31 Mar
2013
31 Mar
'13
2:14 p.m.
Fyi.
http://arstechnica.com/information-technology/2013/03/worlds-fastest-superc…
--
Stephane
FreeDonne <http://www.freedonne.org> Join FreeDonne - Rejoignez FreeDonne.
31 Mar
31 Mar
2:48 p.m.
On 31-mrt-2013 23:14, St?phane Tsacas wrote:
The power bill must've also been in the top 500 of the world...
- MG
2:56 p.m.
On 03/31/2013 05:48 PM, MG wrote:
The power bill must've also been in the top 500 of the world...
No doubt. But that's what power is for. =)
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
5:33 p.m.
Built, used for four years, and now DISMANTLED?
What is wrong with their fundamental architecture?
Can't they design one where more could be added to keep up with needs, and
upgrade incrementally and components while keeping the overall machine?
Are there finite limits on how much space can be created under the Fort?
What is it being replaced with?
--
Grumpy Ol' Fred cisin at xenosoft.com
5:40 p.m.
On 03/31/2013 08:33 PM, Fred Cisin wrote:
Built, used for four years, and now DISMANTLED?
Remember, this is America. If it's old, it's bad, and if it's not new,
it's old!
What is wrong with their fundamental architecture?
Contracts, and suits.
Can't they design one where more could be added to
keep up with needs, and
upgrade incrementally and components while keeping the overall machine?
Of course. This is IBM. They can do damn near anything these days, if
technology is the only limiting factor.
Are there finite limits on how much space can be
created under the Fort?
What is it being replaced with?
13,158 iPads?
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
5:58 p.m.
> Are there finite limits on how much space can be
created under the Fort?
hyperspace?
> What is it being replaced with?
On Sun, 31 Mar 2013, Dave McGuire wrote:
13,158 iPads?
That's less than one and a
quarter kilometers! x 241.2mm x 185.7mm
How much power would they draw?
6:15 p.m.
On 03/31/2013 08:58 PM, Fred Cisin wrote:
On Sun, 31 Mar 2013, Dave McGuire wrote:
http://gigaom.com/2012/06/21/so-how-much-does-it-cost-to-charge-an-ipad-eve…
"The analysis shows that each model of the iPad consumes less than 12 kWh
of elelctricity over the course of a year, based on a full charge every other
day."
...A LOT.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
13,158 iPads?
That's less than one
and a quarter kilometers! x 241.2mm x 185.7mm
How much power would they draw?
6:31 p.m.
>> 13,158 iPads?
> That's less than one and a quarter kilometers! x 241.2mm x 185.7mm
> How much power would they draw?
On Sun, 31 Mar 2013, Dave McGuire wrote:
http://gigaom.com/2012/06/21/so-how-much-does-it-cost-to-charge-an-ipad-eve…
"The analysis shows that each model of the iPad consumes less than 12
kWh of elelctricity over the course of a year, based on a full charge
every other day."
and that is probably based on "normal usage", not 24x7
...A LOT.
6:36 p.m.
On 03/31/2013 09:31 PM, Fred Cisin wrote:
Probably.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
>>
13,158 iPads?
> That's less than one and a quarter kilometers! x 241.2mm x 185.7mm
> How much power would they draw?
On Sun, 31 Mar 2013, Dave McGuire wrote:
http://gigaom.com/2012/06/21/so-how-much-does-it-cost-to-charge-an-ipad-eve…
"The analysis shows that each model of the iPad consumes less than 12
kWh of elelctricity over the course of a year, based on a full charge
every other day."
and that is probably based on "normal usage", not 24x7 
6:37 p.m.
On 3/31/13 7:15 PM, Dave McGuire wrote:
On 03/31/2013 08:58 PM, Fred Cisin wrote:
Of course their retiring super box probably takes 12 kwh in a few
minutes. :D
It's all relative I guess... Heck, it keeps my 6.5kw generator busy,
powering my few machines and a couple of lights when the power goes down
here.
