Exactly! they have to run it at *least* two times.
If they dont get the same answer, they have to
make more runs.
My memory (128MB PC133 DIMMs from crucial, this was several years ago..) had an average of around.01428 errors per day per gigabyte. If bigmac's memory is ten times better than mine, they should see an error on the average of once every 3.82 hours. If their memory is one hundred times better, then the interval would be 1.59 days on the average.
Suppose Apple has this 100 times better memory
that will fail every 2 days.
How many times in avarage will they have to run
a computation that takes a week? A month?
From personal experience the hardware on our Sun boxes at school has little problem. The macs and the PCs have much more problems.
Now, the harddrives are the same used in PC:s, if you think these drives are near the quality of high end scsi drives then I think you are wrong. I did *not* say 'cheap and crappy' I said cheap. This is a supercomputer, and I think it is worth having great uptimes.
Having 9 (right?) fans per node, and having so many nodes, many will break.
When running on thusands of machines for weeks, both runnings will most likely not be the same, thus you will have to do them *many* times until the result is the same.
When running on thusands of machines for weeks, both runnings will most likely not be the same, thus you will have to do them *many* times until the result is the same.
more than that considering some of the building
infrastructure are in that figure
The reverse statement is more true.
Its a good bet too that this thing is going to have lower maintainence costs and higher up-time given the macs attention to cooling, the use of high quality hard drives and power supplies, and high end memory chips. (on our cluster a tenth that size we blew 60 hard drives in the first 6 months and had to replace 10% of the motherboards.
I would guess that they will have more problems with the mac hardware. To many fans, do you know how often those fail? Cheap hard drives, and
no ECC, come on, you can not compare the memory
of the mac with ECC memory.
Also, the VirginiaTech cluster is the only "self-made" supercomputer in the Top50
And I guess the only one not using ECC memory.
This is a *major* problem. Doing computations
on many nodes during a long period *will* cause
memory errors. This of course is both cheeper
and also a bit faster (ECC has a performance hit).
You will though not be able to do long computations.
And short comutations must be run at least two
times in order to check the results
The GPL (and the LGPL?) GCC EMACS GDB parts of GNU Make parts of GNU ls parts of the manual GMP
But I guess also a lot of the basic GNU programs. I know he has done some small hacks on Posix conformance of Bison and that he also was involved with the development of the atomake and autoconf programs.
I do not think he has done much work on the C library, but I may be wrong.
This is why USA do not want a definition on terrorism.
The difference between a "freedom fighter" and a terrorist does not exist.
Its funny looking at Rambo III when he helps the taliban freedom fighters against the "non belivers" of Soviet Russia. Giving them good american missiles and training.
Re:What about those of us
on
CNet on WinFS
·
· Score: 1
I guess that NTFS works like unix inode based filesystems, i.e. a directory is a file containing all the files and subdirs in the dir.
In that case you are wrong. I think you are. The speed increase will be when you want to list all your mp3:s. Or all office documents made by author John.
Slashdot Mirror
Yes, that would of course be the "solution"
to the problem. Would not like to be the person
doing these hacks!!!
Yeah, thanks!
Interesting, do you know what the chance is
that a job running on all nodes for a week
will terminate with no errors?
What I have read these errors are common.
Yes, I think ECC can find two-bit errors and
correct 1 bit errors.
Normal mem modules do not have any parity
(what I know of).
I do not agree.
Most important, what has this to do what I said?
From personal experience the hardware on
our Sun boxes at school has little problem.
The macs and the PCs have much more problems.
Now, the harddrives are the same used in PC:s,
if you think these drives are near the quality
of high end scsi drives then I think you are wrong. I did *not* say 'cheap and crappy' I
said cheap. This is a supercomputer, and I
think it is worth having great uptimes.
Having 9 (right?) fans per node, and having so
many nodes, many will break.
Yes, that is what I sugested.
Running things twice is *not* low-overhead.
When running on thusands of machines for weeks,
both runnings will most likely not be the same, thus
you will have to do them *many* times until the result is the same.
See the thread about ECC on aceshardware.com.
Running things twice is *not* low-overhead.
When running on thusands of machines for weeks,
both runnings will most likely not be the same, thus
you will have to do them *many* times until the
result is the same.
See the thread about ECC on aceshardware.com.
I have made this smart software that makes my
calculator solve NP hard problems in no time,
it is realy great, it has error recovery built-in.
Do you not trust me? why?
Then, as you have read it, you can tell the world how you do this "low-overhead software error correction to compensate for the non-ECC memory".
I am *very* interested in it!
Yeah, and GNU/Solaris.
No.
GNU is Gnu userland on top of the Hurd multiserver
using L4 or Mach as microkernel.
GNU/Linux is GNU userland on top of Linux kernel.
The OS most people use, and many calling Linux.
GNU/Linux is a *version* of the GNU OS.
Linux is the kernel originaly created by Linus
Torvalds. Often run with GNU userland. If run
with BSD userland better called BSD/Linux.
From my head!
The GPL (and the LGPL?)
GCC
EMACS
GDB
parts of GNU Make
parts of GNU ls
parts of the manual GMP
But I guess also a lot of the basic GNU programs.
I know he has done some small hacks on Posix
conformance of Bison and that he also was
involved with the development of the atomake
and autoconf programs.
I do not think he has done much work on the
C library, but I may be wrong.
This is why USA do not want a
definition on terrorism.
The difference between a "freedom fighter"
and a terrorist does not exist.
Its funny looking at Rambo III when he helps the
taliban freedom fighters against the "non belivers"
of Soviet Russia. Giving them good american missiles
and training.
I guess that NTFS works like unix inode based
filesystems, i.e. a directory is a file containing
all the files and subdirs in the dir.
In that case you are wrong. I think you are.
The speed increase will be when you want to list
all your mp3:s. Or all office documents made by
author John.
Yes, it is MS SQL Server (a lite version of it)
on top of NTFS, i.e. the filesystem will be the
same.
It is more like a service on top of the filesystem.
I think GNOME is developing something similar.
>> Or coule it be a Power5? Yes I think it is. I mean the Power4 isnt 64bit. Atleast be correct in your stupidity.
Power4 is 64 bit.
Thats not true, it is basicly a Power4 with
less cache, and an altivec unit.
I can hardly imagine that the gui code Apple
wrote is beeing used in the cluster.
from:
h tm l
Yes! but this is not 640k!
http://pages.prodigy.net/jhonig/bignum/qaearth.
"It is extremely unlikely that the total number of atoms in the earth is higher than 10^51 or lower than 10^48"
So we have less than 2^78 atoms.
I belive I can rest in peace before proven wrong.
We allready have 128 bit machines (SIMD) and
we will not need more than 64 bit address space.
sizeof(void*) == 4
I would say that the Athlon kills the G5 in SPEC.
What benchmark would you use? That is fair and
runns on both platforms?
I would say that it is worse for the AthlonFX
to run with half it registers than the G5 running
classic.