Deep Blitz: taking on Deep Blue with a PC

"Keith" wrote in message
news
On Sun, 29 Jan 2006 07:29:05 -0800, YKhan wrote:

Keith wrote:
Has there been another challenge?

No, they're just trying to see if they can match or exceed Deep Blue
nowadays with commodity hardware with less cost. Deep Blue cost $100M.

It doesn't take long to eat through $100M on a one-off design. That's
only 40-50 warm.body-years.

I gather that once grid-capable chess programs are available, the next
generation of chess machines will be going in the same direction as
supercomputers have gone towards superclusters.

Hmm, maybe that's the real purpose behind the Cell. ;-)

--
Keith

100M/50 = 2M/PY. You guys must make a lot more than us E&TS guys.
damn, I knew I deserved a raise.

Del Cecchi wrote:
100M/50 = 2M/PY. You guys must make a lot more than us E&TS guys.
damn, I knew I deserved a raise.

Don't forget managers and execs count as warm bodies too.

Yousuf Khan

--
There is no failure, only delayed success

Del Cecchi wrote:
"Keith" wrote in message
news
On Sun, 29 Jan 2006 07:29:05 -0800, YKhan wrote:

Keith wrote:
Has there been another challenge?
No, they're just trying to see if they can match or exceed Deep Blue
nowadays with commodity hardware with less cost. Deep Blue cost $100M.
It doesn't take long to eat through $100M on a one-off design. That's
only 40-50 warm.body-years.

I gather that once grid-capable chess programs are available, the next
generation of chess machines will be going in the same direction as
supercomputers have gone towards superclusters.
Hmm, maybe that's the real purpose behind the Cell. ;-)

--
Keith

100M/50 = 2M/PY. You guys must make a lot more than us E&TS guys.
damn, I knew I deserved a raise.


I've had a couple of jobs working for PhD's doing medical
research. By the time you factor in equipment costs, office
space, clerical support, technical support, etc., for every
dollar of PhD salary there was 10+ dollars spent on other things.

On Sun, 29 Jan 2006 16:32:03 -0600, Del Cecchi wrote:


"Keith" wrote in message
news
On Sun, 29 Jan 2006 07:29:05 -0800, YKhan wrote:

Keith wrote:
Has there been another challenge?

No, they're just trying to see if they can match or exceed Deep Blue
nowadays with commodity hardware with less cost. Deep Blue cost $100M.

It doesn't take long to eat through $100M on a one-off design. That's
only 40-50 warm.body-years.

I gather that once grid-capable chess programs are available, the next
generation of chess machines will be going in the same direction as
supercomputers have gone towards superclusters.

Hmm, maybe that's the real purpose behind the Cell. ;-)

--
Keith

100M/50 = 2M/PY. You guys must make a lot more than us E&TS guys.
damn, I knew I deserved a raise.

Ok, it seems I can't get the decades right. Maybe that's why I made
more a decade ago than today. ...and *WE* make money? 72% this year. :-(

Tis' a damned good thin that AMD is making someone money! ;-)
--
Keith

In article . com,
YKhan wrote:
Del Cecchi wrote:
Has Deep Blue been updated? When was the hardware for deep blue built?

1997. According to the article it consisted of 32 Power2 cores, and
about 512 specialized "chess processors", and cost IBM $100M. They're
seeing what they can now manage with off-the-shelf hardware.

Yousuf Khan



IBM's way beyond that, now.

http://www.top500.org/lists/2005/11/basic

131,072 processors.

http://www.research.ibm.com/bluegene/


Late last year I saw a presentation by a couple of the engineers on
that project. Neat.

--
a d y k e s @ p a n i x . c o m

Don't blame me. I voted for Gore.

"Al Dykes" wrote in message
...
In article . com,
YKhan wrote:
Del Cecchi wrote:
Has Deep Blue been updated? When was the hardware for deep blue
built?

1997. According to the article it consisted of 32 Power2 cores, and
about 512 specialized "chess processors", and cost IBM $100M. They're
seeing what they can now manage with off-the-shelf hardware.

Yousuf Khan



IBM's way beyond that, now.

http://www.top500.org/lists/2005/11/basic

131,072 processors.

http://www.research.ibm.com/bluegene/


Late last year I saw a presentation by a couple of the engineers on
that project. Neat.

--
a d y k e s @ p a n i x . c o m

Yeah. I heard the hardware was done by some farmers out on the tundra in
flyover land. Will wonders never cease?.

:-)

Yeah. I heard the hardware was done by some farmers out on the tundra in
flyover land. Will wonders never cease?.


Really? I heard the farmers decided that they should let the cows do
it...

DK

In article . com,
says...
Yeah. I heard the hardware was done by some farmers out on the tundra in
flyover land. Will wonders never cease?.


Really? I heard the farmers decided that they should let the cows do
it...

Wrong farm. The cows are in Vermont. The tundra farmers heard
corn. ;-)

--
Keith

Al Dykes wrote:
IBM's way beyond that, now.

That's a given. However, currently commodity hardware in superclusters
are providing most of the supercomputing. I don't think any of the
chess programs available out there are able work over a grid, although
they might work over an SMP. I think the next generation of chess
programs are going to have to work over superclusters, just like
supercomputers have evolved into.

Yousuf Khan

"Keith Williams" wrote in message
T...
In article . com,
says...
Yeah. I heard the hardware was done by some farmers out on the
tundra in
flyover land. Will wonders never cease?.


Really? I heard the farmers decided that they should let the cows do
it...

Wrong farm. The cows are in Vermont. The tundra farmers heard
corn. ;-)

--
Keith

What did the corn say?

Stick it in your ear?