Windows Breaks Into Supercomputer Top 10 294
yanx0016 writes "Wow, that's some news this week at SuperComputing 08. Apparently Microsoft Windows HPC Server 2008, with a Chinese hardware OEM (Dawning), made #10 on the Top500 list, edging out #11 by only 600 Gflops. Folks were shocked to see Microsoft getting so serious around HPC; I think we are only beginning to see a glimpse of Microsoft in the HPC field."
Yeah, mut how much useful stuff is happening? (Score:4, Insightful)
Re:Retarded (Score:2, Insightful)
I seriously doubt that anyone with enough to spend on a top 10 supercomputer is worried about the Windows tax.
Off topic, but I have to mention it (Score:5, Insightful)
edging out #11 by only 600 Gflops
Emphasis mine.
Maybe I'm suffering from a case of advancing years, but I couldn't help but be amazed by this metric. These days it is indeed small, but another part of me remembers being a fifteen year old kid amazed at how absolutely great his C64 was.
I wonder exactly how many years a C64 would have to run to make up a single seconds worth of that difference. How long would a C64 have to run to perform 600 Gflop? How long would every single C64 ever made have to run? I wonder.
You'd have to run some integer-only 6502 IEEE floating point library or something like that to figure out how long a single floating point operation would take on the C64. Then multiply by 600G.
Would it be a few years? A few millenia? Blue-green algae?
Re:Retarded (Score:2, Insightful)
.
Development tools. Something Microsoft is very, very, good at.
And missing from the summary is this little note: Just a year ago, the best Microsoft could do was 116th place based on rankings from Top500.org, which has been benchmarking supercomputers since 1993 with its bi-annual tests it calls "runs."
Re:Off topic, but I have to mention it (Score:4, Insightful)
Does not compute. M$ is not for HPC. (Score:-1, Insightful)
How can you combine "familiar" and "roll out new language" in the same breath? Even if someone is familiar with all the quirks and shortcomings of Windows programming, why would they risk significant money and time on M$'s next vaporware?
How can a new language or familiarity offset the massive difference in cost and hassle when it comes to building the thing and adding new nodes? I mean really, do you want to write you next grant around how many Windoze licenses you think you can afford rather than what you think you need to get the job done?
Re:Retarded (Score:5, Insightful)
Flagship demo projects like this often get exceedingly big discounts from the vendors.
Re:Does not compute. M$ is not for HPC. (Score:2, Insightful)
Re:there are lots of Windows developers out there. (Score:5, Insightful)
It is possible, even if not entirely likely.
Developing a language and compiler that takes advantage of multiple CPUs (especially if it's scaling the number of CPUs) is something that a lot of research (or money). MS does have this. Whether they use it effectively is another matter.
Also, remember that they are not unfamiliar with HPC abstraction. Direct3D abstracts the architecture of the GPU, and GPUs have been parallel processors for a decade or so.
Re:there are lots of Windows developers out there. (Score:3, Insightful)
Re:there are lots of Windows developers out there. (Score:3, Insightful)
So you're getting your car fixed by the milkman?
Supercomputing and parallel computing are different than building regular apps and websites. Why would you want to get the wrong programmers for the job?
Re:Retarded (Score:2, Insightful)
Re:Potentially bogus (Score:5, Insightful)
Re:Yeah, mut how much useful stuff is happening? (Score:0, Insightful)
Where do you think the overhead is at? GUI, mouse, I/O? Do you think Internet Explorer is dragging it down?
My guess is that when it's running 600 GFLOPS that any "OS pissing" is marginal and doesn't matter one bit at all. Besides, doesn't the benchmark score get reported AFTER all of the OS overhead?
Where is the insight here? Just seems like the usual Microsoft bashing, though this time it doesn't even really make sense.
Re:From the article, pricing is (Score:5, Insightful)
Each node probably has 4 CPUs and 4 cores each, which reduces the price significantly, to only $28 for the commercial version, or about a dollar per node for the academic version.
That's not bad. And of course you don't understand the CALs, but hey, making erroneous statements can get you modded insightful so maybe I should spout something disingenuous about Linux, like it costs $699 to license it from SCO or something.
(For the uninformed, not all CALs are created equally and the parent is assuming that these are named licenses that must be purchased for each user. Many different kinds of CALs exist, and I suspect these are either physical unit licenses or concurrent access licenses, i.e.: you purchase 1 per node, period.)
The OS is very important (Score:5, Insightful)
It is one thing to measure Drystones etc, or some other simple grunt-measuring metric, but that does not realistically stress the OS's influence on how the system will perform on huge complex number crunching models.
