Read more.Quote:
Will arrive in the summer of 2015, according to reports.
Printable View
Read more.Quote:
Will arrive in the summer of 2015, according to reports.
Well this is timely...at the same time, Summer of 2015 is a looong way out...
If true that'll be 3.5 years since their last major silicon refresh. No wonder nVidia is pulling ahead. It better be something special and let's be honest, nVidia will likely be releasing their follow up to the GTX980 around 3 or 4 months after. It needs to blow everything out of the water.
Both AMD and nVidia have 14.1% of the market share as of latest report, Q3 2014.Quote:
Originally Posted by cptwhite_uk
Intel are still leading on 71.8% though.
Well new nvidia 970 and 980, barely beats 290 and 290x and sometimes even looses
It's all about the power consumption though. Here's hoping that AMD's GCN Hawaii refresh in early 2015 brings them up to par with GTX 970 and 980 performance per watt.Quote:
Originally Posted by Smikis
Those figures likely include AMD's APUs and Intel CPU HD rubish. So removing those you'd get Nvidia as the company with the greatest discreet graphics card shareQuote:
Originally Posted by kalnielBoth AMD and nVidia have 14.1% of the market share as of latest report, Q3 2014.Quote:
Originally Posted by cptwhite_uk
Intel are still leading on 71.8% though.
And that matters why? Increasing IGP performance negates the need for many lower-end discreet cards, which make up a large part of the market, especially in mobile. Whether it's part of the same die as the CPU, or connected through PCIe, it's still a GPU at the end of the day. As I said in another thread, you can set cut-off points in various places to make the figures look better or worse for a given company. E.g. why not include the GPU in all three consoles for AMD? That would significantly change things...Quote:
Originally Posted by MustardCutter
Was GCN1.2 on the R9 285 something that I dreamt? It seemed quite a good part, beating the mainstream 760 on performance per watt, managing to improve performance over the 280 despite a reduction in memory bandwidth showing great improvements in architecture & balance & a competitive price, but it is like it never happened.Quote:
Originally Posted by cptwhite_uk
Ofc GCN1.2 was what people like Anandtech called it, because AMD say absolutely nothing about card internals. Perhaps that is the biggest difference here.
The dream then turned into a nightmare with the part being different enough that it is currently all but unusable under Linux, so even if it was real I couldn't have one.
And GCN 1.1 for Hawaii, Bonaire. There's a big difference in marketing though.Quote:
Originally Posted by DanceswithUnix
Quite. People talk as if Nvidia are bounding unstoppably away from AMD, but not at any price point I am willing to look at.Quote:
Originally Posted by watercooled
Is the 760 a fair competitor to the 285? If it is then Nvidia should manage something better this time around because that keeps up the usual leapfrogging we are used to and by Anandtech's figures Nvidia are currently losing.
http://www.anandtech.com/bench/product/1333?vs=1038
I'm not as technically aware as others here. When I say refresh I mean, as a layman, I understand the last major refresh to be the release of the GCN architecture in Dec '11 / Jan '12 with the 7 series of cards. I thought the R9 series was basically a GCN refresh, with tweaks and added features for Audio etc. but the main silicone design remained largely unchanged. As I understand the new stuff coming is going to be based on a die shrink and a re-worked silicone design, hence my assumption it's the first major refresh in 3.5 years? Or am I wrong in what I'm saying?
I'm also an AMD user so don't think I'm trying to be a fan boy on this issue.
But by that metric the last major Nvidia refresh was Fermi. The difference is that if Nvidia issue a bugfix like the 480 to 580 update they call it a new architecture. In reality, both sides hold on to a basic architecture for about 5 years before doing any serious update. Before Fermi we had 8800GT, before that we had I think the 5000 series. The stuff between is evolutionary, not an architecture change. Nothing wrong with that, it is the sane way to develop, but the marketing guys seem to go a bit wonky sometimes.Quote:
Originally Posted by cptwhite_uk
So are we looking at a ground up refresh this time from AMD? But other than the flagship 7 series generation and the R9 generation AMD must have been going 18-24 months between major generation flagship products. I'm sure things never used to progress that slowly, we use to have flagship single core products superseded on rougly a 12 month cycle?
That'd be a change in architecture, and you're perhaps being fooled by the very shortlived VLIW4 architecture that appeared only in the 6970/50 (and was itself transitional, being an intermediate step between VLIW5 and GCN). VLIW5 lasted right from the 2900XT through to the 6870/50 (and is still going for low end parts), with some weight loss and efficiency improvements along the way.Quote:
Originally Posted by cptwhite_uk
For nvidia I think it's fairer to say their last new architecture was Kepler, where they went from a small number of complex cores to a large number of simpler cores - Fermi was kind of a transitional architecture in that sense, much like VLIW4 for AMD. Maxwell is a refresh of Kepler in the same way as Hawii/Bonnaire were a refresh of GCN1.0, but AMD don't seem to be using refresh codenames any more, simply referring to all their cards as GCN. Not so hot for the marketing department, but perhaps more honest, technically....
I think that's for the market to decide. If an APU is good enough for someone's uses then fair play to it. Today's APUs are as powerful in the graphics department as add in boards of previous years.Quote:
Originally Posted by MustardCutter