Breaking News
News & Analysis

Data Centers May Ride on ASICs

Research Targets 2020 System
8/28/2014 06:00 AM EDT
10 comments
NO RATINGS
Page 1 / 2 Next >
More Related Links
View Comments: Oldest First | Newest First | Threaded View
rick merritt
User Rank
Author
Google already eyeing IBM Power 8
rick merritt   8/28/2014 10:28:45 AM
NO RATINGS
Interestingly the talk was given on the Google campus where engineers are developing Power 8 boards, eyeing possibilities for its data centers.

wilber_xbox
User Rank
Manager
Re: Google already eyeing IBM Power 8
wilber_xbox   8/28/2014 12:06:52 PM
NO RATINGS
Yes Rick, and with others also modifying their server and storage profile the market for data centers, warehouses, server and storage is pretty exciting for coming years. The goal will be to make on-chip communication work by integrating silicon based photonic devices for faster communication and computation.

AZskibum
User Rank
CEO
GPUs before ASICs
AZskibum   8/28/2014 1:16:11 PM
NO RATINGS
In the mean time, before all those new custom ASICs get to production, a great deal of acceleration can be realized by incorporating GPUs into the data center.

rick merritt
User Rank
Author
Re: GPUs before ASICs
rick merritt   8/28/2014 4:38:18 PM
NO RATINGS
@AZskibum: I beg to disagree. At Hot Chips, Microsoft and Baidu researchers both said they looked at GPUs but found them too hot and costly.

They are both going with FPGAs for some specific apps, but both complain they are still too hard to program.

rick merritt
User Rank
Author
Re: Google already eyeing IBM Power 8
rick merritt   8/28/2014 4:39:19 PM
NO RATINGS
@Wilbur: Indeed Patterson and team will put optical links (sans lasers) on their ASICs and expects others to do so come ~2020.

docdivakar
User Rank
Manager
Re: GPUs before ASICs
docdivakar   8/28/2014 4:56:31 PM
NO RATINGS
@Rick: not so fast! @AZskibum does have a point in applications similar to those that FBook is pushing thru its Wedge reference design. Wedge does include micro servers that can benefit from GPU's, thermal effects notwithstanding.

MP Divakar

DrFPGA
User Rank
Blogger
SPU not GPU
DrFPGA   8/29/2014 11:01:25 AM
NO RATINGS
Seems like the argument is for software processing units. Identify the core functions that are currently limiting software bandwidth like garbage collection and fix them with specialized hardware functions... Let's not call these hardware elements ASICs let's call them SPUs.

rick merritt
User Rank
Author
Re: SPU not GPU
rick merritt   8/29/2014 12:43:17 PM
NO RATINGS
@DrFPGA: In the short term, Fbook is clear it will support x86 and ARM server SoCs in the Group Hug board in Wedge.

In a side discussion, Yuval suggested FPGAs or other programmable elements are not ruled out by any means depending on what apps develop.

AZskibum
User Rank
CEO
Re: GPUs before ASICs
AZskibum   8/31/2014 10:02:50 PM
NO RATINGS
FPGA's are hardly different from ASICs in this context, besides being reprogrammable. Either way, with FPGAs or ASICs we're talking about implementing designs that specifically target the acceleration of particular processing tasks in a datacenter.

GPUs, on the other hand, have a lot of library support for parallel high-performance computing.

An FPGA or an ASIC could potentially offer lower power and/or lower cost acceleration -- once you have decided exactly what to implement in that silicon and have developed & debugged the library functions to allow your existing software stack ot access the new silicon.

With GPUs, all of the exists already, which is why I titled my post "GPUs before ASICs."

 

 

krisi
User Rank
CEO
Re: GPUs before ASICs
krisi   9/8/2014 3:50:50 PM
NO RATINGS
"FPGA's are hardly different from ASICs"...well, ASICs can pack up way more transistors for one

August Cartoon Caption Winner!
August Cartoon Caption Winner!
"All the King's horses and all the KIng's men gave up on Humpty, so they handed the problem off to Engineering."
5 comments
Top Comments of the Week
Like Us on Facebook

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Times on Twitter
EE Times Twitter Feed
Flash Poll
Radio
LATEST ARCHIVED BROADCAST
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for todays commercial processor giants such as Intel, ARM and Imagination Technologies.