Extending rivalries to HPC

In October, Indiana University announced it would purchase a Cray XK7 named “Big Red II”. With a theoretical peak of just over 1 petaFLOPS, it would be the fastest University-owned (not associated with a national center) cluster. Of course, in state rivals would never let that stand. In the latest Top 500 list, unveiled at at the International Supercomputing Conference, Big Red II ranks a very respectable 46th. Unfortunately for them, Purdue University’s new Conte cluster checked in at 28. Oops! Let’s compare:

Cluster Cost Theoretical performance LINPACK performance Cost per benchmarked TFLOPS
Big Red II $7.5 million 1000.6 TFLOPS 597.4 TFLOPS $12.55k / TFLOPS
Conte $4.3 million 1341.1 TFLOPS 943.4 TFLOPS $4.56k / TFLOPS
Comparison 57.33% 134.03% 157.92% 36.33%

It’s clear that Conte is the winner in performance and cost. But what about value? Both of these clusters have accelerators, Big Red II uses Nvidia GPUs and Conte uses Intel’s Phi (which also powers China’s new Tianhe-2, far and away the fastest cluster in the world). Using the GPU requires writing code in the CUDA language, whereas Phi will run native x86 code. This lowers the barrier to entry for users on Phi, but GPUs seem to win in most benchmarks. This would seem to increase the cost of providing user support, but it may be that IU’s users are already prepared to run on the GPU. All of the performance numbers in the world won’t matter if the clusters aren’t used, and only time will tell which cluster provides a better value. What may end up being a more interesting result is the political ramifications. Will the statehouse be okay with the two main state universities both running expensive high performance computing resources? If not, who will get to carry on? Both institutions have a record of success. Indiana ranked as high as #23 on the June 2006 list, but Big Red II is the first Top 500 system there since November 2009. Meanwhile, Purdue has had at least one system (and as many as three) on every list since November 2008. With Conte and the additional clusters in operation, Purdue has much greater capacity, but that doesn’t mean that IU’s system is a waste. I suspect that as long as both universities are bringing in enough grant money to justify the cost of their clusters, nobody in Indianapolis will care to put a stop to this rivalry. In the meantime, it appears that Purdue will remain the dominant HPC power in the state, as on the football field.

Big news!

There’s big news in Fiasco Land. I don’t just mean Mario Marathon 6 (raising money for Child’s Play Charity) which starts next week. Beginning July 1, I will be a Senior Support Engineer at Cycle Computing. That’s right, after nearly seven years of professional employment at my alma mater, I’m leaving the public sector.

I won’t be leaving Purdue entirely, though. Because my new position is telecommutable, I’ll be able to finish my last semester of classes for my masters degree (which just got a lot more expensive). I wasn’t particularly looking to get away, but this opportunity seems to come at the right time for me.

So what does this mean for this blog? Well, I’ll either be writing more or less, depending on how much work I have on my plate. Expect to see high throughput computing become more prevalent, to the degree I am allowed to talk about what I’m working on. Other than that, it’s business as usual (for now).

Trying various razor blades

Note: the links in this post go to Amazon.com with an affiliate code. If you purchase from these links, Amazon gives me a tiny cut of the revenue. I am otherwise uncompensated for this post.

About two years ago I switched from using a Gillette Mach 3 to a double-edged safety razor. When I first bought my razor, I just bought a pack of whatever blades seemed like a good value. When that supply began to run low, I thought it would be a good idea to shop around a bit. My friend Andy had a sampler of a variety of blades, so he traded me a few of his. Thus begin several weeks of not-at-all-scientific testing.

The blades I tested with.

The blades I tested with.

Personna

I had originally purchased Personna blades, which come in a handy container that stores your used blades for safe disposal. The shave was alright, but I found it tended to irritate my neck quite a bit, especially from the second shave onward. This is what inspired my quest for a better shave.

Shark

My first new blade was the Shark. While I didn’t have the problem this guy had, I was definitely a little more nicked than normal. There was no irritation, though. Subsequent shave attempts with the Shark proved much better.

