV104 GPU (GTX 2070, GTX 2080)
* Volta GV102 GPU (GTX 2080 Ti / GTX TITAN XV?)
* Volta GV110 GPU
in order of power, we have the mainstream GV104 Volta GPU, which will probably be found in the GTX 2070 and GTX 2080 to begin with. Then, we have the GV102 GPU, which will probably be part of the Quadro series GPUs as well as the server/HPC market. This should eventually trickle down into the desktop segment in the form of the GTX 2080 Ti and the new TITAN card
Finally we have the GV110 GPU. Here is the thing however, the nomenclature GV110, translates to a refined iteration of the GV100 GPU since the second numeric stands for the revision no. Since the GV100 will probably be for the server/HPC market only, I can only assume that this GPU will be the last to come to desktop, and will probably be the full blown die whereas the GV102, will be the yield-adjusted die. According to the leaker, the Volta based GeForce 20 series, should finally step into the world of HBM2, as well as offer improved memory speeds on the GDDR5X platform.
Same, but just because I don't have money to burnStill rocking a 970 here and prolly gonna skip this series too. Not planning to upgrade my monitor, until GPUs can handle 4K on 144Hz. 😅
That's pretty crazy.i'm still using a gtx 560 ti lol with a AMD phenom II 2 cores. Then again all i play on pc is Overwatch
Constant 23-40 fps and most of the time come out on top. Suck it all you 144fps players haha![]()
30% max.
There is no actual die-shrink but only an architectural revision like happened with Maxwell... most time die-shrink helps with big jumps in performance.
Thought Nvidia HBM2 cards were 2018? Didn't AMD have some kind of deal for HBM2 first?
Thought thats why Nvidia is going GDDR5x?
A few days ago Nvidia released the first HBM2 GPU. The Tesla P100 for workstations.
The first consumer GPU with HBM2 will be probably the AMD Vega.
About the next Nvidia GPU and the memory type I have see rumors with GDDR5X, HBM2 and even GDDR6 for late 2017 or 2018.
The upgrade cycle used to be every 12 months, not 18.It's not silly if your from a time where video cards had longer cycles. And it use to be like that where the card you bought was the card for at least 2 years until something else came out or was not made irrelevant so quickly as it is now.
Typically NVIDIA moves performance down one tier every cycle, and boosts performance by about 30% at the high end with the faster cards.
They aren't screwing anyone over if they release new GPUs in Q3; it will be business as usual.
The upgrade cycle used to be every 12 months, not 18.
It's slower now than it has ever been.
Like how 1080p was trivial on the PS4?
There's dimishing returns, I don't think we are going go get native 4K as standard next gen.
I guess you weren't around in the late 90s and early 00s, when every year brought not just performance but entirely new graphical features to the table. Those were the times.I really miss the days of gpu's having much larger cycles and having more legs than just a year being the top card for their range.
P100 has been out for a relatively long time. A friend has been running stuff on P100s since fall 2016 (on Piz Daint).A few days ago Nvidia released the first HBM2 GPU. The Tesla P100 for workstations.
May be selling semantics here, but 1080p on PS4 IS technically trivial, as 99% of games are running at 1080p this gen, only those that are trying to push 60fps are downgraded, and even then that's because of the CPU, not the GPU as we're talking about here.
Even on Xbox One base, about 65 to 70% of games released are also 1080p, we're just excluding certain bigger games that are taxing on the GPU.
If devs want to hit 4K next gen, it'll be significantly easier than any time before i feel.
If it was 50% there would be so much outrage on people who bought one last year, or this past fall. Cards are coming out to quickly if you ask me.
It's one thing if it's a refresh with some nicer clocks and such. But if the performance difference is 30%+ then you have a problem with not really considering what people pay up front for your over priced cards.
The fact that 1070's can be had for 350-379 now is crazy since when they launched they were 459.99.
I really miss the days of gpu's having much larger cycles and having more legs than just a year being the top card for their range.
P100 has been out for a relatively long time. A friend has been running stuff on P100s since fall 2016 (on Piz Daint).
60fps/full 4K will take multiple generations until it's in the upper mainstream segment.
Yeah, going from GeForce 256 all the way to the 8 series was pretty rapid.
May be selling semantics here, but 1080p on PS4 IS technically trivial, as 99% of games are running at 1080p this gen, only those that are trying to push 60fps are downgraded, and even then that's because of the CPU, not the GPU as we're talking about here.
Even on Xbox One base, about 65 to 70% of games released are also 1080p, we're just excluding certain bigger games that are taxing on the GPU.
If devs want to hit 4K next gen, it'll be significantly easier than any time before i feel.
For me one of the bigger surprises was Watch_Dogs 2 being 1080p and looking better than the 900p WD1.
You can't​ just make an assumption about what resolution devs target while disregarding the breakdown of their actual choices in current gen.These numbers are baseless
But that's not my point. Ryse could have been 1080p on Xbox One, for example, but devs are mostly likely going to push their titles graphically in other areas causing a drop in resolution.
You can't just make an assumption about what resolution devs target while disregarding the breakdown of their actual choices in current gen.
I think we're relying too much on fuzzy guesses.What do you mean?
We have many examples of devs not going for native 1080p this gen the same way we didn't get native 720p all the time last gen.
What's debatable here is semantics, what 4k native as "standard" means. If it means 99% of games, that's not what we have for 1080p this gen on PS4. Looking at AAA games alone, it's a modest majority.
the main point is that a vast majority of those are 1080p and so the term 'struggling' to hit the res doesn't really apply