On Fri, Apr 03, 2009 at 07:34:26AM +0800, Stuart Midgley wrote:
> In 2019 when Intel releases their 1024core chip, still at 2.5GHz with
> 256GB dimm memory, a lot of people will be surprised that linux works
Ain't going to work. Arguably doesn't work already beyond 8 cores
in a single socket.
Bill, if i use the wording latency versus bandwidth,
what is to your best GUESS the difference?
will you at least have enough respect for the list
that you send your insults by direct email?
___
Beowulf mailing list, Beowulf@beowulf.org
To change your
Vincent Diepeveen wrote:
> Bill,
>
> the ONLY price that matters is that of ECC ram when posting in a cluster
> group.
>
>
> If there is 1 commission that EVER puts a signature underneath a
> production cluster
> without ECC ram using x86 processors (gpu's is yet another new thing
> that is
Bill, if i use the wording latency versus bandwidth,
what is to your best GUESS the difference?
What you report is all bandwidth, not latency.
Vincent
___
Beowulf mailing list, Beowulf@beowulf.org
To change your subscription (digest mode or unsubscrib
Vincent Diepeveen wrote:
> Bill,
>
> the ONLY price that matters is that of ECC ram when posting in a cluster
> group.
Agreed. All the numbers and URLs I mentioned were for ECC ram.
> So in short i can completely ignore your posting.
>
> ECC is a requirement, not a luxury.
Maybe you should re
Bill,
the ONLY price that matters is that of ECC ram when posting in a
cluster group.
If there is 1 commission that EVER puts a signature underneath a
production cluster
without ECC ram using x86 processors (gpu's is yet another new thing
that is interesting
to discuss), then please info
Well Glen,
You're speaking of the new Sun.
After Sun got hammered with some nukes, the remaining survivors had
first some genetic
defects, resulting in niagara. The nuke survivors then released the
x86-64 platform much against
their will initially, yet it really is serving them well by now a
Vincent Diepeveen wrote:
> I wouldn't bet at registered-ecc DDR3 ram to become cheaper.
> To be honest i misjudged that for DDR reg-ecc ram also,
> it still is relative spoken expensive.
I've heard this a dozen times, seems repeated quite often. Yet when I
actually look I see it either so small a
Yes a sad day,
especially considering they were supporting shared memory,
and i can just compile my chess engine, modify 2 settings and
it works right out of the box there at up to 2048 processors,
no problem.
MPI is a lot slower and would require a huge rewrite of my code.
Of course you can com
On Apr 1, 2009, at 11:08 AM, Jan Heichler wrote:
Hallo Mikhail,
Dienstag, 31. März 2009, meintest Du:
MK> In message from Kilian CAVALOTTI
MK> (Tue, 31 Mar 2009 10:27:55 +0200):
>> ...
>>Any other numbers, people?
MK> I beleive there is also a bit other important numbers - prices f
Stuart Midgley wrote:
So, to get back to the original discussion, SGI disappearing from the
landscape means one less option to choose from.
You think SGI designed the hardware for their non-shared-memory
clusters? I don't thnk so.
-- greg
No, SGI use supermicro. But that isn't really th
I did mention the point that the shared-memory machines aren't
selling. If you're missing them, you're in a rather elite and small
group ;-)
In 2019 when Intel releases their 1024core chip, still at 2.5GHz with
256GB dimm memory, a lot of people will be surprised that linux works
and has d
> On Fri, Apr 03, 2009 at 07:11:05AM +0800, Stuart Midgley wrote:
>
>> No, SGI use supermicro. But that isn't really the point. They did
>> design their own shared memory systems and they will be sorely missed if
>> they are not continued. We were really looking forward to a large
>> shared mem
On Fri, Apr 03, 2009 at 07:11:05AM +0800, Stuart Midgley wrote:
> No, SGI use supermicro. But that isn't really the point. They did
> design their own shared memory systems and they will be sorely missed if
> they are not continued. We were really looking forward to a large
> shared memory
> Personally, I don't think there's anything wrong with
> domain-specific processors. Cell, ARM, Sparc T2's, etc all
> have a certain market which they are suited to and work well
> in, but are pooly suited for workloads outside that market.