--
Dave Land
Land Computer Service
Check out my site at http://www.landcomp.net
On Sun, 31 Mar 2013, Dave McGuire wrote:
http://gigaom.com/2012/06/21/so-how-much-does-it-cost-to-charge-an-ipad-eve…
"The analysis shows that each model of the iPad consumes less than 12 kWh
of elelctricity over the course of a year, based on a full charge every other
day."
...A LOT.
-Dave
13,158 iPads?
That's less than one
and a quarter kilometers! x 241.2mm x 185.7mm
How much power would they draw? 
1 Apr
1 Apr
11:14 a.m.
New subject: World^Òs top supercomputer from ^Ñ09 is now obsolete, will be dismantled
On 03/31/2013 08:58 PM, Fred Cisin wrote:
Assuming a year is 365 days, thats 8760 hours. And if 12kWh is the total
energy consumed in those 8760 hours, it gives an average power
consumption of about 1.37W. So your 13158 of them would have an average
power consumption of about 18kW. Big, yes, but not ridiculous.
But... I would gues that 12kWh figure is for the 'norma' use fo the Ipad,
not usein the processor flat-out all the time. Against that, I guess
some of the power goes to the RF communications stuff and even more to
the display and backpligth, which wouldn't bee needed in this application.
-tony
On Sun, 31 Mar 2013, Dave McGuire wrote:
http://gigaom.com/2012/06/21/so-how-much-does-it-cost-to-charge-an-ipad-eve…
"The analysis shows that each model of the iPad consumes less than 12 kWh
of elelctricity over the course of a year, based on a full charge every other
day."
...A LOT. 13,158 iPads?
That's less than one
and a quarter kilometers! x 241.2mm x 185.7mm
How much power would they draw?
6:32 p.m.
New subject: World^�s top supercomputer from ^�09 is now obsolete, will be dismantled
> >> 13,158 iPads?
> > That's less than one and a quarter kilometers! x 241.2mm x 185.7mm
> > How much power would they draw?
> http://gigaom.com/2012/06/21/so-how-much-does-it-cost-to-charge-an-ipad-eve…
> "The analysis shows that each model of the iPad consumes less than 12 kWh
> of elelctricity over the course of a year, based on a full charge every other
> day."
> ...A LOT.
On Mon, 1 Apr 2013, Tony Duell wrote:
Assuming a year is 365 days, thats 8760 hours. And if
12kWh is the total
energy consumed in those 8760 hours, it gives an average power
consumption of about 1.37W. So your 13158 of them would have an average
power consumption of about 18kW. Big, yes, but not ridiculous.
It helps to look at the numbers. 1.37W ??!?
REALLY??!?
But... I would gues that 12kWh figure is for the
'norma' use fo the Ipad,
not usein the processor flat-out all the time. Against that, I guess
some of the power goes to the RF communications stuff and even more to
the display and backpligth, which wouldn't bee needed in this application.
AND that would require that it gets 48 hours of use from each charge!
("full charge every other day") So, it is presumably assuming a
"reasonable" usage of prob'ly no more than 8 hours use per day.
Does anybody know what the actual POWER CONSUMPTION of the device is, not
"what will my electric bill be if I charge it every other day?"
Can it RUN off of a wall-wart? Would that not be a little more efficient
than running down the batteries and then recharging them?
It might even be possible to design a more efficient multi-output power
supply, rather than a few miles of outlet strips with wall-warts.
However, if you were to run it off batteries, and charge them every other
day, it could be PORTABLE! How big a truck?
--
Grumpy Ol' Fred cisin at xenosoft.com
31 Mar
31 Mar
6:07 p.m.
? Contracts, and suits.
Bingo!
________________________________
From: Dave McGuire <mcguire at neurotica.com>
To: On-Topic and Off-Topic Posts <cctalk at classiccmp.org>
Sent: Sunday, March 31, 2013 8:40 PM
Subject: Re: World?s top supercomputer from ?09 is now obsolete, will be dismantled
On 03/31/2013 08:33 PM, Fred Cisin wrote:
Built, used for four years, and now DISMANTLED?
? Remember, this is America.? If it's old, it's bad, and if it's not new,
it's old!
What is wrong with their fundamental architecture?
? Contracts, and suits.
Can't they design one where more could be added to
keep up with needs, and
upgrade incrementally and components while keeping the overall machine?
