News & Analysis
Top 10 disruptive startup tech plays
Peter Clarke
1/23/2013 10:20 AM EST
#7. Adapteva close to sampling
At #7 is Adapteva close to sampling 28-nm, 64-core coprocessor by Peter Clarke.
This article was posted in March 2012 and marked a stage in the development of the company where company CEO and founder Andreas Olofsson was already seeking to get chips to market without the punishing development costs of conventional design cycles. The fourth generation of the Epiphany multicore architecture IP is a 28-nm design with 64 independent RISC cores, each with 32-kbytes of memory on an 8.2 square millimeter die.
Adapteva went on later in the year to raise nearly $900,000 on the crowd-sourcing Kickstarter web-site to fund a mask set for its Epiphany processor and seed a developer community of nearly 5,000 potential customers in the process.
Related links and articles:
www.adapteva.com
News articles:
Adapteva kickstarts $100 supercomputer
Debunking the myth of the $100 million ASIC
From RTL to GDSII in just six weeks
At #7 is Adapteva close to sampling 28-nm, 64-core coprocessor by Peter Clarke.
This article was posted in March 2012 and marked a stage in the development of the company where company CEO and founder Andreas Olofsson was already seeking to get chips to market without the punishing development costs of conventional design cycles. The fourth generation of the Epiphany multicore architecture IP is a 28-nm design with 64 independent RISC cores, each with 32-kbytes of memory on an 8.2 square millimeter die.
Adapteva went on later in the year to raise nearly $900,000 on the crowd-sourcing Kickstarter web-site to fund a mask set for its Epiphany processor and seed a developer community of nearly 5,000 potential customers in the process.
Related links and articles:
www.adapteva.com
News articles:
Adapteva kickstarts $100 supercomputer
Debunking the myth of the $100 million ASIC
From RTL to GDSII in just six weeks
Navigate to related information


daustins
1/25/2013 11:37 AM EST
I am an Anglophile and have been all my life. Still, is it necessary to call a Brit or anyone by a title? Americans do not have nobility or royalty, de jure. Call him, or any other ennobled Brit, by his last name, or first, if you're familiar. As I said, I'm an Anglophile but we don't use fancy titles for Rothchilds and anyone in the Almanach de Gotha.
No reason for an American to fawn over someone, even a cousin.
Sign in to Reply
R. S. Moore
1/25/2013 11:42 AM EST
But he is brave Sir Robin.
Sign in to Reply
TJones
1/25/2013 12:36 PM EST
My name is Inigo Montoya. You killed my father. Prepare to die.
Sign in to Reply
trm1945
1/25/2013 12:53 PM EST
Hello there, what's your name?
"My name is Pedro."
Pedro. That's Mexican, isn't it?
No, ees Swees. Theese way we don offend no buddy."
Sign in to Reply
hari.tadepalli
1/25/2013 1:00 PM EST
Appreciate these techonology development insights. But, . . ., one quarter column worth of reading material in each click? My 1920 x 1080 laptop screen can fit at least six of these slides in a screenful. I understand the need for sponsors and advertisements; but with all the formatted content on this page, inclidng the ad, occupy less than 50% of my screen, with two large white margins left open on either side. For reading convenience, please consider posting at least two of these slides on each page.
Sign in to Reply
RAJEEV.KRISHNAMOORTHY_#1
1/25/2013 1:15 PM EST
I completely agree with the comment about the irritating format. I frequently stop reading articles because I get turned off. In this case, I stopped after just a couple of slides.
Sign in to Reply
William Miller
1/28/2013 2:59 AM EST
Agree. Slides are not convenient for reading..
Sign in to Reply
jjgray
1/28/2013 6:39 AM EST
Agree! These articles always prick my interest and then prick my bubble as I recognise the tedious "1 of 10" format.
Sign in to Reply
TJones
1/25/2013 2:20 PM EST
I, on the other hand, enjoy clicking once for each word in the article.
This is especially effective on slower browsers, where the densely-populated and expertly-coded EET web page takes torturous seconds to load and I stop reading the article, but am _sure_ to read the advertisements, v e r y s l o w l y.
Sign in to Reply
alex7ko
1/27/2013 7:57 AM EST
I used to click "Print" when reading such multi-page articles as this.
Sign in to Reply
hari.tadepalli
2/5/2013 2:22 AM EST
Thanks for the suggestion. I missed the tip. Yes, clicking print takes you to an ad-free version with all the 10 slides on the same page. Some websites don't allow this, but works in EETimes.
Sign in to Reply
joyhaa
1/30/2013 3:15 PM EST
at least there should be a quick jump-to list for the 10 items.
also, on item 7 for servergy.com, I think they're using Freescale silicon, not sure why IBM was mentioned there.
Sign in to Reply
John_Galt
2/5/2013 10:51 AM EST
I read the Cyclos white paper with great interest. The resonant concept is simple but it may be too simple. The white paper focuses heavily on clock tree power and skew - to be sure these are vitally important aspects of IC design - but so is jitter. In certain applications jitter can be a show-stopper and my engineering sense tells me that a parallel resonant tuned clock tree (being of high impedance) would be very susceptible to jitter generated from crosstalk.
Sign in to Reply
selinz
2/12/2013 2:13 PM EST
Perhaps a drop-down menu (similar to those used in product reviews) that enable jumping to an arbitrary page would make the format less irritating.
Sign in to Reply