> There's no point in burning massive amounts of die
So, to get back to the original discussion, SGI disappearing from the
landscape means one less option to choose from.
You think SGI designed the hardware for their non-shared-memory
clusters? I don't thnk so.
-- greg
No, SGI use supermicro. But that isn't really the point. They did
de
When corporate trumped engineering, it was only a matter of time.
When they told Jim Clark to not come back to port, it was only a
matter of time.
When the USSR economic system went into conniptions, repeat purchases
fell in the USA, and the big company eventually couldn't feed itself.
I don'
On Thu, 2 Apr 2009, Andrew Piskorski wrote:
On Wed, Apr 01, 2009 at 04:56:53PM -0700, Bill Broadley wrote:
http://news.cnet.com/8301-1001_3-10209580-92.html
According to that articel, the pictured dual socket server has two
hard drives, but from the second photo, although the view is obscured
With all due respect, rgb, you're somewhat out of date with respect to
(Open)Solaris. I don't think anyone would argue that Solaris x86 wasn't a
mess prior to Solaris 10. However, it took a massive step forward in Solaris
10 once Sun really started pumping out Opteron servers. Additionally, it's
On Thu, Apr 02, 2009 at 10:11:07AM -0400, Prentice Bisbal wrote:
> Or it could be because they make motherboards that convert 12 VDC to 5
> VDC on the motherboard.
All Itanium and some other x86 boxes take a single 48V input to the
mobo. I talked to a mobo designer once and he claimed that there
april 1st?
Ricardo Reis
'Non Serviam'
PhD student @ Lasef
Computational Fluid Dynamics, High Performance Computing, Turbulence
http://www.lasef.ist.utl.pt
&
Cultural Instigator @ Rádio Zero
http://www.radiozero.pt
http://www.flickr.com/photos/rreis/__
Robert G. Brown wrote:
>On Wed, 1 Apr 2009, Ellis Wilson wrote:
>
>>
>> Beyond that, building an entire cluster of that size into a large
>> shipping container is genius - given access and resources to a crane and
>> proper machining expertise. But then again if your building a cluster
>> of that
On Thu, Apr 02, 2009 at 10:18:39AM -0400, Mark Hahn wrote:
>> April 1st, 2009 : Rich Miller
>
> was this just a very droll and not very funny april fools article?
This April 1 was full of odd but real announcements.
Strange... perhaps that is the joke in and of itself.
>> uninterruptible power su
Greg Lindahl wrote:
The hard part in a 48-disk system is the software, not the hardware.
We've been using thumpers and zfs since they debuted. They have proven
reliable and efficient for our usage.
--
Mike Davis Technical Director
(804) 828-3885 C
> > According to that articel, the pictured dual socket server has two
> > hard drives, but from the second photo, although the view is obscured
> > by the UPS battery, it looks like four drives to me, stacked two high.
> > Or if it's really only two drives, what are they sitting on top of?
>
> I
On Thu, Apr 02, 2009 at 06:40:47PM +0200, Loic Tortay wrote:
> Ah ah ah, you've obviously never seen a "Thumper". :)
OK, I must have ass-u-me-d something when it came out -- they were
touting "48 drives in 5U!!", right at the same time that a commodity
chassis with 48 drives in 5U (and room for a
> -Original Message-
> From: Donald Becker [mailto:bec...@scyld.com]
> Sent: Thursday, April 02, 2009 11:36 AM
> To: Lux, James P
> Cc: Greg Lindahl; beowulf@beowulf.org
> Subject: RE: [Beowulf] Interesting google server design
>
> On Thu, 2 Apr 2009, Lux, James P wrote:
> > > On Thu, Ap
> However, I would imagine this is a trivial amount at 12 VDC
> vs 5 VDC and the incredibly short distances the electricity
> travels on a PCB trace.
Actually, it's a non-trivial amount. Consider that a processor that draws 100W
at 3.3V is drawing 30Amps.
If you're using 2oz copper (typical
> However, I would imagine this is a trivial amount at 12 VDC
> vs 5 VDC and the incredibly short distances the electricity
> travels on a PCB trace.
