Researchers at MIT have successfully built a 36-core processor
Share on Twitter.
Get the most reliable SMTP service for your business. You wished you got it sooner!
June 28, 2014
Research scientists at MIT say they have cleverly built a new 36-core processor that uses an
internal networking system to get maximum data throughput from all the processing cores.
Unveiled at the International Symposium on Computer Architecture, the original chip design
gets around some of the issues with multicore processors, namely bus sharing between cores, and
maintaining cache coherence, among other things.
To be sure, most conventional designs use a single bus to connect various cores, meaning
that when two cores communicate together, they typically use the entire bus and leave other
The MIT design borrows from the internet's concept and allows all chips to share data with
their neighbors using their own router.
"Typically, you can reach your neighbors really quickly," said Bhavya Daya, an MIT graduate
student in electrical engineering and computer science, and the first author on the new paper.
"You can also have multiple paths to your destination. So if you're going way across, rather
than having one congested path, you could have multiple ones," he added.
The network is also used to distribute various data between each core's cache without having
to shift it too far, potentially speeding up the system even further.
"Their contribution is an interesting one-- They're saying, 'Let's get rid of a lot of the complexity
that's already in existing networks. That will create more avenues for communication, and our clever
communication protocol will sort out all the details'," said Todd Austin, a professor of electrical
engineering and computer science at the University of Michigan.
"It's a simpler approach but a much faster approach at the same time. It's a really clever idea
of doing things."
The blueprints for the new chip design aren't being released as of yet, since the team first
wants to develop an operating system capable of using it to its full advantage.
The team is now adapting a version of Linux to use the new chip before releasing the designs.
In other IT news
The president of the Open Data Centre Alliance (ODCA) has given what some would call smart
advice to CIOs contemplating how they migrate their legacy platforms into the cloud-- forget
it and just dump your old code!
However, Correy Voo, whose job is the infrastructure CTO at UBS, added this was likely a temporary
dilemma as the coming wave of technology bosses, who’ve grown up with theoretically unlimited
resources at their fingertips, will not even contemplate such a move.
“Don't try to lift a legacy platform and make it a cloud process. There are some things that
just will not fit a cloud environment,” Voo told an audience at the Cloud World Forum event in London
An application developed to deal with the constraints of the 80s and 90s, or even earlier, would never
take full advantage of the new platform, he argued.
Banks and other financial institutions, many of which are members of the ODCA, will have
stacks of such apps languishing on mainframes or other aging platforms.
“I’m not saying it’s impossible. It’s certainly possible if you want to throw a lot of time
and effort at it. But if I want to buy something brand new like a $3,000 TV I don’t want to play
low definition material on it."
“In my view, the users in some ways are being cheated because they don’t get the full capability
of that investment.” He cited mobility as one example of this mismatch.
The discussion about whether to migrate applications designed to cope with the hardware and
infrastructure constraints of the 80s and 90s has a parallel in the differing attitudes towards
infrastructure between younger and old tech pros.
“That discussion is coming to a head in some respects,” Voo added. “Volume wise we’re now at
the stage where there’s sufficient volume of new people in the industry who think differently
to the old people.”
“It’ll be very fractious early on and there’ll be a lightbulb moment where everybody will realise
that’s probably the right way of doing it.”
“When it becomes easy to consume cloud technology without any of the complication, any of the
cost implications, when that happens it’s a no-brainer-- you simply move on to the next issue.”
As for what that problem that is? Voo said it will be about how we manage and access data.
“Just because we have the ability to use a piece of data, doesn’t necessarily mean we should
ethically, so I think the IT industry and business in general have a number of questions to resolve
in relation to data growth the type of data that’s out there.”
In other IT news
MongoDB has inked new agreements that will place the company's NoSQL database into Google and
Microsoft's Cloud data centers.
The company announced on June 24 that it had sealed pacts with the two companies to make it
easier for developers to launch MongoDB databases on top of their respective clouds.
With these two new relationships, MongoDB has gained a foothold in all three of the major clouds.
It already had a relationship with Amazon that made it available on the company's eponymous
By partnering with big internet and software companies, MongoDB has cut the steps it takes
for developers to get the database working in public clouds, with MongoDB now available as an
Add-On in the Microsoft Azure Store, and also via a preview "Click to Deploy" button on Google.
"Fundamentally, we want developers to be able to bring up a MongoDB cluster on any infrastructure
they want very easily," explained MongoDB's CTO, Eliot Horowitz.
"Technically, people have been running databases on their cloud infrastructure for
a while so it was just a matter of the integration," he added.
As any entrepreneur already knows, if you aren't sticking your product in front of users it's hard
to sell it, so MongoDB's deals are a big factor in its broad developer adoption.
These agreements follow a spate of activity in the past few months that has seen the company
join Pivotal's "Cloud Foundry Foundation" to help tie its system closer to the company's eponymous
platform-as-a-service, link its sales and marketing and technology organizations with Hadoop-expert
Cloudera, and team up with IBM to make it easier for MongoDB apps to work with data kept in DB2 and
With the new Google and Microsoft agreements, MongoDB appears just as keen to make its database
prevalent in the cloud as it is to be ubiquitous in on-premises technology.
Though MongoDB isn't yet in the same league as the database incumbents it's trying to displace,
it's doing a much better job of getting its name out than its young rivals DataStax, Couchbase and
In other IT news
Microsoft said earlier this morning that it has fixed a nine-hour service outage in its
online Exchange service that crippled North American customers’ Office 365 and hosted Outlook
The service outage struck companies in the U.S., Canada and Mexico, leaving entire organizations
unable to read, write or retrieve emails or seeing their email slow down to a snail's pace.
Users manned the Office 365 community message boards and Twitter accounts to complain and
vent out their frustration, singling out a flat-footed response from both Microsoft’s technical
and customer service staff dealing with the issues.
Customers were angry at being kept on hold by customer support for up to two hours at a time
and at not being given a set ETA.
Microsoft finally announced that the issues were resolved late Tuesday evening Eastern time.
David Zhang of Microsoft support told the Office 365 community message board-- “Our investigation
determined that a portion of the networking infrastructure entered into a degraded state. Our engineers
made some configuration changes on the affected capacity to resolve end-user impact. The issue
was successfully fixed on June 24, 2014 at 9:50 PM EST.”
Zhang also issued the boilerplate corporate apology for any inconvenience caused but he could
offer little to satisfy users or downplay the magnitude of the problem.
During the day, Jim1001 wrote on the Office 365 community page-- “Our entire corporation cannot
send or receive emails from Outlook (Office 365 Exchange) or even the OWA web browser as of 8 AM MST
time this morning June 24, 2014! I have never seen a world-wide email go down like this before.
MadBuffalo wrote-- “Office 365 is beginning to look like a very poor choice for mission critical
Trec posted-- "Tried to contact support by phone, and after a half hour, the call was dropped on
both occasions. It seems the support guys are saturated and there are not enough staff for an
huge issue like this one."
For many, it was the second outage to Microsoft’s business-cloud service in less than 24
hours and it was proving way too much.
Additionally, Monday saw Lync Online, Microsoft’s unified communications service, slow down and crash.
Lync provides voice over IP, corporate instant messaging, presence, meetings and video conferencing.
Problems surfaced around 11 am EST and were addressed by the Microsoft Lync team via Twitter
later in the afternoon.
Service was restored by 5.30 pm EST. MMS Infotech on Twitter blamed the problem on a botched
data centre migration by Microsoft technicians, moving the relevant network infrastructure.
Source: The Massachusetts Institute of Technology.
Get the most dependable SMTP server for your company. You will congratulate yourself!
Share on Twitter.