cypeq :
somebodyspecial :
For clarification in case you can't understand the bars: they're showing they have 0% share of desktops and servers and WANT TO OWN SOME share
Everything in that pic that is GREEN they are coming for. Of course an updated pic would show 25% or more of mobile computers as Chromebooks are 20% of NOTEBOOKS already. You don't have to believe me...ARM IS TELLING YOU.
My comment and the article is about K1.
K1 comes with denver just later this year. 64bit custom cores. Unless they amp mhz on it (or a special model to say 4ghz), it will probably only assault better than lowend notebooks and likely crap end desktops. But @20nm that changes, which I'm guessing will come with at least 1-3 more smx units and 3ghz+. But there is nothing stopping NV from slapping together a SOC box that uses their gpus. At worst, these will be pretty much as fast as AMD's APU's so well suited to desktops at 20nm Q1 next year, and this is before AMD has an Arm mobile or desktop part. Seattle is for servers-and it's based on the same thing as Denver (ARM V8 A57/A53...LOL). I guess you should email AMD and tell them they should trash Seattle today because Arm V8 isn't for servers
Not all desktops come with 2 gpus...LOL. Less than 2% do, but they can throw a K1 into a box with 2 gpus if they'd like to design one. Heck they may be able to make a small box work with thunderbolt and use an external GPU that is easily upgradable for a while...ROFL. Again, Denver is for desktops (no matter how weak you might think they are, or they turn out to be).
How does talking K1 change my story or your comment? It's for desktops/laptops/tablets/phones (at some point if they add a modem at 20nm, but that might be M1 based on maxwell or something).
http://www.theverge.com/2013/3/19/4123636/nvidia
March 2013 - Nvidia wants to be on every machine and OUST INTEL. They're talking CPU & GPU here.
"
the message was clear: Nvidia isn't just about entertainment and mobility; it can build serious personal computers if it wants to, and it might be able to do so
without help from Intel or AMD."
MESSAGE IS CLEAR. So without them YOU are powering the whole machine right? They are talking DENVER 20nm finfet with maxwell! There is no difference here between a SOC and an APU from AMD. What happens if I amp up the watts on a K1 (any arm) to 95w with an actual heatsink/fan attached? How fast do you think it is then? What makes you think they HAVE to be in a tablet/phone? A chromebook already is a lowened notebook (and it at 21% of their market share...ROFL). Now if the gpu isn't enough for a gamer, sell them a PCIE gpu for that box too
I don't understand your point. NV isn't alone trying to kill x86:
http://www.brightsideofnews.com/news/2012/9/20/nvidia-project-boulder-revealed-tegras-competitor-hides-in-gpu-group.aspx
At the same time, similar story is taking place at Apple, who wants to replace Intel products from its consumer notebook and desktop products.
What do you think they bought PAsemi for? A cpu team. What do you think a high watt SOC goes into? Laptops/desktops as they intend to do eventually at apple. Just waiting for iOS to replace MacOS now. Probably 2015/2016 as he suggest (iOS 8/9).
Nvidia's PROJECT DENVER page (K1 is DENVER):
http://blogs.nvidia.com/blog/2011/01/05/project-denver-processor-to-usher-in-new-era-of-computing/
"ARM is already the standard architecture for mobile devices. Project Denver extends the range of ARM systems upward to PCs, data center servers, and supercomputers. ARM’s modern architecture, open business model, and vibrant eco-system have led to its pervasiveness in cell phones, tablets, and other embedded devices. Denver is the catalyst that will enable these same factors to propel ARM to become pervasive in higher-end systems."
Higher end than tablets, UPWARD to PC's, supercomputers etc. What part of it do you not understand? Read the K1 article, this thing comes with Denver later.
"Denver frees PCs, workstations and servers from the hegemony and inefficiency of the x86 architecture."
Umm, so no INTEL? As he said before they'll either run it or stream to it, that is the goal. Nothing stopping them from slapping more than one in a larger die either (octa core would already have 750mflops then right? 2smx), and at 20nm this would easily become 4 smx at worst, up to 8 probably at best before blowing past AMD's ~245mm die of Kaveri. T4 was 80mm. If you tripled that size it would eat K1 for lunch. If K1 is the same size, triple that too. See the point, it's a Kaveri competitor at that point right? Iris too. I don't understand your logic. Nvidia already said they could have built the proc for Xbox1/PS4, it was not technical issues that stopped them, it was profits. Both consoles are just lowend pc's.
My comment was about K1 too...It's DENVER soon. Did you read the K1 article? Comes in two flavors.