Clearspeed Technology Share Discussion Threads

[mdrans1] 16/1/2007 07:38

see the header:-

RNS Number:5577P

Trading update

[mdrans1] 06/1/2007 12:04

The first time you hear it it's a "major deal" the next time it's mentioned it's "regurgitated marketing news" .


" IBM System Cluster 1350 will also offer ClearSpeed Advance™ accelerator boards" but this is not positive promotion of ClearSpeed boards ?

Go on Tradx666, as Brian said "Always look on the bright side of life de-dum de-dum Always look ......... "

[jcn1066] 06/1/2007 11:47

"Interestingly, and pretty much as I stated at the time (June) there appears to be no mention of Clearspeed in IBM's marketing literature,"

[tradx666] 06/1/2007 11:26

jcn,

I saw the regurgitated marketing news, but not anything in the product specs! My point was why...given that this was heralded as such a major deal, one would have thought that IBM would at least be positively promoting it?

regards

T..

[jcn1066] 06/1/2007 11:16

IBM System Cluster 1350 will also offer ClearSpeed Advance™ accelerator boards, which expands hybrid cluster architecture options. The PCI-X adaptors are designed to improve the performance of many numerically intensive workloads by routing math library routines to the ClearSpeed accelerator board. This acceleration process is transparent to the end user and the application, except for improved speed of calculation.

[tradx666] 03/1/2007 11:13

mdrans1,

Interestingly, and pretty much as I stated at the time (June) there appears to be no mention of Clearspeed in IBM's marketing literature, see;

ftp://ftp.software.ibm.com/common/ssi/rep_sp/n/CLD00221USEN/CLD00221USEN.PDF

Unless of course, they have hidden it somewhere else! But this latest marketing splurge was dated October, so I wonder if they have hit difficulties in validating the solution fit?

Given that they trumpeted this as a major tier1 OEM win at the time, surely there should be an update on this from CSD?

In any event, the absence of win/revenue news would suggest that revenue growth may well be a bigger struggle now? Perhaps this is why Tom and the non-exec sold significant stakes when the share price was around 255....?

Lack of news leading upto the year end, suggests that all may not be as rosy as it once was? Well it shouldn't be too difficult to get the full year accounts out swiftly, to take this uncertainty away - should it?

regards

T..

[mdrans1] 12/12/2006 15:21

Cheer up Trad. On the up at last ?

We haven't seen this sort of volume for a while. I note they are in Japan today.

[tradx666] 25/11/2006 13:53

56789,

I suspect the trigger will be the expected lack of growth in real underlying revenues, and the fact that in three major uk high visibility deployments they have not secured even a single card order, - very worrying...

regards

T..

[Abuse team] 25/11/2006 13:48

Post removed by ADVFN

[tradx666] 25/11/2006 13:46

Even more curious is that all that seem to be involved in this have very public collaborations with CSD!

Surely, CSD must be targetting institutions like cambridge as 'must win' projects?

regards

T..

Cambridge University supercomputer 'one of the world's fastest'

Publisher: Jon Land
Cambridge University
supercomputer 'one of the
world's fastest' A new supercomputer used by the University of Cambridge has been ranked as one of the top 20 fastest in the world and 7th fastest in Europe according to the latest top 500 list released yesterday.

The supercomputer, built by ClusterVision, uses Dell's 9th generation servers and storage units with Dual-Core Intel Xeon processors and will provide a central compute and storage facility to all scientists of the University of Cambridge.

Dr Paul Calleja, Director of High Performance Computing at the University, said: "Hosting one of the fastest supercomputers in the world is a major milestone for the Cambridge High Performance Computing Facility.

"I have been very pleased with the working partnership between Dell, Intel and ClusterVision which supplied the computer cluster.

"Dell's Dual-Core Intel Xeon processor-based hardware has proven to be very robust under the computational load already placed on it for the top 500 benchmark runs.

"ClusterVision's role has been key in rapidly turning the Dell-supplied hardware into a manageable cluster, ready for the top 500 benchmarks. We will soon make the cluster available to the university's academic community as a whole.

"Dell is pleased to be teaming up with Cambridge University on the UK's fastest academic supercomputer.

