Right.
I'm going to act the dumb blonde here.
I can see what's the cut/paste/rescale on the left-hand side of the pic, that it looks like a Fermi thing blah. Which is kind of what you might expect if they graft a Fermi thing onto another thing.
Where's the cut-and-paste of the right-hand side from Fermi stuff though??? Which bit of Fermi is that if the whole thing is fake?!
I realise that BSN is BS, that that article is 31% vitriol towards an unmentionable individual who is also a rumour-monger, blah blah, and so on. But I'm struggling to see the mapping from the right-hand-side of the upper supposed die-shot to the lower supposed die-shot, regardless of the dotted white lines.
?!
I'm going to act the dumb blonde here.
I can see what's the cut/paste/rescale on the left-hand side of the pic, that it looks like a Fermi thing blah. Which is kind of what you might expect if they graft a Fermi thing onto another thing.
Where's the cut-and-paste of the right-hand side from Fermi stuff though??? Which bit of Fermi is that if the whole thing is fake?!
I realise that BSN is BS, that that article is 31% vitriol towards an unmentionable individual who is also a rumour-monger, blah blah, and so on. But I'm struggling to see the mapping from the right-hand-side of the upper supposed die-shot to the lower supposed die-shot, regardless of the dotted white lines.
?!
This puppy here is Denver!
DarthShader
Regular
Fermi shaders might be pretty potent, but I doubt 256 of them are able to emulate an ARM CPU, particulary the Load/Store and L1DCache on the left side.
It's indeed a bit harder to spt, since the right side is spammed with coloured rectangles. But if you look for standing out points in the shader groups on the Fermi die, you will be able to see them on the "Denver die" too, so it's another group of 4SMs, just like the left hand side.
Sorry to bump an old thread.
But I was wondering what the speculation is for how many arm cores would be integrated into the Maxwell GPU?
Would they still be using the more traditional but updated stream (Fermi style) processors along side this?
If so whats the thinking on number of units stream processors needed to achieve the flops count quoted by nvidia.
Would the number of flops from the Arm cores be a significant contribution to the overall power of the GPU?
But I was wondering what the speculation is for how many arm cores would be integrated into the Maxwell GPU?
Would they still be using the more traditional but updated stream (Fermi style) processors along side this?
If so whats the thinking on number of units stream processors needed to achieve the flops count quoted by nvidia.
Would the number of flops from the Arm cores be a significant contribution to the overall power of the GPU?
My understanding is that Maxwell is just the name of the GPU architecture, just as Fermi is the name of the architecture, not GF100 specifically. So many combinations would be possible, including no ARM cores at all. And I actually think that's very likely for GeForces. ARM cores are small, especially on 20nm, but I still think that NVIDIA won't want to waste any silicon on CPU cores that wouldn't be useful for gaming.
Yes, most likely. Well, the SMs would be closer to what you'll find in Kepler, but that's the same lineage, so to speak.
Have they actually given any hard numbers? In any case, 2 operations per FMA unit, clocks probably somewhere in the 1.4~2.0GHz range, so I'll let you do the math. If the ARM cores are Cortex A15s, they would be capable of 4 operations per cycle (8 if they support FMA, I can't recall if they do) and would probably operate around 2.0GHz. So with, say, 8 cores, that would be 64 GFLOPS, or 128 with FMA.
Currently, GF110 is over 1.5 TFLOPs, and I think Maxwell should exceed 3.
According to: http://gruntthepeon.free.fr/blog/index.php/2011/05/26/54-arm-cortex-a9-peak-flops
3.99 G/flops per core per 1GHz A9 (dual or single not mentioned).
I would hazard a guess and say no.
spacemonkey
Newcomer
According to Charlie we'll have to wait until 2014... (at least)
http://semiaccurate.com/2011/12/30/nvidia-gpu-roadmap-slips-a-year-too/
http://semiaccurate.com/2011/12/30/nvidia-gpu-roadmap-slips-a-year-too/
And the "new" roadmap has been floating around the net for several months. This is old news either way.
Yeah, the new one with 2014 has been public for months. Don't know why it took Charlie so long to catch up.
The charitable explanation. Another is trolling for page hits. Looks like it worked.rpg.314 said:
spacemonkey
Newcomer
In Charlie's defense, if you do a google image search for 'nvidia roadmap' his is the first (and only?) roadmap image that shows Maxwell in 2014. I know that there's a lot of speculation and provocative language in his stories, but for some of us he does actually provide useful news.
Which is a truly sad thing (no metaphor even). It is akin to saying that an entity like Fox News, for example, does actually provide useful anything.
Richthofen
Newcomer
I didn't took him that long to catch up. He knows the roadmap for months. The thing is he doesn't know much about Kepler and has missed quite a few tapeouts. And now he is desperatly looking for page hits - journalism at its worst. I wouldn't even call it journalism.
DarthShader
Regular
oooh, pls share!
As I know the "2014" roadmap is known since June/July 2011 (here are all 3 versions I found - from above: 2010, June/July 2011 and the last one: May 2011). Anyway, I agree that people are bashing Charlie unjustly. There were many websites spreading fakes and bullshit befor the GCN's release, but nobody criticises them.
Richthofen
Newcomer
Well i think OBR has been critisized a lot, which usually happens when your are running an agenda with your website. Same happens to S|A.