Actually, it's a non-trivial amount. Consider that a processor that draws 100W
at 3.3V is drawing 30Amps.
If you're using 2oz copper (typical
On Wed, 1 Apr 2009, Ian Dillon wrote:
> Likewise Joe, watching the demise of SGI has been incredibly hard to fathom.
> This was a great company with tons of energy/potential and did right by
> their employees back in the day. I think this is why so many of us feel for
> the company and its curren
On Thu, 2 Apr 2009, Lux, James P wrote:
> > On Thu, Apr 02, 2009 at 10:11:07AM -0400, Prentice Bisbal wrote:
> > > Or it could be because they make motherboards that convert
> > 12 VDC to 5
> > > VDC on the motherboard.
> > Greg Lindahl wrote:
> > All Itanium and some other x86 boxes take a sing
Bill Broadley wrote:
Andrew Piskorski wrote:
On Wed, Apr 01, 2009 at 04:56:53PM -0700, Bill Broadley wrote:
http://news.cnet.com/8301-1001_3-10209580-92.html
According to that articel, the pictured dual socket server has two
hard drives, but from the second photo, although the view is obscured
Greg Lindahl wrote:
> On Thu, Apr 02, 2009 at 10:11:07AM -0400, Prentice Bisbal wrote:
>
>> Or it could be because they make motherboards that convert 12 VDC to 5
>> VDC on the motherboard.
>
> All Itanium and some other x86 boxes take a single 48V input to the
> mobo. I talked to a mobo designer
Andrew Piskorski wrote:
> On Wed, Apr 01, 2009 at 04:56:53PM -0700, Bill Broadley wrote:
>> http://news.cnet.com/8301-1001_3-10209580-92.html
>
> According to that articel, the pictured dual socket server has two
> hard drives, but from the second photo, although the view is obscured
> by the UPS
Greg Lindahl wrote:
On Thu, Apr 02, 2009 at 10:32:22AM +0200, Kilian CAVALOTTI wrote:
Anyway, I was thinking about companies like Penguin or Scalable, designing
their own hardware products, like Relions or JackRabbits, and not just
reselling Supermicro boxes. Unless I'm mistaken on the fact th
> -Original Message-
> From: beowulf-boun...@beowulf.org
> [mailto:beowulf-boun...@beowulf.org] On Behalf Of Greg Lindahl
> Sent: Thursday, April 02, 2009 9:31 AM
> To: beowulf@beowulf.org
> Subject: Re: [Beowulf] Interesting google server design
>
> On Thu, Apr 02, 2009 at 10:11:07AM -
We have one query on vbuf_lock in mvapich2
Here for uDAPL code (located inside src/mpid/ch3/channels/mrail/src/udapl)
do not maintain any protection like “vbuf_lock” where as similar code for IB
(located in src/mpid/ch3/channels/mrail/src/gen2) we have used vbuf_lock
protection.
Does any one know
Orion Poplawski wrote:
Jan Heichler wrote:
Don't forget DDR3-ECC memory which must be registered.
http://www.siliconmechanics.com/i22789/2U-4-node-server.php shows
options for "ECC Unbuffered" or "ECC Registered". No idea what the
practical difference would be. Anyone?
Registered DIMMs
Rahul Nabar wrote:
On Tue, Mar 31, 2009 at 6:43 PM, Don Holmgren wrote:
Instead of logging into the node directly, you might want to try an
interactive
job (use "qsub -I") and then try your mpirun. This may give you messages
that
for some reason aren't getting back to you in your job's .o or
Likewise Joe, watching the demise of SGI has been incredibly hard to fathom.
This was a great company with tons of energy/potential and did right by
their employees back in the day. I think this is why so many of us feel for
the company and its current employee base.
-Original Message-
Dear John and others,
first of all, I am reading the list now for some years and I gained quite a
bit of knowledge here, so why shall I leave? Don't worry, I got no intention
to do so.
I can see you point and I have already contacted the cluster support.
However, I have mixed experiences with s
Killian,
As it turns out, Joe Landman and I had a very good conversation about the
state of HPC on Monday. Please note that I do not want to turn this thread
into a sales and marketing pitch but rather make a couple suggestions to
your concern about EMEA support/resources for HPC.