"ClusterVision played a key part in the installation and integration of the supercomputer, and will be providing first-line support services to Cambridge University."

[tradx666] 25/11/2006 13:43

I wonder why CSD didn't manage to get in on this?

regards

T..

Cambridge boffins unveil Dell supercomputer


OK computer? I'm super, thanks for asking...

By Gemma Simpson

Published: Wednesday 22 November 2006

The boffins at the University of Cambridge have joined forces with Dell to unveil one of the world's fastest supercomputers.

The machine, called Darwin, consists of 2,340 computers arranged in a cluster. It has been ranked as Europe's seventh fastest supercomputer, and the 20th fastest in the world, according to the top 500 list.

Get the latest public sector news straight to your inbox. Sign up for the PS newsletter today!
Dr Paul Calleja, director of high performance computing at Cambridge told silicon.com he chose Dell based on price. Darwin only cost £2m to install but the fastest supercomputer in the UK (ranked 12th in the world) cost £40m with only double the speed of Cambridge's supercomputer, Calleja added.

The supercomputer uses Dell's ninth-generation servers and storage units with Dual-Core Intel Xeon processors to provide scientists with super-fast simulations.

Calleja added: "We are now in discussions with Dell and Intel for development and research efforts."

The supercomputer will be used across departments at the university for various applications including weather simulations, modelling how air flows over a car and attempts to determine the origins of the universe.

Stephen Hawking is one of the supercomputer's most prolific users, Calleja added.

[jcn1066] 25/11/2006 13:14

ClearSpeed commits to 5x floating point boost


ClearSpeed also contends that its TiTech result has put server accelerators on the map for good, after years of unfulfilled hype around similar technology.

At $8,000 per board, ClearSpeed will need to keep a close eye on how it stacks up from a price/performance perspective moving forward.

It should be noted though that the company claims to offer large discounts on volume purchases.

Between those two options, ClearSpeed seems the most practical choice for customers looking to take advantage of cheaper x86 servers while showing dramatic floating point and performance per watt improvements.

[jcn1066] 25/11/2006 13:05

Arrived at SC06


The big news at SC06 is in the rise of the accelerators. Last year there were a few brave FPGA people, and some custom chip people. Now we have lots of accelerated projects: GPGPU, ClearSpeed, Cell, PeakStream, Celoxica, ....

What is interesting to me is how quickly the market has switched. I had expected it to take years to grow an interest. The interest here is a strong ground-swell, literally a grass-roots effort.

[jcn1066] 25/11/2006 12:57

AMD and Clearspeed team up in Top 500 supercomputer list



One of the most interesting facets of the new Top 500 list is the ninth slot in the Top 10 (see chart below), a Japanese system based on Opteron processors (via Sun) that now includes Clearspeed coprocessors.

The Clearspeed product is only an HCP coprocessor, so it doesn't need all of the additional graphics-specific hardware that even a more generalized "stream processor" GPU like NVIDIA's new G8800 has. Because the NVIDIA and ATI offerings are still chock full of specialized graphics hardware, they can't touch a much leaner, dedicated solution like Clearspeed in terms of performance per watt.

[jcn1066] 25/11/2006 12:52

Top Five Companies to watch in 2007



We picked ClearSpeed for several reasons: Technology, Messaging, Focus, and Management.

When partners like Intel applaud you for helping them deliver "breakthrough" capabilities, it's a pretty good sign that you are doing something right. ClearSpeed has an impressive portfolio when it comes to performance acceleration. It's what they do. And they do it really well.

Increasing performance without impacting the power requirements. Do you think the HPC community finds this interesting? Darn right they do.

And there's no confusion here. When you're looking for a company that really "gets it" in terms of messaging, we have to tip our hats to ClearSpeed. This is a company that knows what they do best and they tell their story well. Conversations with their employees and even their external PR agency are consistent and on point. It's actually quite impressive.

ClearSpeed's challenge in 2007 will be competing with commodity alternatives. The high-volume, lower-cost solutions that are being marketed to the same companies looking for improved floating point performance. ClearSpeed is well aware of this - and their marketing efforts are focusing on distancing the company from its commodity competitors.

