As many of you know by now, there’s a new computer at the tippy top of the just-published Top500 list. I found out a bit earlier than most, via a webcast I was recording at 3:00 a.m. my local time on Sunday. It was a bit of a struggle to record it; their net connection kept breaking, and the sound quality on my side was like a …
But will it run
I know, I know. I'll get me coat.
the questions should be...
How many users playing Crysis simultaneously?
Could it run every copy of Crysis at once?
Either way, its an amazing machine.
I would love to test drive my giga and terapixel image processing routines on that beast. 93% efficiency is particularly neat, and I should consider these sparc processors for our next machine.
The circle is completed
In the very beginning there was no Linux, of course, next the Linux machines where not mentioned at all, or they where called "home made". After several years when Linux passed the 40-50% level the top500 list started to mention the OS too. Today when the level is +90% it is
again, sort of, unnecessary to mention Linux in articles about the top500 computers.
Linux has become a de facto standard in supercomputing.
In embedded devices Linux has probably reached the 40-50% level too, but in those devices the mention of the OS is totally irrelevant and not mentioned.
It will be interesting to see how fast LInux will reach the 40-50% market share in smart phones.
Even now Linux is very seldom mentioned in that market as it is an unnecessary word in marketing.
If you finf this topic interesting then have a look at:
Why mention Linux?
The only reason I mentioned Linux in the article is because the Fujitsu box is using SPARC chips. I know that Linux is the o/s of choice for HPC and particularly Top500. But SPARC is closely tied to Solaris in a lot of reader minds so I wanted to make sure that no one jumped to the conclusion that it was running Solaris. It's not something I'd mention when the system is based on x86 processors.
and it's rubbish
Linux is far, far too big to run on supercomputers. Take a look at some of the research that's been done, you'll see that Linux is a huge, noisy system compared to the other options. The problem is that people think that Linux and MPI (preferably with FORTRAN and Python in the mix) are all we'll ever need for supercomputing, even at exascale.
Re: and it's rubbish
this is why the Blue-Gene here really hasn't much of an OS at all (on the compute nodes that is, the front-end does have an OS). This does generate a new set of problems (OK, challenges) for programmers.
From what I've learned in follow-up calls, they aren't prepared to release Crysis results yet, but are optimistic that the system should be able to run a highly modified version of it (ported to SPARC, additional parallelism, etc.). With luck, they're hoping to be able to host, as they put it, 'most of their Crysis krew at the same time.' I'll let you know what I hear....
My take on the Top500
In the June 2011 list (the 37th TOP500 List) for the first time, all of the top 10 supercomputers demonstrated more than a Petaflop of computing performance (more than a quadrillion calculations per second…). Here is my view on the top 10 systems (the truth, if you can handle the truth….).
Geography: USA with 5 systems, Europe (France) with 1 system, Japan with 2 systems and China with2 systems. Interestingly enough, in the top 5, one system is from the US, and 4 are from Asia. Is Asia taking the lead in HPC? Or will we see an increase in funding to enhance the HPC development in the US?
Vendors: Fujitsu, NUDT, Cray (3 systems), Dawning, HP, SGI, Bull and IBM. Complete diversity.
CPUs: 5 systems use Intel, 3 systems use AMD, 1 system with Sparc and 1 system with Power. 80% use commodity x86 solutions. The new appearance of Sparc in the top 10 is due to the new system made by Fujitsu. Some folks see that as the second Earth simulator – a nice demonstration of capability, no much spread beyond that.
Accelerators: 5 systems use accelerators (GPGPU, Cell) to achieve the desired performance. Interesting trend due to the compute/dense/economical efficiency of the GPGPUs. My prediction is that more GPGPUs will be used (that is a safe bet…. ;-) ), and we will definitely find them as part of off-the-shelf CPUs in the not-to-far future.
Interconnect: 5 systems use Mellanox InfiniBand, 3 systems use Cray proprietary interconnect, 1 system uses NUDT proprietary interconnect and 1 system uses the Fujitsu Tofu proprietary interconnect. Looking in previous lists, InfiniBand as a standard has gain momentum in the top 10 systems. From 3 systems in the top 10, to 4 systems (last list) and to 5 systems in the current list. Win for the standard-base solutions is a win for all of us – most of the high-performance computing systems are based on standard solutions (there are much more than 500 HPC systems in the world you know….) therefore development around standard solution in the high-end of HPC platforms brings better capabilities and feature set to the rest of the HPC arena.
For more info, you are welcome to check my blog - HPC-Opinion - http://hpc-opinion.blogspot.com/.
- Breaking news: Google exec veep in terrifying SKY PLUNGE DRAMA
- Geek's Guide to Britain Kingston's aviation empire: From industry firsts to Airfix heroes
- Analysis Happy 2nd birthday, Windows 8 and Surface: Anatomy of a disaster
- Google CEO Larry Page gives Sundar Pichai keys to the kingdom
- Something for the Weekend, Sir? SKYPE has the HOTS for my NAKED WIFE