I head up th
On Mon, Mar 30, 2009 at 02:14:50PM +0100, J?rg Sa?mannshausen wrote:
> Dear all,
>
> I am having this rather anoying problem with the parallel execution of
> one of the programs (GAMESS US version) on our cluster. The error
> message is:
>
> TCP connect error: ECONNREFUSED.
> TCP: Connect fai
On Wed, Apr 01, 2009 at 04:56:53PM -0700, Bill Broadley wrote:
> http://news.cnet.com/8301-1001_3-10209580-92.html
According to that articel, the pictured dual socket server has two
hard drives, but from the second photo, although the view is obscured
by the UPS battery, it looks like four drives
Hallo Greg,
Donnerstag, 2. April 2009, meintest Du:
GL> On Thu, Apr 02, 2009 at 10:32:22AM +0200, Kilian CAVALOTTI wrote:
>> Anyway, I was thinking about companies like Penguin or Scalable, designing
>> their own hardware products, like Relions or JackRabbits, and not just
>> reselling Supermi
2009/4/2 Greg Lindahl :
>
> You think SGI designed the hardware for their non-shared-memory
> clusters? I don't thnk so.
>
Greg, the Atoka motherboard in the ICE cluster was designed in a
cooperative effort by SGI and Supermicro,
as far as I know. Given that they have pinouts on the rear for the S
On Thu, Apr 02, 2009 at 10:11:07AM -0400, Prentice Bisbal wrote:
> Or it could be because they make motherboards that convert 12 VDC to 5
> VDC on the motherboard.
All Itanium and some other x86 boxes take a single 48V input to the
mobo. I talked to a mobo designer once and he claimed that there
On Thu, Apr 02, 2009 at 10:32:22AM +0200, Kilian CAVALOTTI wrote:
> Anyway, I was thinking about companies like Penguin or Scalable, designing
> their own hardware products, like Relions or JackRabbits, and not just
> reselling Supermicro boxes. Unless I'm mistaken on the fact that those
> syst
April 1st, 2009 : Rich Miller
was this just a very droll and not very funny april fools article?
uninterruptible power supply (UPS). The design shifts the UPS and battery
backup functions from the data center into the server cabinet (see our
well, that sounds reasonable - I guess the point i
Ellis Wilson wrote:
>
> Also interesting is that they use Gigabyte - I haven't been entirely
> impressed with them and that is for purely desktop use. Perhaps their
> server grade boards are better quality enough to make them worthwhile at
Google probably uses Gigabyte because they have the b
John Hearns wrote:
2009/4/1 Robert G. Brown :
Here, let me add a little spot of brandy to that cup. They deserve at
least that... for it was upon their shoulders and out of their egg that
Linux was borne, as all the SunOS and Irix users in the Universe,
frustrated with high prices and incompati
Glen Beane wrote:
On 4/1/09 10:45 AM, "Joe Landman" wrote:
This is a current religious mantra within Sun, that Solaris/Sparc will
(eventually) kill Linux and x86/x64. Um. No.
Isn't Sun's biggest business x86-64 hardware running Linux? My cluster is
all AMD64 Sun gear and they gave us a k
On Thu, 2 Apr 2009, Kilian CAVALOTTI wrote:
On Wednesday 01 April 2009 17:48:56 you wrote:
A point I
didn't make in my blog post on this stuff is that the academic side,
while a nice market to play in, isn't terribly profitable. You have to
give "killer" discounts to play in many cases. Which
While I really appreciate the elegy, the summation of this is: The
hardware vendors we once knew have, or are, slowly doing themselves in.
The ones we're going to recognize in the foreseeable future are Intel,
AMD and Freescale, plus a few others around the edges (Via, for
instance), who learn
Jan Heichler wrote:
And if a small company does something new you won't sell a huge
quantity. So the components get (usually) more expensive. The
engineering isn't quite as advanced (because it is expensive) and many
errors just show when you have a larger install-base.