In August of 2006, ClearSpeed added a key member to its management team - Stephen McKinnon - who joined ClearSpeed as Chief Operating Officer. McKinnon is an Intel veteran with more than 25 years of semiconductor industry experience. He's a pro. So now, add a solid management team to that powerful combination of technology, messaging and focus, and you have a very impressive combination that makes ClearSpeed a company worth watching. www.clearspeed.com.

[jcn1066] 25/11/2006 12:46

HPCwire's 2006 Readers' and Editors' Choice Awards


Readers' Choice Award Recipient:

ClearSpeed Technology
Visit ClearSpeed Technology

ClearSpeed Technology is a specialist semiconductor company focused on delivering double precision high performance coprocessors and boards to be used alongside general purpose processors in the world's most compute-intensive applications. ClearSpeed's advanced multithreaded array processing technology provides the ability to significantly accelerate data-intensive applications at extremely low power. Products include chips, boards, software tools, applications and support. ClearSpeed has offices in San Jose, California and Bristol, UK and has 84 patents granted and pending. For more information on ClearSpeed, visit www.clearspeed.com.

[jcn1066] 25/11/2006 12:43

IDC Meeting at SC06 Includes Five-Year Forecast



Willard presented highlights of IDC's January 2006 cluster end-user study, which found that the three primary buying criteria are price/performance, system throughput, and total cost of ownership. The top challenges are facilities issues (e.g., power and cooling), and system management capability. To expand on this, about equal numbers of the users buy new clusters or add more nodes. The mean number of nodes was 180, of CPUs/core 360, and of sockets 256. Nearly half of all sites (47 percent) used in-house codes, while 45 percent used third-party codes and 10 percent used open source codes.

[jcn1066] 25/11/2006 12:37

ClearSpeed Technology Named One of the "Top Five Companies to Watch in 2007" by HPCwire at SC06



ClearSpeed also received the HPCwire Readers' Choice award for the most significant new HPC hardware product for its Advance™ Accelerator boards


According to the official HPCwire announcement of the awards posted on the HPCwire Web site, the editors stated, "When partners like Intel applaud you for helping them deliver 'breakthrough' capabilities, it's a pretty good sign that you are doing something right. ClearSpeed has an impressive portfolio when it comes to performance acceleration. It's what they do. And they do it really well. Increasing performance without impacting the power requirements. Do you think the HPC community finds this interesting? Darn right they do."

[jcn1066] 25/11/2006 12:35

Understanding the Different Acceleration Technologies
by John L. Gustafson, CTO
High Performance Computing, ClearSpeed Technology Inc.



But the issue is no longer just speed. It's speed per watt, speed per square foot, and speed per cubic inch. Oh, and speed divided by total cost of ownership. This is where accelerators can sometimes be of great benefit compared to simply adding more general-purpose processors to a system. With a few exceptions, accelerators that are performing the function for which they were designed can far exceed the ratio of performance to power and space that you can achieve with a general-purpose processor.

[jcn1066] 25/11/2006 12:30

Clearspeed give supercomputer a 24% boost
by David Manners
Friday 17 November 2006

ClearSpeed Technology, the Bristol co-processor company which works with Intel and AMD to soup up their processors performance, has had its technology adopted by the world's ninth fastest supercomputer.

"Tokyo Tech GSIC worked with ClearSpeed to achieve a 9 TeraFLOPS increase in performance from our previous result, with very little addition to power requirements and no overhead in space," said Professor Satoshi Matsuoka of the Global Scientific Information and Computing Centre (GSIC) of Tokyo Institute of Technology (Tokyo Tech) which built the supercomputer which it calls TSUBAME.

The ClearSpeed accelerated result of 47 TeraFLOPS is a 24 per cent performance boost from the non-accelerated performance of 38 TeraFLOPS.

The increased performance is delivered with only a one percent increase in energy consumption. Tokyo Tech has estimated that the TSUBAME's architecture will save them up to a $1m per year in combined facilities and energy costs, amounting to a reduction of between 10 and 20 per cent of the system's total cost of ownership.