Astra

The Astra blades gave a pretty nice shave. There was nothing particularly memorable about them, but I recall liking the experience. I should have taken notes.

Durablade 7 A.M. Plus

Initially, I didn’t notice these blades since the color of the wrapper was very similar to the other Durablades (below). I discovered them after I had already ordered more blades. The shave was great: very smooth, with no irritation or nicking.

“Israeli Personna”

This unlabeled blade, which is known as the “Israeli Personna“is one of Andy’s favorites. There is no universal “best” blade, and this is a prime example. The blade felt like it was too short, and it took several passes to get a shave that was sufficiently smooth.

Durablade Sharp

This blade was so bad that I had to force myself to use it more than once. It pulled fiercely. I hated it.

The verdict

I ended up buying 100 Shark blades, which will probably end up being about a two-year supply. Had I found the 7A.M.s before I ordered the Sharks, I would have had to flip a coin or something. The Astras would have been a good choice, too. The others I probably won’t use again.

The downside to the Sharks is that they come in cardboard boxes, which left me needing some way to dispose of used blades. I found an empty Arizona tea cannister (metal) in the garage. A little bit of JB Weld to keep the lid on and a small slit cut with the Dremel and I now have many years’ worth of safe blade storage.

Student speech rights

To continue the legal theme from a few days ago (with the addition of some “old news is so exciting!”), a high school in Kansas suspended the senior class president for comments he made on Twitter. What did he say? ““Heights U” is equivalent to WSU’s football team“. WSU’s football team doesn’t exist. That’s it. For that, the school deemed his initial tweet and responses were disruptive to the school.

It’s not clear to me if the Heights High School is acting in accordance with legal precedent (their decision is certainly unjust, but that’s another matter). The Supreme Court has affirmed and re-affirmed restrictions on the free speech rights of students. Bethel School District v. Fraser, Hazelwood v. Kuhlmeier, and Morse v. Frederick have all served to limit what students can say.

In Tinker v. Des Moines, the Court protected non-disruptive political speech, with the disruption being the critical factor. In Bethel, Hazelwood, and Morse the speech in question was part of a school-sanctioned activity even if the activity was not on school grounds (as in Morse). It would be a great stretch to consider Mr. Teague’s Twitter account to be a school-sanctioned activity, as it appears to be his personal account. To my knowledge, no Supreme Court ruling has ever addressed a school’s ability to restrict speech that occurs outside of school events.

Arguably, the concept of in loco parentis could be used to support the ability of schools to respond to behavior that happens outside the school. I don’t agree with this, but it would be interesting to see how this argument played out in the courts. In the meantime, I expect that this may end up being discussed in court rooms for years to come. If no suit is filed, it should at least be used as an exercise in high school government classes across the country.

Remembering Tim Samaras

I woke up this morning to learn that veteran tornado researcher and storm chaser Tim Samaras and two others were killed by a tornado near El Reno, Oklahoma. I never knew Tim in person, but I had the pleasure of interacting with him on the wx-chase mailing list and on the Stormtrack forum. Tim was of the old breed of chasers: safety-conscious, focused, and a serious scientist. This makes his death all the more jarring; Tim Samaras is about the last person you’d expect to die in a tornado.

That’s why this is so upsetting for me. I’ve always held to the belief that chasers are safe so long as they’re not stupid. I don’t know what happened in those last minutes, but it’s safe to say Tim was not being stupid. Did he make a mistake? Did he lose situational awareness? Was this a completely unavoidable accident? I can’t answer any of these, which means I’m face-to-face with the lethality of my sometimes-hobby.

To my knowledge, Tim and his companions are the first people to die while actively chasing. The other deaths that I’m aware of were due to roadway accidents on the drive home. That nobody has been killed is a surprise in itself given some of the crazy antics of those who have taken up the hobby inspired by “Twister” or Discovery’s “Storm Chasers”.

Tim can no longer contribute to the scientific study of tornadoes. Perhaps his death will serve to remind us all that even the best are vulnerable.