? Of course.? This is IBM.? They can do damn near anything these days, if
technology is the only limiting factor.
Are there finite limits on how much space can be
created under the Fort?
What is it being replaced with?
? 13,158 iPads?
? ? ? ? ? ? ? -Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
1 Apr
1 Apr
3:14 a.m.
New subject: Re[2]: World’s top supercomputer from ‘09 is now obsolete, will be dismantled
------ Original Message ------
From: "Dave McGuire" <mcguire at neurotica.com>
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk at classiccmp.org>
Sent: 01/04/2013 01:40:51
Subject: Re: World?s top supercomputer from ?09 is now obsolete, will be
dismantled
I believe that the interconnect is the limiting factor....
On 03/31/2013 08:33 PM, Fred Cisin wrote:
I think more bean counters. You can get a newer faster box that uses
much less power and so you actually can re-coup the investment in a few
years. Looking at that link I see that box needs 2345Kw to do just over
1.0 Pet Flops. The UK's fastest computer, which is slightly faster at
1.2 Peta Flops only uses around 575kw so only 25% of the power...
http://www.stfc.ac.uk/News+and+Events/39207.aspx
http://www.top500.org/system/177723
http://www-01.ibm.com/software/success/cssdb.nsf/CS/STRD-8YREBG?OpenDocumen…
http://en.wikipedia.org/wiki/Blue_Gene
(More links to IBM research for those that get seriously excited by fast
computers)
Built, used for four years, and now DISMANTLED?
Remember, this is America. If it's old, it's bad, and if it's not
new,
it's old!
What is wrong with their fundamental
architecture?
Contracts, and suits.
Can't
they design one where more could be added to keep up with
needs, and
upgrade incrementally and components while keeping the overall
machine?
Of course. This is IBM. They can do damn near
anything these days, if
technology is the only limiting factor.
Na they have upgraded from PlayStation II to Playstation III chips....
Are there finite limits on how much space can be
created under the
Fort?
What is it being replaced with?
13,158 iPads?
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Dave
G4UGM
31 Mar
31 Mar
5:56 p.m.
On Mar 31, 2013 5:34 PM, "Fred Cisin" <cisin at xenosoft.com> wrote:
What is it being replaced with?
It's being replaced by the WOPR.
Apparently its replacement has been in service for a while:
http://www.lanl.gov/orgs/hpc/cielo/
8:03 p.m.
On Mar 31, 2013, at 5:33 PM, Fred Cisin <cisin at xenosoft.com> wrote:
Built, used for four years, and now DISMANTLED?
What is wrong with their fundamental architecture?
Can't they design one where more could be added to keep up with needs, and
upgrade incrementally and components while keeping the overall machine?
There's nothing wrong with the fundamental architecture (other than power
consumption). The newer systems are faster and consume a lot less power
and are therefore a lot cheaper to run.
RoadRunner hit 1.042PFlops and consumes 2.3MW (and currently sits at #22
on the super computer list). #21 consumes only 1.2MW, and #23 @ 1.035PFlops
consumes 493KW (ie almost as fast but consumes ~1/5 the power).
TTFN - Guy
1 Apr
1 Apr
12:56 p.m.
On 03/31/2013 10:03 PM, Guy Sotomayor wrote:
The newer systems are faster and consume a lot less
power
and are therefore a lot cheaper to run.
Pfft. Just hook a solar panel or two up to it. Or rather, lots. And only
run it during the daytime, when it's a nice weather.
4:02 p.m.
On Apr 1, 2013, at 12:56 PM, Jules Richardson <jules.richardson99 at gmail.com>
wrote:
On 03/31/2013 10:03 PM, Guy Sotomayor wrote:
It's in Los Alamos, so it's always sunny (ok 80% of the time). ;-)
TTFN - Guy
The newer systems are faster and consume a lot
less power
and are therefore a lot cheaper to run.
Pfft. Just hook a solar panel or two up to it. Or rather, lots. And only run it during
the daytime, when it's a nice weather. 
2 Apr
2 Apr
1:43 p.m.
On 04/01/2013 02:33 AM, Fred Cisin wrote:
Built, used for four years, and now DISMANTLED?
