Clearspeed Technology Share Discussion Threads

[speedieruk] 15/3/2006 21:20

Tradx666 - More outlets are reporting the news - e.g.



One of those snippets of information which could have little or no effect, but on the other hand, as you rightly point out, it could transform the company.

[tradx666] 15/3/2006 19:03

SpeedierUK,

Frankly, I'm surprised that this has driven a bigger spike, or perhaps the market simply hasn't understood the ramifications of this?

regards

T..

[speedieruk] 15/3/2006 17:53

Tradx666 - you're right there, and a similar link here:



I wonder whether this came about because of previous collaboration.

Anyway, as you say, could lead to good things.

[tradx666] 14/3/2006 11:18

this from electronics weekly...the implications of this are simply huge!!!


Clearspeed plans AMD chip link up

By David Manners [e-mail]


Clearspeed,the Bristol parallel processor company, is in talks with AMD about developing a closely-coupled co-processor to soup up the AMD multi-core family of x86 processors.

"Clearspeed is talking to AMD about producing a closely coupled co-processor," said Jeff Underhill, business development manager for 64-bit embedded applications at AMD.

Behind the talks are worries that it is not achieving sufficient performance out of its multi-core processors. Stoking those concerns was the stunning performance of the Sony/Toshiba/IBM microprocessor Cell.

"Cell was a wake up call. Absolutely, things are underway at AMD as a result in looking at future products," said Underhill.

The immediate future product for AMD is QuadCore, due out next year, a four-core
microprocessor delivering 20Gflops which doubles the performance of AMD's current two-core product, DualCore, which delivers 10Gflops.

However, the Quadcore/ DualCore level of performance lags behind the blazing speed of Cell which uses8 synergistic processors around a PowerPC core to deliver a quarter of a Teraflops or 256bn flops.

Of course, x86 is Cisc, and Cell is Risc and so can run faster, nonetheless the difference in performance between Cell and QuadCore is such that AMD has recognised that it needs to crank up the performance of its microprocessors.

Clearspeed's experience of producing ultra-high-speed accelerator chips to boost computing power, while maximising power per Watt, is the answer. Clearspeed's CSX600 chip has 96 processor cores, runs at 250MHz, dissipates 10W, and adds 25Gflops to a computer's performance.

The problems in implementing multicore processors in practice, according to Chris Rowen, CEO of Tensilica, which automates microprocessor generation, are: "A need for new tools and software for application-centric energy management and processor programming; and improvements for low-voltage and capacitance."

[mdrans1] 10/3/2006 10:04

Small player is fast enough for supercomputer
by Alan Cane
Published: March 7 2006 18:54 | Last updated: March 7 2006 18:54

IBM must have thought it was on to a shoo-in. Late last year the creator of Blue Gene, the world's fastest supercomputer, was ­confidently expecting a positive response to its bid to build a machine that would be almost as powerful for the Tokyo Institute of Technology.

But instead, the institute opted for a consortium led by NEC, the Japanese computer group. Alongside big names such as Sun Microsystems, US chip-maker AMD and US networking specialist Voltaire Infiniband, the grouping includes UK-based Clear-Speed Technology, which is providing $2m worth of specialised microprocessor technology which will give the supercomputer its blinding speed. Ironically, IBM makes ClearSpeed's chips on a contract basis in its US fabrication plants.

The Japanese order represents a huge step forward for the company. Alex Jarvis, research analyst with KBC Peel Hunt, the company's broker, says the contract is a "superb first win" for ClearSpeed and vindication of its decision to stick with its business strategy during recent years when the semiconductor industry was going through its worst recession.

The management bet the company, which listed on Aim, London's junior market, in July 2004, on a belief that a market would emerge for its esoteric product.

Tom Beese, chief executive, a very experienced former Hewlett-Packard executive, recalls he told the investors and the team: "Classically this will feel like the worst of times to do this, but if we are right in believing that this market will emerge – and if we are right in how to address the needs of that market – then this is potentially the best of times to do this. We might emerge with significant first-mover advantage." He adds: "And that is how it has played out".

Brave words for the dark days of 2002, which is when the company was founded – although its roots go back to Inmos, the UK-based ­inventor of the transputer, through Division, the virtual reality group, and Pixel­fusion, a high-end graphics company. All were specialists in parallel processing – computer systems that can handle two or more streams of computation simul­taneously. This is inte- gral to high-performance ­computing.

