This week we look at some preliminary HPC econimy survey numbers and intoduce a chance for you to smackdown the HPC experts
I was determined not to continue my “Hitting The …” theme this week, but I could not. I promise next week to bring a more festive holiday message. (And no, it will not be called “Hitting The Bottle”). What do I mean hitting the experts? I’ll get to that in minute, first I want to follow-up on last weeks survey. For those too lazy to click over to last weeks column, here are the results (as of 2PM Eastern Time, Tuesday December 16th). In response to the question
How has the recent economic events effected your HPC budget plans for 2009?
- No effect (40%)
- Pushing back major HPC acquisitions (19%)
- Cutting back on people (14%)
- Cutting back on maintenance hardware/software (12%)
- Medium across the board cut back (10%)
- Large across the board cut back (2%)
- Small across the board cut back (2%)
There have been 42 votes cast so far (You can still vote. As it is not too late to skew my analysis). I prefer at least 100 votes in ad-hoc surveys, but beggars cannot be choosy. Let’s take a look at the results. The first number that stands out are the 40% of the people who said “No Effect.” In this era of economic doom and gloom, I suppose that is good news. If you multiply the previous stated IDC growth rate of 9.25% by 40%, you get 3.7% growth. I consider that good news as any growth in this economy has got to be cause for celebration. Of course, I’m basing this on a tiny survey and limited knowledge of the respondents, so be careful how much you trust my analysis — probably about as far as two or three of you could throw me. What about the other respondents. It seem the other 60% are experiencing some changes with the largest push back coming on major HPC acquisitions. That new cluster may have to wait a bit. The cut back on people is disheartening. I have long lamented the need to for qualified people in HPC and I believe that this will further constrict market growth when the economy gets back in gear. Cutting on maintenance expenses seem to be logical thing to do. As far as the across the board cuts, I suppose that is very industry specific. Overall, I would say that the market is going to get beat up a bit, but overall it should be able to stand on it’s own. No bail-out needed here.
As for comments, well there were none. I did receive an email from someone who wished to be anonymous. The comment was one of fortune rather than policy:
The economic downturn hasn’t hurt my HPC plans, but only because we just bought our cluster before the downturn, and don’t plan on buying another for at least 3-4 years. Yes, we just beat the economy to the punch! Since we are funded through an endowment that is invested in the market, there is concern about it shrinking as a result of the economy.
Well, there you have it, this weeks snapshot of the HPC economy. Let’s not forget however, this is arm chair advice, the best advice comes from economists who seem to be offering much more sage advice, to wit, Who the hell knows.
Moving on, we have a new feature. We are fortunate to have access to some high powered HPC mavens from IBM and Intel. While you are at it, you can check out the resource center sponsored by IBM and Intel for some fine HPC cluster content. We have been posting soem good content over the last year. Check it out. Back to mavens. Since we had all these smart people, I thought it would be great to engage them in some form of civilized cluster/HPC discussions. Then it it hit me. Why not introduce a topic each week and allow the experts to post an opinion or two. And while we were at it, let’s allow registered users to respond as well. It sounds like a great place for calm measured discourse, which is why I decided to call it HPC Smackdown. Kind of fits don’t you think. Check it out and don’t forget to add your opinion. I may even jump into the mix. This week we are addressing “Are Multi-cores really bad for HPC?” Let’s keep it clean and on topic, no biting, hair pulling, or chairs unless you submit a video.