What is wrong with their fundamental architecture?
Can't they design one where more could be added to keep up with needs, and
upgrade incrementally and components while keeping the overall machine?
What would you replace? The CPU? Then you need a new socket and then you
need a new motherboard. I suppose you could hold it against them that
they didn't make the CPU's pin compatible, but that's only viable for a
few years. The pin count changes as more and more features are moved
onboard the CPU. And I suppose you could hold it against them that they
didn't put the CPU on a daughterboard, but then the cost of
manufacturing of the daughterboard would probably be close to a whole
motherboard. Much easier to throw out and replace and let somebody else
worry about assembly.
Ram speeds go up, so you want to replace those when you get a new CPU as
well.
What is left? Power supplies and fans? yeah, those probably have a four
more years in them. Probably would make sense to reuse.
/P
2:31 p.m.
> Built, used for four years, and now DISMANTLED?
> What is wrong with their fundamental architecture?
> Can't they design one where more could be added to keep up with needs, and
> upgrade incrementally and components while keeping the overall machine?
On
Tue, 2 Apr 2013, Pontus wrote:
What would you replace? The CPU? Then you need a new
socket and then you
need a new motherboard. I suppose you could hold it against them that
they didn't make the CPU's pin compatible, but that's only viable for a
few years. The pin count changes as more and more features are moved
onboard the CPU. And I suppose you could hold it against them that they
didn't put the CPU on a daughterboard, but then the cost of
manufacturing of the daughterboard would probably be close to a whole
motherboard. Much easier to throw out and replace and let somebody else
worry about assembly.
Ram speeds go up, so you want to replace those when you get a new CPU as
well.
What is left? Power supplies and fans? yeah, those probably have a four
more years in them. Probably would make sense to reuse.
I'm having a little difficulty visualizing "The World's Top
Supercomputer"
as being a single chip CPU on a motherboard.
When did "supercomputers" become single board devices?
"Put the CPU on a daughterboard"?
I'm kinda stuck thinking in terms of adding another rack that
supplants PART of the CPU functions with faster, and relegating
the rack(s) that it replaced to anciliary functions.
I never realized that inside one of those racks there was a
CPU chip plugged into a socket.
2:43 p.m.
New subject: World's top supercomputer from '09 is now obsolete, will be dismantled
-----Original Message-----
From: cctalk-bounces at classiccmp.org [mailto:cctalk-bounces at classiccmp.org]
On Behalf Of Fred Cisin
Sent: Tuesday, April 02, 2013 4:32 PM
To: General Discussion: On-Topic and Off-Topic Posts
Subject: Re: World's top supercomputer from '09 is now obsolete, will be
dismantled
You guys are making me wish I had hung onto the old Cray supercomputers that
we had abt 15 years ago. Sold the hard drives, and scrapped the rest.
The IBM Big Blue cabinets, we shipped via common carrier, and some idiot ran
his forklift through the cabinets!
The DEC refrigerator-sized cabinets, we pulled the drives and boards,
scrapped the rest.
If I had only known then what I know now....
-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2013.0.3267 / Virus Database: 3161/6218 - Release Date: 04/01/13
2:55 p.m.
On 04/02/2013 04:31 PM, Fred Cisin wrote:
I expect it's lots of CPUs per board, probably soldered direct to the PCB,
and then lots of boards per rack, with custom interconnects between boards
and racks. To a point you may be able to add racks and boards, but once the
individual CPUs (or the memory, or the interconnects) are deemed to be too
slow then it's game over.
As Paul says, it may be possible to keep the power supplies (and the empty
racks) but that's such a small part of the whole - and the topology might
not even be suited to whatever comes next anyway - that there's little point.
I suppose the sheer logistics of moving something like this as a going
concern - particularly if the space is needed for a replacement - are what
prevents the whole lot being sold on to some other entity (i.e. just
because it's not fast enough for you doesn't mean that it's useless to
someone else)
cheers
Jules
> Built,
used for four years, and now DISMANTLED?
> What is wrong with their fundamental architecture?
> Can't they design one where more could be added to keep up with needs, and
> upgrade incrementally and components while keeping the overall machine?