ClearSpeed's product family is described as "double-precision, floating-point accelerator boards", jargon for printed circuit boards and chips that can manipulate very long numbers with very high accuracy.

They are not the main micro­processors but co- processors that relieve the central processing unit of the heaviest arithmetical load and permit computational speeds that have only ever been beaten by Blue Gene.

John Fowler, chief software technology officer for Sun Microsystems, which is developing the Tokyo supercomputer, says: "The critical thing that ClearSpeed adds is the ability for Sun to meet efficiently a customer's most intense floating-point require­ments in a way that cannot be done economically with central processors alone."

ClearSpeed has no plans to move its headquarters from Bristol, three hours' drive west of London, although its US operations in San Jose, California, are vital for staying in touch with the all- important US supercomputer industry. With revenues of only £146,000 ($256,000) at the half-year point last June, the $2m japanese deal represents a breakthrough.

Mr Beese points out how quickly the kinds of projects that ClearSpeed wants to tackle can develop. At the start of August last year, he had no information about the Tokyo contract. Within two weeks, he was approached by possible bidders who realised conventional microprocessors would not meet the specification. "Sun engaged with us less than 10 days before the bids had to be in and the consortium was still able to win."

Supercomputing, however, will remain a small market. Mr Beese believes the future lies in linking off-the-shelf microprocessors to accelerator boards such as the ClearSpeed CSX 600 which will open high-powered, low-cost computing to a wide range of industries – aerospace, white goods, automotive, oil and gas.

"We saw that was likely to happen. The collapsing of the cost of entry is producing a whole new wave of demand," he says.



Find this article at:


PS I think it was also in the FT (paper) on Wednesday.

[taylor20] 06/3/2006 10:35

Crikey, no pleasing some people! ;-)

I thought the point was that this is the first time they have started shipping boards in volume to any customer [that they can tell us about of course!](excluding prototypes, A & B model boards), i.e. this is very much a finished product with all issues ironed out.

This is one of the company's major milestones, hence the announcement.

What annoys me is that the share price always start to move up about a week or two before the RNS is released.

[tradx666] 06/3/2006 09:55

Cerrito,

To be frank, the fact that they seem to need to announce the shipment of a relatively small number of boards is to me a worrying sign, as it suggests that they have little else win news to crow about!

I hope I'm wrong, we desperately need to see a British company succeed in the this space.

regards

T..

[tradx666] 01/3/2006 16:39

broadwood,

I've been watching them carefully, and have to say that their strategy seems to have some big holes in it now...but I guess we will all know soon enough..

regards

T..

[broadwood] 01/3/2006 16:12

I was beginning to get worried about this. They seem to talk a good game - but need to transfer it into firm orders,
Still, comforting move up today.

[mdrans1] 22/2/2006 15:49

and another one :-

February 21, 2006 11:00 AM London Timezone
ClearSpeed Expands U.S. Office and Adds Training Centre; Company moves to a larger office space as the number of employees has doubled in the past year

"The San Jose office is a strategically important location for ClearSpeed during this time of growth," said Tom Beese, ClearSpeed chief executive officer. "With the close of a new round of funding and our planned shipments to major customers, the expanded office will enable us to maintain a high level quality in our ongoing training and support of customers."

[scrutable] 21/2/2006 16:23

mdrans
That was a public service rendered. But I was delighted to see that it has also gone up on the company web site.

The investors also need to know when the boards start going out in production quantities, and when the japanese order is scheduled for delivery

[cerrito] 17/2/2006 19:34

mdrans
Thanks for posting that article
Taking into account the price weakness been looking at topping up, but the lack of news is rather spooky(even though I take the point that it has been emphasized on this board that we need patience) and I have decided to pass.
My hunch is that in the next 2/3 weeks we will get a preclose season trading update with hopefully tangible news.

[mdrans1] 16/2/2006 13:32

February 13, 2006 08:00 AM London Timezone
ClearSpeed Partners with Nissho Electronics to Bring the World's Fastest and Most Power-Efficient Multi-Core Co-Processors and Accelerator Boards to Japan
SAN JOSE, Calif.--(BUSINESS WIRE)--Feb. 13, 2006--
Nissho Electronics, One of Japan's Top Distributors, to Resell ClearSpeed's CSX600 Co-Processors and Advance(TM) Accelerator Boards

ClearSpeed Technology (LSE:CSD), a developer of high-performance, low-power, programmable co-processor solutions, today announced a partnership with Nissho Electronics Corporation to deliver the company's chips, boards and software tools to Japan. Under the terms of agreement, Nissho will market and distribute ClearSpeed's CSX600 co-processors and Advance(TM) accelerator boards to Japanese high-performance computing (HPC) customers immediately. ..........