"From this point forward, mainstream HPTC systems will be architected by combining industry standard-platforms with purpose-designed acceleration technology," said Stephen McKinnon, ClearSpeed's COO, "ClearSpeed is paving the way for a new era of energy-efficient high performance and technical computing."

[tradx666] 21/11/2006 18:19

The bang-for-buck area just got tougher..

regards

T..


Ahead of next week's Supercomputing 2006 show
in Tampa, Fla., Intel Corporation revealed new products and technology
building blocks for system manufacturers, system builders and software
developers to more easily create high-performance supercomputers that
run faster and save space.


New Intel-based server products introduced this year are delivering
breakthroughs in technical computing. As a point of reference, 10 years
ago, a supercomputer with 9000 Intel® Pentium®
Pro processors took up 1,500 square feet, delivered 1.80 teraflops of
peak performance and consumed some 800,000 watts of power. Today, the
same level of performance can be achieved by a server cluster with 44
Quad-Core Intel® Xeon®
processor 5300 series, codenamed "Clovertown"
and due to ship shortly, while taking up just 16 square feet and
consuming less than 10,000 watts.1


"The latest Intel technology innovations have
enabled new computing models and form factors that were previously
unavailable for high-performance computing which, in turn, hastens new
scientific discoveries, product designs and simulations among many other
discoveries around the world," said Boyd
Davis, director of marketing, Intel's server
platform group. "Over the past year we have
seen the development of the first personal supercomputers,
double-density rack mount servers and the increase of the industry's
largest, most scalable systems with dual-core Intel Itanium processors
and Intel Xeon processors."


New Products


This month, Intel will launch the first industry-standard quad-core
server processor for dual processor systems, the aforementioned
Quad-Core Intel Xeon processor 5300 series. These processors will
deliver more than six times the performance over single core Intel Xeon
processors for high- performance computing applications based on the
Linpack benchmark.2


New Software Tools for Clustering


Today, Intel made available new software development tools to make it
easier to create, debug and optimize parallel applications for server
clusters. These new products -- Intel®
Cluster Toolkit 3.0 and Intel Cluster OpenMP for Intel compilers -- are
designed to satisfy developer demand for high-performance tools and to
facilitate the adoption of multi-core processor server clusters. The new
Intel Cluster Toolkit combines multiple products into one software
bundle that includes new releases of the Intel MPI Library, Intel Math
Kernel Library Cluster Edition, and Intel Trace Analyzer and Collector.
The Cluster OpenMP for Intel compilers is a new offering that extends
OpenMP to be applicable to distributed memory clusters, helping OpenMP
become a programming method that works well for dual-core and quad-core
processors as well as clusters. For more information, go to www.intel.com/software/products.


New Building Blocks Reduce Racks and Costs


Intel's dual- and quad-core technology is
making it easier for system builders to construct rack mount servers
that save space in a data center and deliver high performance and better
performance per watt across every price point.


For applications where cost is a key consideration, Intel has developed
the S3000PT -- a purpose-built, small form factor server board for
high-performance and high-density computing. This powerful,
energy-efficient board is based on the Dual-Core Intel®
Xeon processor 3000 series and is compatible with the quad-core
processor, codenamed "Kentsfield."
The small 5.9"x13"
form factor is well suited for personal clusters for individuals and
workgroups as well as high- density datacenters such as ISPs and Web
hosting. For more information go to www.intel.com/go/hpcservers.
Ciara Technologies is using the S3000PT in their Nexxus 4000 personal
cluster which is targeted at researchers and engineers. The cluster
delivers up to 340 gigaflops of performance in the size of a desktop PC
and starts at under $12,000. For more information on the Nexxus 4000 go
to www.vxtech.com.


Supermicro, Intel and Mellanox have developed the world's first
high-volume compact server board for high-end, dual processor servers.
This high-density, energy-efficient board, codenamed "Atoka," supports
the Quad-Core Intel Xeon processor 5300 series and the Dual-Core Intel
Xeon processor 5100 series. A key advantage of this new form factor is
the flexibility to configure twin boards with quad-core processors in a
1U configuration, which essentially quadruples (16 cores and 32 threads)
computing density. Atoka includes InfiniBand networking and dual gigabit
Ethernet to provide exceptional I/O capabilities, making it ideal for
the development of workgroup, departmental and enterprise clusters
targeted across a wide range of industry segments and application
requirements. For more information, visit www.supermicro.com/products.


