Whither SGI

Obviously SGI has existential challenges ahead. This means something quite simple. No cow is sacred. No ego’s can get in the way of doing the right thing by the shareholders.
I will be frank. This is about 9 years too late.
First, will SGI recover? Possibly, though I am not going to bet on it. The challenges are not just internal, their competitors have wasted no time in making use of their situation. High performance computing is an unforgiving market. Single mis-steps can be fatal. Running a marathon of past mis-steps like SGI did should be fatal. I wouldn’t count them out, but the odds are against them regaining good will and customer trust.

Read moreWhither SGI

Something cluster-like this way comes

It appears that Microsoft is about ready to release CCS. This might be interesting, depending upon what was done, and how it all works. Some things to note. It includes a job scheduler. Built into the OS. This is either a really good thing, or a really bad thing. I can see arguments both ways. It includes bits that Linux has had for a while. Remotable/scriptable installation. Multiple security models.
Some thoughts. The Platform folks in Toronto cannot be happy about the integrated scheduler. Microsoft is firing across the bow of them, Altair, and others. We will have to see how interested the customer base is in this.

Read moreSomething cluster-like this way comes

Interesting NUMA issues in current SuSE kernel

One of our development systems is a dual socket system with 2 dual core Opteron 275 chips. 4 GB ram, nice disk config, and a quadro fx/1400. This is a good machine t work on.
I had set it up with SuSE 9.x, and had left it at 9.3 for quite a while. Recently we upgraded it to SuSE 10.0 Pro. More modern kernel, somewhat updated apps. I thought it would be nice to stay somewhat current.
We use this for building and testing our accelerated HMMer code. This code is 1.6-2.5x faster than the binaries from the WUSTL site. Source code changes, and we are finalizing a new release of it.

Read moreInteresting NUMA issues in current SuSE kernel

SCSI vs FC vs SATA

I have heard this argument come up again several times recently. Lots of folks out there from the enterprise storage realm still love their FC drives. The SCSI crowd like their units. Both handily disparage SATA as being inferior, poorly performing, or with higher failure rates.
This is an interesting point.
As far as I am aware, all the drives come physically off the same manufacturing production line. The only significant difference between the units that I am aware of (modulo newer motors on newer units) are the electronics that connect to the bus.

Read moreSCSI vs FC vs SATA

In the limit, as N(cores) -> infinity …

So way back in the good old days, programming a single core CPU in a high performance manner was a challenge. Compilers promised much and delivered small fractions of maximum theoretical performance. To get nearly optimal performance, you had to hand code assembly language routines. You would never be able to achieve 100% utilization of the processor capabilities, but you might be able to sufficiently balance memory operations with floating point and integer operations so that you were utilizing a sizeable fraction of the chips subsystem capabilities.

Read moreIn the limit, as N(cores) -> infinity …

MEDC update

According to the MEDC site, 505 applications were turned in for mostly commercial efforts. 505… The mind boggles. Of those 505, 139 are commercialization. Another smattering are also commercial, though hidden. Call that 150 commercial ones. In all the previous competitions; the MLSC (Michigan Life Science Corridor), the MTTC (Michigan Tri Technology Corridor), the commercial … Read moreMEDC update