On
Tue, 2 Apr 2013, Pontus wrote:
What would you replace? The CPU? Then you need a
new socket and then you
need a new motherboard. I suppose you could hold it against them that
they didn't make the CPU's pin compatible, but that's only viable for a
few years. The pin count changes as more and more features are moved
onboard the CPU. And I suppose you could hold it against them that they
didn't put the CPU on a daughterboard, but then the cost of
manufacturing of the daughterboard would probably be close to a whole
motherboard. Much easier to throw out and replace and let somebody else
worry about assembly.
Ram speeds go up, so you want to replace those when you get a new CPU as
well.
What is left? Power supplies and fans? yeah, those probably have a four
more years in them. Probably would make sense to reuse.
I'm having a little difficulty visualizing "The World's Top
Supercomputer"
as being a single chip CPU on a motherboard.
When did "supercomputers" become single board devices?
"Put the CPU on a daughterboard"?
I'm kinda stuck thinking in terms of adding another rack that
supplants PART of the CPU functions with faster, and relegating
the rack(s) that it replaced to anciliary functions.
I never realized that inside one of those racks there was a
CPU chip plugged into a socket.
3:34 p.m.
>> Built, used for four years, and now
DISMANTLED?
>> What is wrong with their fundamental architecture?
>> Can't they design one where more could be added to keep up with needs, and
>> upgrade incrementally and components while keeping the overall machine?
?As time goes on the density of the processors goes up but its not just the CPU's. Its
what goes on behind the processors such as SGI's MIPS supercomputers of a few years
ago. They used a distributed shared memory that required circuitry that connected
processors, local memory, a network interface, and the
I/O. These systems can be difficult to add onto brick wise and the only way to increase
speed is to add racks which adds power consumption.
Some of the technologies like Infiniband does have an upgrade path where older network
technologies like HIPPI were not.
Some of the Crays are field upgradable to new processor blades.
Like this:
https://www.youtube.com/watch?v=ehCgG9yZQmg
6:03 p.m.
On Tue, Apr 2, 2013 at 6:34 PM, Brian Roth <abacos_98 at yahoo.com> wrote:
As time goes on the density of the processors goes up but its not just
the CPU's. Its what goes on behind the processors such as SGI's MIPS
supercomputers of a few years ago. They used a distributed shared memory
that required circuitry that connected processors, local memory, a network
interface, and the
I/O. These systems can be difficult to add onto brick wise and the only
way to increase speed is to add racks which adds power consumption.
The other problem with Roadrunner was it was a hybrid design - AMD Opterons
combined with cell processors, like the one in the Playstation 3. At the
time these were screaming fast, but difficult to program.
6:15 p.m.
On Apr 2, 2013, at 9:03 PM, Jason McBrien wrote:
On Tue, Apr 2, 2013 at 6:34 PM, Brian Roth
<abacos_98 at yahoo.com> wrote:
They're about as difficult to program as any other supercomputer
processor. The whole point of the original Cray architecture
(and other Cray-designed CDC supercomputers, IIRC) was that there
were I/O processors performing the data transfers to provide the
CPU(s) enough data to chew on at full throttle. Anything less,
and you were wasting processor power. The Cell's ring data
structure wasn't all that different.
Which, I guess, is not to say that they weren't difficult to
program, but that supercomputer programmers probably should have
felt right at home.
- Dave
As time goes on the density of the processors goes up but its not just
the CPU's. Its what goes on behind the processors such as SGI's MIPS
supercomputers of a few years ago. They used a distributed shared memory
that required circuitry that connected processors, local memory, a network
interface, and the
I/O. These systems can be difficult to add onto brick wise and the only
way to increase speed is to add racks which adds power consumption.
The other problem with Roadrunner was it was a hybrid design - AMD Opterons
combined with cell processors, like the one in the Playstation 3. At the
time these were screaming fast, but difficult to program.
3 Apr
3 Apr
11:29 a.m.
On Tue, Apr 2, 2013 at 9:15 PM, David Riley <fraveydank at gmail.com> wrote:
Which, I guess, is not to say that they weren't difficult to
program, but that supercomputer programmers probably should have
felt right at home
The complaint I heard was the tools were primitive relative to the other
architectures out there. You had a somewhat spartan gcc toolchain and BLAS.