Tyan, Intel, Mellanox and Microsoft have developed the next-generation
personal supercomputer, Typhoon 600, based on the Quad-Core Intel Xeon
processor 5300 series. Typhoon 600 is a turnkey system that delivers 256
gigaflops of performance and only requires a 110V/15 amp standard wall
outlet. The system is designed for engineers and scientists who need the
computational horsepower of a supercomputer, yet in the size of a PC.
The system starts at $15,000 and will generally be available early next
year from Tyan. For more information go to www.tyanpsc.com/products/products.html.


Intel, the world leader in silicon innovation, develops technologies,
products and initiatives to continually advance how people work and
live. Additional information about Intel is available at www.intel.com/pressroom.


Intel, Intel Xeon, Itanium and the Intel logo are trademarks or
registered trademarks of Intel Corporation or its subsidiaries in the
United States and other countries.


Other names and brands may be claimed as the property of others.


1Source: Intel. ASCI Option Red system that
Intel delivered to Sandia National Laboratories in 1996. 10,000-watt
system based on Quad-Core Intel Xeon processor system 5300 series at
2.66 GHz.


2Source: Intel. Intel S5000PAL with two
Quad-Core Intel Xeon Processors 5355, 8GB FBDIMM DDR2 5300, Red Hat
Linux AS4 update 4, LINPACK built with Intel MKL 9.0 reporting 64
gigaflops at 30K matrix size.


Ahead of next week's Supercomputing 2006 show in Tampa, Fla.,
Intel Corporation revealed new products and technology building blocks
for system manufacturers, system builders and software developers to
more easily create high-performance supercomputers that run faster and
save space.

¶ New Intel-based server products introduced this year are
delivering breakthroughs in technical computing. As a point of
reference, 10 years ago, a supercomputer with 9000 Intel(R) Pentium(R)
Pro processors took up 1,500 square feet, delivered 1.80 teraflops of
peak performance and consumed some 800,000 watts of power. Today, the
same level of performance can be achieved by a server cluster with 44
Quad-Core Intel(R) Xeon(R) processor 5300 series, codenamed
"Clovertown" and due to ship shortly, while taking up just 16 square
feet and consuming less than 10,000 watts.(1)

¶ "The latest Intel technology innovations have enabled new
computing models and form factors that were previously unavailable for
high-performance computing which, in turn, hastens new scientific
discoveries, product designs and simulations among many other
discoveries around the world," said Boyd Davis, director of marketing,
Intel's server platform group. "Over the past year we have seen the
development of the first personal supercomputers, double-density rack
mount servers and the increase of the industry's largest, most
scalable systems with dual-core Intel Itanium processors and Intel
Xeon processors."

¶ New Products

¶ This month, Intel will launch the first industry-standard
quad-core server processor for dual processor systems, the
aforementioned Quad-Core Intel Xeon processor 5300 series. These
processors will deliver more than six times the performance over
single core Intel Xeon processors for high- performance computing
applications based on the Linpack benchmark.(2)

¶ New Software Tools for Clustering

¶ Today, Intel made available new software development tools to make
it easier to create, debug and optimize parallel applications for
server clusters. These new products -- Intel(R) Cluster Toolkit 3.0
and Intel Cluster OpenMP for Intel compilers -- are designed to
satisfy developer demand for high-performance tools and to facilitate
the adoption of multi-core processor server clusters. The new Intel
Cluster Toolkit combines multiple products into one software bundle
that includes new releases of the Intel MPI Library, Intel Math Kernel
Library Cluster Edition, and Intel Trace Analyzer and Collector. The
Cluster OpenMP for Intel compilers is a new offering that extends
OpenMP to be applicable to distributed memory clusters, helping OpenMP
become a programming method that works well for dual-core and
quad-core processors as well as clusters. For more information, go to
www.intel.com/software/products.