Most large (and smal
Hallo Geoff,
Donnerstag, 2. April 2009, meintest Du:
>> Also it's not like big discounts don't exist in the corporate world.
>> Back when I worked for $multinational our group rates with Dell were
>> astonishingly cheap.
GG> There was one university I worked for that Dell sold a medium-size
Geoff Galitz wrote:
Also it's not like big discounts don't exist in the corporate world.
Back when I worked for $multinational our group rates with Dell were
astonishingly cheap.
There was one university I worked for that Dell sold a medium-sized (approx
100 nodes) cluster complete with
Huw Lynes wrote:
On Thu, 2009-04-02 at 10:49 +0200, Kilian CAVALOTTI wrote:
On Wednesday 01 April 2009 17:48:56 you wrote:
A point I
didn't make in my blog post on this stuff is that the academic side,
while a nice market to play in, isn't terribly profitable. You have to
give "killer" discoun
Kilian CAVALOTTI wrote:
On Wednesday 01 April 2009 17:48:56 you wrote:
A point I
didn't make in my blog post on this stuff is that the academic side,
while a nice market to play in, isn't terribly profitable. You have to
give "killer" discounts to play in many cases. Which in the case of a
low
On Wed, 1 Apr 2009, Ellis Wilson wrote:
Beyond that, building an entire cluster of that size into a large
shipping container is genius - given access and resources to a crane and
proper machining expertise. But then again if your building a cluster
of that size I suppose the crane, shipping co
On Thursday 02 April 2009 11:13:28 Jan Heichler wrote:
> KC> My bad here for not being explicit enough. By "their own hardware", I
> meant KC> "hardware they put a lot of thought in, that they carefully
> conceived and KC> designed, and have manufactured according to their own
> needs and guideline
On Thursday 02 April 2009 06:44:53 Joshua Baker-LePain wrote:
> AFAIK, the only difference is available capacity (well, and price of
> course). You can't get unbuffered DIMMs bigger than 2GB. Also note that
> memory speed goes down as you add more DIMMs per channel (it's also a
> function of CPU
Hallo Kilian,
Donnerstag, 2. April 2009, meintest Du:
KC> On Wednesday 01 April 2009 16:19:57 John Hearns wrote:
>> > There are a few services/integration HPC
>> > companies in the EU, but not any that I'm aware of selling their own
>> > hardware, as Scalable or Penguin do.
>> ? I know that
> Also it's not like big discounts don't exist in the corporate world.
> Back when I worked for $multinational our group rates with Dell were
> astonishingly cheap.
There was one university I worked for that Dell sold a medium-sized (approx
100 nodes) cluster complete with myrinet network
On Thu, 2009-04-02 at 10:49 +0200, Kilian CAVALOTTI wrote:
> On Wednesday 01 April 2009 17:48:56 you wrote:
> > A point I
> > didn't make in my blog post on this stuff is that the academic side,
> > while a nice market to play in, isn't terribly profitable. You have to
> > give "killer" discounts
http://www.datacenterknowledge.com/archives/2009/04/01/efficient-ups-aids-googles-extreme-pue/
Efficient UPS Aids Google’s Extreme PUE
April 1st, 2009 : Rich Miller
Google continues to improve its energy efficiency, and is telling the
industry how it’s doing it. After years of secrecy surroun
On Wednesday 01 April 2009 17:48:56 you wrote:
> A point I
> didn't make in my blog post on this stuff is that the academic side,
> while a nice market to play in, isn't terribly profitable. You have to
> give "killer" discounts to play in many cases. Which in the case of a
> low margin commodity
On Wednesday 01 April 2009 16:19:57 John Hearns wrote:
> > There are a few services/integration HPC
> > companies in the EU, but not any that I'm aware of selling their own
> > hardware, as Scalable or Penguin do.
>
> ? I know that both Streamline and Clustervision will do you a
> Supermicro bu
> Hmm... I take issue with "Irix sucked" ... then again I am biased as I
> learned Irix right after Unicos and Sunos. Compared to either of those,
> Irix (from a user perspective) definitely did not suck.
I have to agree. I think IRIX was my fourth UNIX (if you count XENIX as an
UNIX var
69 matches
Mail list logo