Compare that with the dozens of highly optimized libraries and toolchains
for even generic x86, much less specific chips if you're talking about
Intel, especially when you're dealing with threading.
2 Apr
2 Apr
11:15 p.m.
On Tue, Apr 02, 2013 at 02:31:39PM -0700, Fred Cisin wrote:
I'm having a little difficulty visualizing "The World's Top
Supercomputer"
as being a single chip CPU on a motherboard.
When did "supercomputers" become single board devices?
"Put the CPU on a daughterboard"?
Do you really think I'm that stupid? I've poked my head into plenty of
supercomputers, in fact I have the very rack of a top500 list entry in
my garage.
I'm not even going to reply to that...
I'm kinda stuck thinking in terms of adding
another rack that
supplants PART of the CPU functions with faster, and relegating
the rack(s) that it replaced to anciliary functions.
What anciliary functions? http servers? storage? The performance per
watt is way of for that kind of application.
Hopefully, the machines won't hit the recyclers but end up in hands of
students, less fortunate and hobbyists.
I never realized that inside one of those racks there
was a
CPU chip plugged into a socket.
Again? This is close to insulting.
Regards,
Pontus.
11:38 p.m.
New subject: World?s top supercomputer from ?09 is now obsolete, will be dismantled
Quoting Pontus Pihlgren <pontus at Update.UU.SE>:
On Tue, Apr 02, 2013 at 02:31:39PM -0700, Fred Cisin
wrote:
Just for entertainment. That's exactly how Intel introduced the i860
back then.
"cray on a chip" if I remember correctly ;-)
When did "supercomputers" become single
board devices?
"Put the CPU on a daughterboard"?
Do you really think I'm that stupid? I've poked my head into plenty of
supercomputers, in fact I have the very rack of a top500 list entry in
my garage.
3 Apr
3 Apr
9:41 a.m.
On Tue, Apr 02, 2013 at 02:31:39PM -0700, Fred Cisin
wrote:
> I'm having a little difficulty visualizing "The World's Top
Supercomputer"
> as being a single chip CPU on a motherboard.
> When did "supercomputers" become single board devices?
> "Put the CPU on a daughterboard"?
On Wed, 3 Apr 2013, Pontus Pihlgren wrote:
Do you really think I'm that stupid?
Of course not. I acknowledge and respect your expertise.
And, I certainly had no intention of offending you.
I've poked my head into plenty of
supercomputers, in fact I have the very rack of a top500 list entry in
my garage.
I'm not even going to reply to that...
That's why I was so incredulous and taken aback by your OFF THE WALL
statements!
(YOU were the one who made those statatements - full post below)
I even speculated that my whole perception of such machines was wrong and
that things had gone through extraordinary changes that I was unaware of!
I'm kinda
stuck thinking in terms of adding another rack that
supplants PART of the CPU functions with faster, and relegating
the rack(s) that it replaced to anciliary functions.
What anciliary functions?
http servers? storage? The performance per
watt is way of[f] for that kind of application. Hopefully, the machines won't hit the recyclers
but end up in hands of
students, less fortunate and hobbyists.
Let's hope so, but I am cynical about
the prospects.
I never
realized that inside one of those racks there was a
CPU chip plugged into a socket.
Again? This is close to insulting. Built, used for four years, and now DISMANTLED?
What is wrong with their fundamental architecture?
Can't they design one where more could be added to keep up with needs,
and
upgrade incrementally and components while keeping the
overall machine?
What would you replace? The CPU? Then you need a new socket and then you
need a new motherboard. I suppose you could hold it against them that
they didn't make the CPU's pin compatible, but that's only viable for a
few years. The pin count changes as more and more features are moved
onboard the CPU. And I suppose you could hold it against them that they
didn't put the CPU on a daughterboard, but then the cost of
manufacturing of the daughterboard would probably be close to a whole
motherboard. Much easier to throw out and replace and let somebody else
worry about assembly.
Ram speeds go up, so you want to replace those when you get a new CPU as
well.
What is left? Power supplies and fans? yeah, those probably have a four
more years in them. Probably would make sense to reuse.