¶ New Building Blocks Reduce Racks and Costs

¶ Intel's dual- and quad-core technology is making it easier for
system builders to construct rack mount servers that save space in a
data center and deliver high performance and better performance per
watt across every price point.

¶ For applications where cost is a key consideration, Intel has
developed the S3000PT -- a purpose-built, small form factor server
board for high-performance and high-density computing. This powerful,
energy-efficient board is based on the Dual-Core Intel(R) Xeon
processor 3000 series and is compatible with the quad-core processor,
codenamed "Kentsfield." The small 5.9"x13" form factor is well suited
for personal clusters for individuals and workgroups as well as high-
density datacenters such as ISPs and Web hosting. For more information
go to www.intel.com/go/hpcservers. Ciara Technologies is using the
S3000PT in their Nexxus 4000 personal cluster which is targeted at
researchers and engineers. The cluster delivers up to 340 gigaflops of
performance in the size of a desktop PC and starts at under $12,000.
For more information on the Nexxus 4000 go to www.vxtech.com.

¶ Supermicro, Intel and Mellanox have developed the world's first
high-volume compact server board for high-end, dual processor servers.
This high-density, energy-efficient board, codenamed "Atoka," supports
the Quad-Core Intel Xeon processor 5300 series and the Dual-Core Intel
Xeon processor 5100 series. A key advantage of this new form factor is
the flexibility to configure twin boards with quad-core processors in
a 1U configuration, which essentially quadruples (16 cores and 32
threads) computing density. Atoka includes InfiniBand networking and
dual gigabit Ethernet to provide exceptional I/O capabilities, making
it ideal for the development of workgroup, departmental and enterprise
clusters targeted across a wide range of industry segments and
application requirements. For more information, visit
www.supermicro.com/products.

¶ Tyan, Intel, Mellanox and Microsoft have developed the
next-generation personal supercomputer, Typhoon 600, based on the
Quad-Core Intel Xeon processor 5300 series. Typhoon 600 is a turnkey
system that delivers 256 gigaflops of performance and only requires a
110V/15 amp standard wall outlet. The system is designed for engineers
and scientists who need the computational horsepower of a
supercomputer, yet in the size of a PC. The system starts at $15,000
and will generally be available early next year from Tyan. For more
information go to www.tyanpsc.com/products/products.html.

¶ Intel, the world leader in silicon innovation, develops
technologies, products and initiatives to continually advance how
people work and live. Additional information about Intel is available
at www.intel.com/pressroom.

¶ Intel, Intel Xeon, Itanium and the Intel logo are trademarks or
registered trademarks of Intel Corporation or its subsidiaries in the
United States and other countries.

¶ Other names and brands may be claimed as the property of others.

¶ (1) Source: Intel. ASCI Option Red system that Intel delivered to
Sandia National Laboratories in 1996. 10,000-watt system based on
Quad-Core Intel Xeon processor system 5300 series at 2.66 GHz.

¶ (2) Source: Intel. Intel S5000PAL with two Quad-Core Intel Xeon
Processors 5355, 8GB FBDIMM DDR2 5300, Red Hat Linux AS4 update 4,
LINPACK built with Intel MKL 9.0 reporting 64 gigaflops at 30K matrix
size.

[tradx666] 14/11/2006 17:53

mdrans1,

be careful what you wish for! I see nothing that will reverse the trend at the moment, but plenty of reasons why it may well accelerate.

regards

T..

[badday] 14/11/2006 16:26

New interview, not too sure if this is direct link. On a different pc to normal:

[mdrans1] 14/11/2006 13:36

Cheer up Trad, we might get some bad news.

[tradx666] 14/11/2006 11:17

DAFAD,

I see a small uptick against the trend today, perhaps it's attendees at SC getting excited again!?

an interesting article; I'm slightly amused by the reference to Stephen McKinnon, as I've known him for a few years, and one can only hope that he can make the transition. As for the flattery from Intel, you shouldn't read too much into it, they did and said the same about infiniband and myricom in their time. The only thing that matters now is revenue growth, and of course keeping a watchful eye on nvidia and what they intend to do next, as always the battle is bang for bucks...

regards

T..