Microsoft has only been in this game for a short time and only recently got support for 256 cores. Getting support is one thing, getting **good**, optimised, support is quite another and that will take some time to get right.
Re:Retarded (Score:5, Insightful)
I hate to defend Microsoft, but...
Crap hardware support? Who cares - you're running numerical calculations, not a bloody game on some tossy video card.
Crap vendor support? This vendor will have been given full support by Microsoft, and will be equally supportive of their users.
Performance? They're in the top 10.
Stability? If you're not dealing with odd hardware / crappy drivers, Windows Server versions are actually fairly stable.
Why not run your compute nodes under Windows?
You can actually run Windows Server 2000 and above headless, removing any GUI overhead - so why not?
I still agree that on any particular hardware configuration, Linux or another *nix will likely be faster, but your experience of desktop applications doesn't necessarily translate to HPC.
Re:there are lots of Windows developers out there. (Score:2, Insightful)
No, he very clearly gave an analogy, using Direct3D to show that similar (albeit on a far smaller scale) parallelization problems have already been solved, that this is not completely unknown territory. One could also infer from this analogy that the only real difference that might be a major obstacle is scale.
My question to you, and to folks who ask questions like the one you just asked, is what happened to your reading comprehension? Many people ask questions like this that have very obvious answers which should be apparent to anyone who actually read the comment in question. The ability to know that this is the case is a very valuable thing; did you ever think to mentally reclaim it from the public school system (or whomever) that took it away from you?
You could get upset with me for asking you that, maybe you can throw a few insults my way or maybe the moderators can play the knight-in-shining-armor and mod me down to -1 for it so that they can feel better about themselves, but that would constitute an evasion of the issue I am raising and I believe you know it. I'm sorry to appear to single you out -- I actually see a lot of this and you happened to be the first one I saw today. May you have the patience and forebearance to understand what I am really saying instead of getting on your high horse and reacting to a perceived insult.
Re:Yeah, mut how much useful stuff is happening? (Score:5, Insightful)
The FLOPS are measured by a benchmark program, The Linpack Benchmark [top500.org] that runs under OS overhead, so it would differ with different OSes and probably different configuration parameters in the same OS. I wouldn't be surprized if MS hadn't supplied significant engineering support to get the system tweeked to the T to nail down good numbers on the benchmark suite.
Re:Yeah, mut how much useful stuff is happening? (Score:3, Insightful)
Where do you think the overhead is at? GUI, mouse, I/O? Do you think Internet Explorer is dragging it down?
File system would be my guess. Those context menus are pretty instant, the population with all those headers is what takes the time. If you had a shortcut method (say, analogous to VMS' old "Install" header cache forcing mechanism) for locating and ID'ing the items in the menu list it would probably help. As it is, I think it's lumped in with the "file index" capability on Windows that everybody turns off to cut back on the annoying disk activity. Just my A$0.02
Re:Off topic, but I have to mention it (Score:3, Insightful)
Another scary thing: I clocked the GPU in my desktop at 450 GFlops yesterday (nbody crap). Admittedly, single precision, but still.
Re:Does not compute. M$ is not for HPC. (Score:5, Insightful)
What is wrong with " M$ "?
Nothing is wrong with "M$", in the same sense that nothing is wrong with someone referring to Linux as "linsux" and open-source as "open-sores". The thing is, it tends to make you look somewhat immature.
If you can present a compelling coherent argument, you don't need to use lame decade old snipes about whatever subject matter you are discussing. If you use them in a compelling argument, it usually just makes the people you are out to persuade have a lesser opinion of what you wrote, and thus, you have sacrificed persuasive power.
It comes down to maturity for the most part and just simply putting forward a good argument.
Re:Retarded (Score:5, Insightful)
Development tools. Something Microsoft is very, very, good at.
Microsoft development tools are in the category "If this helps you, you are not qualified for this job to begin with". An equivalent would be multiplication table on mathematician's desk or marathon runner on crutches.
Re:there are lots of Windows developers out there. (Score:3, Insightful)
Problem is they've missed the boat. Windows already has all the tools needed for desktop computing
Look at this chart..
http://marketshare.hitslink.com/report.aspx?qprid=8 [hitslink.com]
Windows has 90% of the market
Linux has less than 1%
Which Operating System do you think is going to have better tools for desktop computing?
__________________________
For all of you angry mods... i know it's a flawed analogy (and it's not even about cars...), but the parent was saying Microsoft should not even try since they're not at the top of the game. I think it's a pretty stupid way to think. Let them all try and develop new things. Someone will eventually come out with something that makes parralel computing easier and we'll all gain something from it even if it come from MS.