/P
----------------------- End ----------------------------
noon
On 04/03/2013 06:41 PM, Fred Cisin wrote:
That's why I was so incredulous and taken aback by
your OFF THE WALL
statements! (YOU were the one who made those statatements - full post
below) I even speculated that my whole perception of such machines was
wrong and that things had gone through extraordinary changes that I
was unaware of!
I apparently made the false assumption that we both shared the view on
what the Road Runner really is, a collection of CPU's each in its own
motherboard each in its own case. So I carelessly thought that when I
referred to the possibility of replacing the CPU it would be clear that
each CPU would be replaced from it's socket.
I seems that you view Roadrunner as _one_ computer and then it certainly
becomes ridiculous to talk about _the_ CPU in _a_ socket and I
understand why you think I was off the wall.
I'm sorry for the confusion and I hope you are not just being sarcastic
with me, that translates badly in text.
Going back to topic. It certainly is possible to design a super computer
that is modular. The SGI Origin2k is a wonderful example of that, though
it only scales to a certain limit (128 CPUs I think). Heck, even the
Roadrunner is quite modular, you could probably cut it up into a hundred
different pieces and each piece would be quite a competent cluster of
computer for a researcher willing to pay the power bill.
Kind Regards,
Pontus.
2:26 p.m.
On Wed, 3 Apr 2013, Pontus wrote:
I apparently made the false assumption that we both
shared the view on
what the Road Runner really is, a collection of CPU's each in its own
motherboard each in its own case. So I carelessly thought that when I
referred to the possibility of replacing the CPU it would be clear that
each CPU would be replaced from it's socket.
I seems that you view Roadrunner as _one_ computer and then it certainly
becomes ridiculous to talk about _the_ CPU in _a_ socket and I
understand why you think I was off the wall.
Yeah, that's most of the confusion. I honestly don't know anything about
Roadrunner, nor anything about the machine being dismantled after 4 years.
Since "top supercomputer" is CERTAINLY going to be upgraded (or replaced)
it seemed that upgradeability and/or expansion would be an extremely high
priority in the design.
I'm sorry for the confusion and I hope you are not
just being sarcastic
with me, that translates badly in text.
No, no sarcasm at all.
I know that I don't know anything about high end machines, but I was
shocked that the design life was only 4 years.
Going back to topic. It certainly is possible to
design a super computer
that is modular. The SGI Origin2k is a wonderful example of that, though
it only scales to a certain limit (128 CPUs I think). Heck, even the
Roadrunner is quite modular, you could probably cut it up into a hundred
different pieces and each piece would be quite a competent cluster of
computer for a researcher willing to pay the power bill.
I like the idea of that, but honestly have no idea of the practicality.
Could multiple 128s be "clustered"?
ANY finite limit will eventually get in the way.
--
Grumpy Ol' Fred cisin at xenosoft.com
10:25 p.m.
On Wed, Apr 03, 2013 at 02:26:36PM -0700, Fred Cisin wrote:
Yeah, that's most of the confusion. I honestly
don't know anything about
Roadrunner, nor anything about the machine being dismantled after 4 years.
Then that is settled, nomhard feelings i hope.
Since "top supercomputer" is CERTAINLY going
to be upgraded (or replaced)
it seemed that upgradeability and/or expansion would be an extremely high
priority in the design.
To some degree at least. The Titan (I think it was) got an upgrade by replacing all of
it's
CPUs.
I like the idea of that, but honestly have no idea of
the practicality.
Could multiple 128s be "clustered"?
Now I had to go and read up. It turns out the O2k could be run in 256 and 512 CPU
configurations as well but it was not marketed and supported.
And certainly such configurations could then be clustered. In fact it has been done with
48
128 CPU O2k computers interconnected with HIPPI using MPI. Read more here, the linked PDF
in the article has a nice picture.
http://en.wikipedia.org/wiki/ASCI_Blue_Mountain
regards,
Pontus
2 Apr
2 Apr
6:15 p.m.
On 3/31/2013 7:33 PM, Fred Cisin wrote:
Built, used for four years, and now DISMANTLED?
What is wrong with their fundamental architecture?