See:-



Note "distribute ......immediately"

[mdrans1] 16/2/2006 13:12

Prelims due on 21/04/06, ( based on last year)

[shoee62] 16/2/2006 09:39

Seems to have run out of steam for the moment. All that volume at the beginning of Jan made it look like a test of the top was likely.

When are the next results out ?

[broadwood] 02/2/2006 14:24

Fidelity have just bought 7% of the Company. Thats good enough for me.

I'm in - at 281p.

[supernumerary] 20/1/2006 16:15

tradx

I agree there's a lack of quality, but surely that's the point of AIM? Most of the warning signs are there at listing, and if people want to take a punt that's their right - better for everybody concerned than spending it on lottery tickets or online poker or whatever the latest gamble is. It only bothers me when there's active deception ie lies are told, or passive deception ie the full truth is not told, and the lack of transparency on trades falls into the second category imo.

[tradx666] 20/1/2006 15:57

supernumeracy,

The basic issue on the 'attacks' on many small companies is in my opinion a distinct lack of quality, and a headlong rush for the founding investors to capitalise on their original investment. The huge growth in aim companies is I believe characterised by many coming to market at too early a stage and with very poor corporate advisers, who take the fees at listing time, and about as good as the proverbial pork pie at a Jewish wedding thereafter!

If only they brought in new rules in working capital/profitability/length of being etc, then many of these targets for shorting wouldn't be in the firing line, and yes that includes 3DM.

And I do agree that complete and utter transparency and market disclosure must be made when short selling, and it is to our detriment that we refuse to follow the US markets in this respect...

regards

T..

[supernumerary] 20/1/2006 00:32

I think the answer (if there is one at all) is not to criminalise shorting, but to make it a lot more transparent. At the moment, no ordinary investor has a clue about either shorts or derivatives being taken out on a share, so an enormous amount of action is simply invisible. If ALL transactions were declared in the same way that ordinary trades are, the playing field would surely be a lot more level?

[scrutable] 20/1/2006 00:09

Tradex666
it makes an interesting debate. Do you really want to prevent companies with good prospects coming to the market to raise cash to achieve their aims?
Surely this is the prime function of a stock exchange?

And there's another angle...Do you support institutions persistently short selling ( or just selling) with inside knowledge of a placing they have been canvassed to support. With the typical discount for the placing it is a licence to print money, to take up an allocation at the lower price level at the expense of the PI.

I agree that it's great discipline against an inept management, but others are badly hurt as well.

And what good has it done anyone at 3DM driving the compoany into the hands of the loan sharks?

[scrutable] 19/1/2006 23:57

supernumerary
thanks -of course - since corrected.

[tradx666] 19/1/2006 14:19

Scrutable,

I will look and I will be a subscriber soon - I have just been far too lazy to do it!

Interestingly, and slightly off-topic, what do you make of AFN's purchase another similar led service to your own? Although, in ED's case, the company seems to pay the costs of the analysts reports.

Sadly, I have to disagree in respect of supporting anything that would stop the markets from being artificially bound against (what I deem to be) normal commercial pressures.

If a small business is susceptible to being a 'short' target, it is normally because of solid reasons, be they the lack of maturity in their business model, the incompetence of the management, or the business as attacked a highly optimistic rating that the underlying fundamentals cannot support.

For those that are vulnerable in any of these respects, I see no reason to protect them against these realities, typically 'shorting' of a company will not break it, and may often be the making of it.

To me, protecting listed companies against this possibility would almost be a licence for them to under-perform, over-promise and is a charter for mediocrity. I for one prefer to see the weak weeded out sooner rather than later.

This may well be seen as overly harsh, but I have seen the other side of this equation through 20 years of successful business building and management, and sooner or later, weak management, or weak business models cost other businesses and individual investors dearly.

Perhaps the question that needs to be more properly answered first is this; How do we stop early stage, immature companies being listed at all - or if they are, some sort of Wealth warning scale that should be assessed by the listing authorities?!

regards

T..