[Beowulf] [email@example.com: Re: Intel?]
landman at scalableinformatics.com
Thu Jun 9 08:38:14 PDT 2005
Robert G. Brown wrote:
>>However, this machine is an AMD64 based laptop from HP, that they will
>>have to pry from my cold dead fingers. If the choice comes between
>>spending $4000 for a new laptop with about the same features,
>>functionality, and 1/2 the bits as this bad boy at $1500, versus
>>spending $300 for a new OSX to run on this bad boy, guess which one I
>>will do ...
> Hey, Joe -- is it true that the AMD64 laptop will actually burn your lap
> if you hold it while it is running a real computation? That's why I
> didn't get one...
(smell of barbeque)
Actually there is a nice side vent for the fan. Gets hot there under
heavy load, about 55C exhaust. Worst I ever cranked it to was about 68C
running an electronic structure calc, and that was pounding on the disk
It runs quite cool most of the time.
> BTW, why aren't you running e.g. FC on your laptop? I'm running FC3 on
> this, my brand new Dell Latitude D610. It has everything my desktop
> has. I'm still messing with the ipw2200 wireless (intel is supporting a
> linux driver, but I'm having a hard time getting linux to decide whether
> the integrated device is an ordinary NIC or a PCMCIA device that has to
> be "inserted").
I am running XP, SuSE 9.3 as of yesterday, and a small FC3 partition.
FC3 has a 4k stack bit which did nasty things to the wireless drivers.
I am waiting to see if FC4 is good next week.
I have to admit to being pleasantly surprised by SuSE 9.3. 9.2 on one
of my other systems bit the dust in such a manner after a crash that it
induced me to switch to Centos 4 (which is pretty good BTW).
I have FC3 on my experimental desktop. I might put it on an Opteron I
am buy/uilding. Still thinking about that.
>>OSX on every laptop should be the goal. Not OSX on Apple laptops.
> Or if they actually make the price right, OSX on every system anywhere.
There appears to be some threading issue with it, so you might not want
it on servers until they fix that. Not compute nodes, but highly
threaded process servers (NFS, web, ...).
> Ultimately this market is going to reject high margin OSs, and MOST high
> margin software. The first real corp to realize this survives, however
> much it will pain them to kiss those huge margins goodbye. Hey, if Sun
> Microsystems had marketed x86 Unix (which they HAD way back in the
> 80's!) for $50 a seat (that is, roughly what MS was charging for >>MS
> DOS<<) in the early 90's, Microsoft would be a tiny company, Apple would
> be out of business, linux would never have been born, and Sun would own
> the universe. OS-2, WinXX, Apple OS all would have disappeared like a
> bad dream. If Stephen Jobs had done the same thing with NextOS in the
> early 90's, HE would own the Universe.
Well, Sun is sort of trying to position Java like that. They cannot
decide if it is a platform, an OS, or a language. They are (sort of)
giving it away. Not really though, as I cannot run JDK 5 on my Itanium2
box. Or a Cray SV1. Oddly enough, I can run Perl/Python/... on them.
I don't think Sun could market to the masses effectively, their cost
structures are wrong, as are their focus. Right now they push Solaris
10 everywhere. Not a bad idea if you like Solaris 10. But I simply do
not see ISVs porting to it enmasse, nor do I see device driver folks
writing to it enmasse, nor do I see end users requesting it enmasse.
IMO that game is over. Linux has won (at least I have not had any
customers ask me for Solaris 10 clusters as of yet, could happen, but I
don't think it is likely).
> MS has a monopoly and can charge 2x to 3x what an OS "should" cost. Its
> competitors cannot. People will pay $30 for an OS/software distribution
> and not even think about it. At $100 up a seat, they start to think.
> At $100/seat PLUS another $500 for all the software you need to actually
> make your $600 computer run (Microsoft's dream is to ULTIMATELY charge
> as much for software as a customer pays for hardware, right?) they
> really feel the burn when the money comes out of their own pocket.
Yup. This is part of the reason I was asking so many questions a while
ago on price points for high performance bits (hardware, software,
otherwise). The cell processor bits on a card in a PCI-e slot: what
would a person pay for hardware like that (if it gave their application
a 10-100x speedup)? Would they pay 10k$ (2 high end compute nodes today
for 10-100x faster performance)? Would they pay $5000? Same question
for software. If you have software which makes your life dramatically
better, what are you willing to pay for it?
I don't think there is anything wrong with 100-500$/seat for really good
high quality software that strongly positively impacts your ability to
get work done (much like the vaporous hardware above). I am not sure I
would spend more than 500$ unless there was a really good business case.
And this gets to a point you have made in the past. How do software
vendors price things for clusters? Per seat will be painful for
consumers. Per user will be less painful, but also reduce usage. Per
cluster is probably a better compromise (consider the cluster to be one
machine, and the compute nodes are not independent machines). Of
course, this doesn't fit in most current software vendors pricing models.
>>Actually, for the past decade, IBM has been slowly transmogrifying into
>>a services company with a small hardware division... :( There is
>>little hard research being done at T.J. Watson anymore (and that is one
>>cool building to go to work in, so it is a bitter shame).
> This sort of transmogrification is why IBM has outlived most of its
> competitors by a factor of two or more. Apple should use them as a
> model. DEC should have USED them as a model. SGI too. NeXT three.
> Diversify. Evolve. Change. It's what the IT marketplace is all about.
> Anticipate the future and alter your corporation to MEET it head on, not
> get left as a vulture-picked bit of road-kill crushed into the pavement
> by the onrushing tires of what is coming. Metaphorically speaking, of
:) Sink or swim. Its all about how you respond to evolutionary
pressure (read as business model changes, structural economic changes).
IBM will likely be here for another several decades/centuries. Their
model changes as times change. They used to build deli meat scales.
They adapt, they move on.
Others who do not adapt will likely die off.
I hope no one writes of Microsoft, as they have shown a remarkable
propensity to adapt. Their model is under assault, and the economy
appears to be changing in such a way to disfavor their business model.
Doesn't mean they are going out of business, but I do expect them to
That said, I haven't seen/heard about the economics of the windows HPC
solution. If it is like their desktop system model (price per seat,
limited number of client connectivity), it is designed to fail.
Joseph Landman, Ph.D
Founder and CEO
Scalable Informatics LLC,
email: landman at scalableinformatics.com
web : http://www.scalableinformatics.com
phone: +1 734 786 8423
fax : +1 734 786 8452
cell : +1 734 612 4615
More information about the Beowulf