<snip>
Nothing really. People learn. Four years is a long time at the high
end of computer science.
I worked on the first BlueGene (BlueGene/L) to make the list. Not too
long after that machine fell off the list RoadRunner made the top of the
list. (I had a lot of friends who worked on RoadRunner. Both were IBM
machines.)
BlueGene/L was quite a bit different than the NEC machine that held the
top slot before it. There were 64K "compute nodes" and something on the
order of 1K "I/O" nodes. Both variants used an embedded style PowerPC
system on a chip, but they were wired a little differently. Besides the
massive number of processors there was a lot of thought put into the
communications between processors ("tree" and "torus") to make it
suitable for parallel workloads that still required some coordination.
RoadRunner was a much different beast. It was essentially IBM QS22
Blades (Cell processor with enhanced double precision support compared
to the variant used in the Sony PS3) and another blade (AMD x86 I
think). Two QS22s and the other blade were packaged together in what we
called a "tri-blade", which served as the basic building block. The x86
blade handled I/O and management and left the CPU intensive work to the
QS22s.
It's really easy to throw rocks a something you don't understand. Some
light reading on Wikipedia would have shown you how much different the
different #1 machines are from each other.
Wait until we see the "exascale" style machines ...
31 Mar
31 Mar
4:17 p.m.
On 03/31/2013 02:14 PM, St?phane Tsacas wrote:
It's probably my age, but I just can't get excited about the massive
multiprocessor boxes.
--Chuck
4:52 p.m.
On 3/31/13 5:17 PM, Chuck Guzis wrote:
On 03/31/2013 02:14 PM, St?phane Tsacas wrote:
It's probably my age, but I just can't get excited about the massive
multiprocessor boxes.
--Chuck
LOL, yeah, now they can probably do the same work with 2 iPads and a
couple of 100tb. NAS arrays
--
Dave Land
Land Computer Service
Check out my site at http://www.landcomp.net
5 p.m.
On 03/31/2013 07:52 PM, Dave Land wrote:
With a supercomputer from four years ago? Uh, no.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
It's probably my age, but I just can't get excited about the massive
multiprocessor boxes.
LOL, yeah, now they can probably do the same work with 2 iPads and a couple
of 100tb. NAS arrays 
5:10 p.m.
On Sun, 31 Mar 2013, Dave McGuire wrote:
Date: Sun, 31 Mar 2013 20:00:06 -0400
From: Dave McGuire <mcguire at neurotica.com>
Reply-To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk at classiccmp.org>
To: General Discussion: On-Topic and Off-Topic Posts <cctalk at classiccmp.org>
;
Subject: Re: [windows-1252] World~Rs top supercomputer from ~Q09 is now
obsolete, will be dismantled
On 03/31/2013 07:52 PM, Dave Land wrote:
With a supercomputer from four years ago? Uh, no.
-Dave
--
Dave McGuire, AK4HZ
New Kensington, PA
Looks like using the GPU in the latest IPAD at full blast will give you ~77
GFlops (but no way high bandwidth way to connect) so about 13,158 ipads to
equal the decommissioned ~1 PF machine in bare CPU horsepower
Peter Wallace
Mesa Electronics
It's probably my age, but I just can't get excited about the massive
multiprocessor boxes.
LOL, yeah, now they can probably do the same work with 2 iPads and a couple
of 100tb. NAS arrays
4049
days inactive
4053
days old
36 comments
19 participants
participants (19)
-
abacos_98＠yahoo.com -
ard＠p850ug1.demon.co.uk -
cclist＠sydex.com -
cisin＠xenosoft.com -
dave.g4ugm＠gmail.com -
emu＠e-bbes.com -
fraveydank＠gmail.com -
ggs＠shiresoft.com -
glen.slick＠gmail.com -
jbmcb1＠gmail.com -
jules.richardson99＠gmail.com -
marcogb＠xs4all.nl -
mbbrutman-cctalk＠brutman.com -
mcguire＠neurotica.com -
pcw＠mesanet.com -
pontus＠update.uu.se -
sales＠elecplus.com -
stephane.tsacas＠gmail.com -
xmechanic＠landcomp.net