Intel Gunning To Challenge Nvidia At HPC With ‘Knights Landing’ Xeon Phi Processor
Intel's Knights Landing is set to offer three times the amount of performance as the current-gen Knights Corner.
Intel (NASDAQ:INTC) has shed further details on its second-generation Xeon Phi CPU, known as Knights Landing.
The processor features several technical achievements, starting with a 14nm manufacturing process, which is a first in this series. Designed to offer high-performance computing, Knights Landing differs from other server-based CPUs in that it uses lots of low-energy cores to run parallel tasks, whereas offerings from IBM (NYSE:IBM) or Oracle (NYSE:ORCL) use fewer but more powerful cores.
Built on Intel’s MIC (Many Integrated Core) architecture with a total of 8 billion transistors, Knights Landing runs a modified version of the Atom Silvermont x86 core in a tile configuration, with a single tile featuring two cores and vector execution units along with shared L2 cache as well as a circuitry that connects the tile to the rest of the mesh network. Intel has mentioned that each Knights Landing package would include a processor with 30 or more tiles and eight on-chip memory modules. Another major highlight with Knights Landing is that it would be able to function as a host processor, meaning that it can boot and run x86 operating systems and application code without any need for recompilation. It can also act as a co-processor.
Talking about memory, the chip vendor has announced that Knights Landing would feature eight 2GB stacks of memory, totaling up to 16GB. The chip is manufactured at Micron, and looks to be a variant of the manufacturer’s Hybrid Memory Cube, which involves stacking memory and using an embedded logic chip to deliver higher bandwidth at a lower power. Micron has mentioned that its HMC modules will be able to transfer data 15 times faster than a standard DDR3 module, while utilizing 70%
"14nm Core-Ms, Cherry Trails, Broxtons, Sofias are all going to wreak havoc
Like the 22nm, 32nm parts before them eh?:)"
Don't you know INtel is coming out with something next year ?
Only this time... the competition is thru :)
"BK isn't bad with a nimble forward-looking outlook"
Nimble was exactly what INtel needed. BK is more williing
to open INtel fabs. Not there yet, but embraced it more than Otellini.
"Otellini wasn't bad either considering the mess he inherited"
My friends at INtel greatly respected him, said he modernized INtel internally.
"They crushed Nvidia's tablet marketshare so what did I get wrong "
NVidia lost it all on their own. But Tegra business is growing now...
Anyway, evidence is $4B in bribes works :)
"I told getanid61 that already"
Yes, for years & years you've been repeating & repeating:
INtel is coming out with something next year... only this time, the competition is thru
Instead of INtel crushing the competition, x86 needed $4B in bribes to stay relevant.
What is stupid about..."Global graphics card shipments in 1Q15 to fall 20-25%" ??
"Rediculous message in your link. "
The point was, which you seemed to have missed entirely:
"Global graphics card shipments in 1Q15 to fall 20-25%"
More than 48 core ARMv8 chips to unlock the data centre door
ARM Holdings is telling investors it will take 20 per cent of the server silicon market by the year 2020.
The company made the claim last week at an event in Taiwan, sending ripples through the Asian media.
A little searching shows that presentations like this one (PDF) shows that the 20 per cent claim has been used in chats with investors since mid-February 2015, with the diagram below offering a hint of how the company thinks it will get there.
Elsewhere in the presentation, ARM makes it plain that it thinks core data centre workloads are within its reach, especially in hyperscale data centres. The slide deck also guesstimates that by 2020 ARM chippery will be present in just 20 devices, suggesting that the company is wooing white box operators and/or vendors' niche servers
ARM's long looked a data centre contender thanks to its chips' low power consumption and cool runnings, ideal features for the dense rigs needed to run efficiently at very large scales. That folks like Cavium have backed the company's platform by building a 48-core ARMv8 chip, and making sure it plays nice with the Xen hypervisor, doesn't hurt either.
There's lots of interest in ARM
semi_punk, again I call your bluff !!!
"a guy who claimed Trigate/FinFet would be only applicable to CPUs and FPGAs but not to SoCs -"
why do you make stuff up ??
as expected you can't back up anything you've claimed...
You claim I say things... but can't back up a single comment you make !!
"play" ARM by buying chip equipment ..."
anybody understand this ?? anyone ??
"but you are thick as a brick "
"G61 dd not understand how 2 D defect density would apply an Intel 3 D "
First, you gotta learn English...
semi_punk, again I call your bluff !!!
Find any comments of mine... or I guess you'll just have to make up some more...