Showing posts with label ARM supercomputer. Show all posts
Packet.net strong-ARMs cloud for $0.005 per core per hour
CEO
Zachary Smith told The Register that the company's cooked up the cloud for a
few reasons. Price is one: Packet will offer ARM cores at a tenth of the price
it charges for Intel cores, at US$0.50 per hour per server, or $0.005 per core
per hour. Smith thinks that will be a head-turner by itself.
Packet.net,
a bare-metal cloud aimed at developers, has flicked the switch on cloud-running
servers powered by a pair of Cavium's 48-core ARMv8-A ThunderX processors.
He
also thinks developers will appreciate the chance to try native Docker on
many-cored machines and appreciate the opportunity an ARM-powered cloud
represents as they pursue 100 per cent portable software. He believes open
source folk will see the arrival of an ARM-powered cloud as incentive to
accelerate cross-platform versions of their pet projects.
Even
ARM will benefit, he says, because having a working cloud on the market will
give both it and licensees more reason to innovate for the data centre.
ARM's
recent purchaser, SoftBank, recently tipped some money into Packet.net, but Smith
swears he's had a long-term ambition to offer an ARM-powered cloud, if only
because he enjoys having multiple ARM server CPU vendors willing to do deals.
That kind of competition is not currently possible in the x86 world, at least
until AMD returns to servers in 2017.
Smith
also feels that ARM clouds are inevitable, probably thanks to telcos looking to
offer cores to rent at the edge of their networks. The CEO feels that telcos
will build edge clouds because they're sick of over-the-top players having all
the fun and profits: this time telcos want to build a revenue-generating
platform beyond mere carriage.
For
now, Packet's ARM cloud offers 64-bit Ubuntu 16.04, but promises that CoreOS,
FreeBSD and CentOS are in the pipeline. Four different ARM server
configurations are also in the works.
The
cloud will have an API, a portal, and will also be accessible from DevOps
favourites likes Terraform and Ansible. Four of the company's bit barns – in
Parsippany New Jersey, Sunnyvale California, Amsterdam and Tokyo – will offer
the service as of Tuesday.
"We
want to offer a super-cheap, 'you would be stupid not to try it'
offering," Smith told The Register. "If we can get the open source
ecosystem rebooted, I think Intel's grip on the data centre will be
shattered." ®
Japan's exascale Flagship 2020 looking more like Flagship 2022 after CPU design headaches ®
Fujitsu's
monster ARM-powered supercomputer, the Post-K, will miss its 2020 deadline.
The
$910m project has stalled because its engineers need more time to design, test
and perfect the machine's processors, we can confirm.
Fujitsu
was hired by Japan's boffinry nerve-center RIKEN to produce a roughly 1,000
peta-FLOPS supercomputer to supersede the nation's K Computer, which today is
the fifth fastest known supercomputer.
When
the Post-K emerges, and if it lives up to its near-exascale performance
promise, it will be eight times faster than today's most powerful known
supercomputer in the world, China's Sunway TaihuLight. The Post-K system will
be used to model climate change, predict disasters, develop drugs and fuels,
and run other scientific simulations.
Fujitsu,
which built the K Computer predecessor using heavily customized SPARC64 VIIIfx
chips, switched to the ARMv8-A architecture for its Post-K beast. However, The
Register understands that the new supercomputer will be one to two years late,
thus arriving in 2021 or 2022, which is kinda awkward because the Post-K's
official name is Flagship 2020.
The
delay may also take some of the shine off Tokyo-based Softbank's just-completed
acquisition of ARM, the UK chip designers that licensed their 64-bit technology
to Fujitsu for things like the Post-K beast. ARM and Fujitsu worked together to
add large vector instructions to ARMv8, bringing the architecture up to scratch
for supercomputer applications.
"There
has been a delay to development due to us applying a new CPU related
semiconductor technology. As a result, the time required for making system
prototypes and detailed designs has now been extended," Fujitsu spokesman
Rishad Marquardt told El Reg on Tuesday.
"Currently
it is expected the delay will push back the operation start time by between 12
and 24 months."
Engineers
working away on the Post-K were expected to put the finishing touches to their
blueprints by early 2018, and fire up the completed system in the first quarter
of 2020. Now that looks likely to happen up to a year or two after. Fujitsu
would not comment on exactly what the problem is: whether it's that moving from
SPARC64 to ARMv8 is proving trickier than expected, or that shrinking designs
down to 10nm is becoming a real headache, or both.
The
Post-K's ARM processors are likely to be 10nm FinFET chips fabricated by TSMC,
and will feature high-bandwidth memory and the Tofu 6D interconnect mesh [PDF]
that's used in the K Computer. It could be that getting decent yields of
processors including all these bells and whistles is going to be harder than
expected.
"We
expect more exascale projects and more delays as the engineering challenges
mount," said The Next Platform's Tim Prickett Morgan, who was first to
report the Flagship 2020 delays after hearing from sources at RIKEN. The
science organization was really hoping to beat America to powering up the first
publicly known exascale supercomputer: the US government doesn't expect to do
so until 2023.
"We
also think that compromises will be made in the power consumption and thermals
to get workable systems that do truly fantastic things with modeling and
simulation," Prickett Morgan added.
Meanwhile,
Fujitsu announced today it will build a 2,380-core x86-based computer system
for the Institute for Cosmic Ray Research at the University of Tokyo.
