I have seen you repeat this misinformation so many times, I have to respond with this. Maybe you ascribe to the notion "if you repeat a lie often enough, it is accepted".async is only a small feature in the global scale of things. Even on AMD hardware, it provides at most 10% performance in the best case "sponsorised tech demo" AoS. 5% will be the norm in the 2016 DX12 games. Not a big deal...
are you really believing what you wrote ?But remember, Polaris 10 will be 90% of GTX 1080 in DX12 for half the price, if not less.
Dunno about you guys, but if AMD's partners aren't going to show off anything at Computex, that's a red flag. But I'm sure to the AMD hopefuls, this just mean AMD is funneling all their stock to Apple or Dell and AMD is about to pull that magic 3pointer that will bring the score up 20 to 145.
How many 1080 & 1070 cards do you estimate have been sold so far and how many by the time of early-mid June?I think AMD has had their A1 silicon for a few months now. But they got wind of Pascal's clocks and realized that they arent even in the same ballpark. So they've been mad scrambling to get those clocks up. They have obviously failed to do so. Otherwise, dont you think they would be showing off something? From this day forward, every day of delay represents millions in lost revenue and surely they know this. The same thing happened a year ago with Fury. If they actually had a winning product, would they have waited so long to release it? I'm sure they had Fury silicon ready to go in April 2015 but they caught wind of 980ti and said "we're screwed" and spent the next two months doing damage control before finally giving up.
are you really believing what you wrote ?
seriously
I put in bold the important part of the quote for youI have seen you repeat this misinformation so many times, I have to respond with this. Maybe you ascribe to the notion "if you repeat a lie often enough, it is accepted".
Note everyone, this is an AoS developer saying that they had little async compute compared to what other developers are doing and getting.
Xpea claims the opposite, saying that AoS is the best case with 10% and the norm will be around 5%. Complete and utter misinformation repeated continually.
Oxide dev quote:
"Our use of Async Compute, however, pales with comparisons to some of the things which the console guys are starting to do. Most of those havent made their way to the PC yet, but Ive heard of developers getting 30% GPU performance by using Async Compute."
Sony was instrumental in adding async compute to GCN way back when it was being designed. I wonder what is in store for the next Playstation PS5 model. For certain it's being worked on now.
http://www.gamasutra.com/view/feature/191007/inside_the_playstation_4_with_mark_.php?print=1
Sony was instrumental in adding async compute to GCN way back when it was being designed. I wonder what is in store for the next Playstation PS5 model. For certain it's being worked on now.
http://www.gamasutra.com/view/feature/191007/inside_the_playstation_4_with_mark_.php?print=1
Indeed, we don't like FUD.sources for the Pascal yield issue ? otherwise you are spreading FUD :\
Pot, meet kettle.I've been saying for weeks that something's wrong with Polaris.
Dunno about you guys, but if AMD's partners aren't going to show off anything at Computex, that's a red flag.
Sony was instrumental in adding async compute to GCN way back when it was being designed. I wonder what is in store for the next Playstation PS5 model. For certain it's being worked on now.
Thirdly, said Cerny, "The original AMD GCN architecture allowed for one source of graphics commands, and two sources of compute commands. For PS4, we’ve worked with AMD to increase the limit to 64 sources of compute commands -- the idea is if you have some asynchronous compute you want to perform, you put commands in one of these 64 queues, and then there are multiple levels of arbitration in the hardware to determine what runs, how it runs, and when it runs, alongside the graphics that's in the system."
Just like NVidia launching a direct only reference card and skipping partners for who knows how long is a red flag? AMD made it abundantly clear they want Polaris in mobile and OEM markets. You need AIB offerings for neither of those.
I agree that something has changed. As to the cause, ?????????It's interesting how the tide of public opinion shifts.
From like Nov of 2015 to about March of 2016, it was AMD and the DX12 train wreck'n'effect hype mobile. Countless posts of how AMD was superior, how DX12 was going to change everything. Then the few DX12 titles came out, some blocked by NV "bribes" others falling on their own face unrelated to DX12, well depending on who you ask.
AMD being 6+ months ahead, no way NV can launch in before Q3 2016, and etc. Then AMD went dark.
Fast foward to May and NV finally talks about Pascal outside of some custom parts for markets unrelated to gaming. So NV was in trouble because they were silent while AMD was ringing their bells, but now NV is active and it's NV is in trouble and AMD is being smart by being quiet. Which is it?
Dunno about you guys, but if AMD's partners aren't going to show off anything at Computex, that's a red flag. But I'm sure to the AMD hopefuls, this just mean AMD is funneling all their stock to Apple or Dell and AMD is about to pull that magic 3pointer that will bring the score up 20 to 145.
But remember, Polaris 10 will be 90% of GTX 1080 in DX12 for half the price, if not less.
I stand corrected.That isn't what your article says:
IOW, GCN already existed and had async compute by the time Sony considered using it. They just improved the async compute capabilities.
Yep, and there it is again. "For how long." For "months" according to another poster.
Yet, in this same thread, partners are going to show custom cards at Computex. Them flags so red they purple!
Not a single thing you wrote here changes the fact that you are posing False information.I put in bold the important part of the quote for you
In consoles, we may see some improvements for xmas 2016 titles, and for PC, much later, maybe 2017, maybe 2018...
I talked about 2016 DX12 PC games ()
I said many times that AMD strategy is wrong when it comes to new features. They are too much ahead of the time. When the feature becomes mainstream, its too late to enjoy it on the initial hardware. Customers want benefit now. Immediately. They play games now. They buy hardware to enjoy now, not with the idea of maybe perhaps whoknows it will be better in the uncertain future...
()()()If NVidia announced the release date for non-founders cards, simply point me to the press statement and be done with it.
we can all see how this is a non founders card right there()()()
And I stand corrected because you forgot to mention "for 2016 DX12 games". Then I took one paragraph to explain why it's not important (ie AMD brings features too early with no immediate benefit)Not a single thing you wrote here changes the fact that you are posing False information.
By you repeatedly:
"Async compute provides improvements of 5% average and 10% max. It's no big thing"
You do whatever you think it's best for you but overall market share has spoken...By the way what happens next year 2017, If I buy a card now?
If NVidia announced the release date for non-founders cards, simply point me to the press statement and be done with it.
Why would they announce the release date of products that would cannibalize their sales? They aren't AMD, they're going to try to maximize their profits.
But if you don't want a Founder Edition, there are a few articles pointing at some already in the pipeline.