Cray Supercomputer?

Quote from prt_systems:

I think the thread was a joke .... Very few things in the world of finance would require the old style crays as you correctly point out .....


Actually the thread wasn't a joke, but some of the posters thought it was.

I was bascially trying to find out if large Hedge Funds run their algorithms on these type of computers?

The Cray Supercomputer is used to model and predict global climates and weather patterns into the future, why couldn't it be used to predict stock prices or market action into the future? The Cray could look back through countless amounts of historical data and determine a "what if" scenario. They say that history repeats itself.

I know that these computers are very expensive, but then again some of these large Hedge funds could easily buy one to run all their many diffferent algorithms.

Does anyone else see any use for the Cray other than crunching numbers?
 
Quote from skepticaltrader:

Actually the thread wasn't a joke, but some of the posters thought it was.

I was bascially trying to find out if large Hedge Funds run their algorithms on these type of computers?

The Cray Supercomputer is used to model and predict global climates and weather patterns into the future, why couldn't it be used to predict stock prices or market action into the future? The Cray could look back through countless amounts of historical data and determine a "what if" scenario. They say that history repeats itself.

I know that these computers are very expensive, but then again some of these large Hedge funds could easily buy one to run all their many diffferent algorithms.

Does anyone else see any use for the Cray other than crunching numbers?

Most - but not all - finance problems can be run on parallel computing clusters at lower cost than on a cray - hence the reason crays are far less popular today than 20 years ago. Some problems still require its processor architecture and for those they are still invaluable.
 
Quote from skepticaltrader:

Actually the thread wasn't a joke, but some of the posters thought it was.

I was bascially trying to find out if large Hedge Funds run their algorithms on these type of computers?

The Cray Supercomputer is used to model and predict global climates and weather patterns into the future, why couldn't it be used to predict stock prices or market action into the future? The Cray could look back through countless amounts of historical data and determine a "what if" scenario. They say that history repeats itself.

I know that these computers are very expensive, but then again some of these large Hedge funds could easily buy one to run all their many diffferent algorithms.

Does anyone else see any use for the Cray other than crunching numbers?

most hedge funds buy dell - like everyone else.

:D

personally, i used to have access to a cray. it was great for surfing porn on the internet - it could find the biggest tits in no time at all. other than that, i really couldnt see the point. who cares what the weather is when theres porn :)
 
Quote from FredBloggs:

most hedge funds buy dell - like everyone else.

Are these Hedge Funds buying Dell PC's like the rest of us? If so they are seriously underpowered.

When you have 'deep pockets' why would you mess around with a whimpy Dell?
 
Quote from skepticaltrader:

Nice... Where do you get this funny stuff?
Vague recollections of conversations/texts from years gone by (when I was a grad student), and of course Google! :)
 
Quote from skepticaltrader:

Quote from FredBloggs:

most hedge funds buy dell - like everyone else.

Are these Hedge Funds buying Dell PC's like the rest of us? If so they are seriously underpowered.

When you have 'deep pockets' why would you mess around with a whimpy Dell?

why would a dell be underpowered?

for wintel solutions, youre able to cram as much in there as anything else.

ok-ok i know what you are talking about isnt the desk top.

the big multi-processor mainframe based systems tend to be used for large batch jobs still - and i think trade matching (exchange engines)

youre asking about modeling applications that fire off trades though (i think). i dont think that requires much processing power by todays standards - not enough to merit any mainframe. remember that simple is always the best approach. they may be looking at a couple of correlated markets outside the complex - and the depth, but thats it.

remember though that a lot of institutions would rather trade in the otc market where possible rather than on exchanges that are too transparent for their benefit. an otc market will generate less data (for processing) than an exchange auction market.

most fund managers ive met are in their 50's and spend most of the day at lunch with customers, and potential ceo's etc of investment companies. they tend to be fundamental old school types - they would run from a computer pdq.
 
Back
Top