Breaking News
News & Analysis

Nvidia describes 10 teraflops processor

11/17/2010 03:47 PM EST
4 comments
NO RATINGS
Page 1 / 2 Next >
More Related Links
View Comments: Threaded | Newest First | Oldest First
ChuckLippmeier
User Rank
Author
re: Nvidia describes 10 teraflops processor
ChuckLippmeier   11/17/2010 11:31:19 PM
NO RATINGS
Ya so what. I've been trying to get ahold of a sales person at the NVIDIA webstore site for two weeks and I've been sent on wild goose chases to PTC and Microsoft service organizations, neither of which knows anything about NVIDIA. I'm not interested in anything NVIDIA has to say anymore.

ChuckLippmeier
User Rank
Author
re: Nvidia describes 10 teraflops processor
ChuckLippmeier   11/18/2010 7:47:06 PM
NO RATINGS
Well, I finally received a response with the information I requested. If this posting was the reason, thank you but it's a heck of a way to get into NVIDIA's sales/Tech support.

vivekv80
User Rank
Author
re: Nvidia describes 10 teraflops processor
vivekv80   11/18/2010 6:23:07 PM
NO RATINGS
awesome is this Kepler or Maxwell? Hope they allow DMA and GPUs will make a mark in embedded processing :)

kjdsfkjdshfkdshfvc
User Rank
Author
re: Nvidia describes 10 teraflops processor
kjdsfkjdshfkdshfvc   12/3/2012 2:33:52 PM
NO RATINGS
Maxwell, with ARM 64bit cores included,probably Neon SIMD ?, and using the existing ARM one Terabit of usable system bandwidth per second interconnect http://www.arm.com/products/system-ip/interconnect/corelink-ccn-504-cache-coherent-network.php and its also probably why your only now seeing Intel talk about their proposed one Terabit Non cache coherent interconnect in an upcoming paper, as they missed that ARM innovation to start with :) i do find it a little odd that Intel are not making use of their in house "Light Peak" optical fiber research here though, didn't they manage to get it on-chip YET by now and cheaper... if only for the higher speed bus and not the ultra low power optical information processing yet.

Most Recent Comments
michigan0
 
SteveHarris0
 
realjjj
 
SteveHarris0
 
SteveHarris0
 
VicVat
 
Les_Slater
 
SSDWEM
 
witeken
Most Recent Messages
9/25/2016
4:48:30 PM
michigan0 Sang Kim First, 28nm bulk is in volume manufacturing for several years by the major semiconductor companies but not 28nm FDSOI today yet. Why not? Simply because unlike 28nm bulk the LDD(Lightly Doped Drain) to minimize hot carrier generation can't be implemented in 28nm FDSOI. Furthermore, hot carrier reliability becomes worse with scaling, That is the major reason why 28nm FDSOI is not manufacturable today and will not be. Second, how can you suppress the leakage currents from such ultra short 7nm due to the short channel effects? How thin SOI thickness is required to prevent punch-through of un-dopped 7nm FDSOI? Possibly less than 4nm. Depositing such an ultra thin film less then 4nm filum uniformly and reliably over 12" wafers at the manufacturing line is extremely difficult or not even manufacturable. If not manufacturable, the 7nm FDSOI debate is over!Third, what happens when hot carriers are generated near the drain at normal operation of 7nm FDSOI? Electrons go to the positively biased drain with no harm but where the holes to go? The holes can't go to the substrate because of the thin BOX layer. Some holes may become trapped at the BOX layer causing Vt shift. However, the vast majority of holes drift through the the un-dopped SOI channel toward the N+Source,...

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed