Do SETI Right

The whole point of SETI is to get some valuable work done using your computer. While it
is set up in a way to encourage friendly competition, the work is what’s important, not the
competition.

There is absolutely no point to generating work that is worthless to SETI. That is just a waste
of everybody’s time and energy.

Just because your machine seems stable doesn’t mean it runs SETI right. The processor can be making subtle errors all the time without crashing. You may think you’re generating work units
when all you’re doing is generating worthless units.

Let me give you a quick test:

Q1: What is better, to generate a correct work unit in six hours, or an incorrect, worthless work unit in five-and-a-half hours?

A) Generate a correct one in six hours
B) Generate an incorrect one in five-and-a-half hours because I get credit for more work units, and that’s all that matters.

Q2: What is better, to check to make sure your overclocked machine works right, or not?

A) It is better to check.

B) It is better not to check. This is my machine. It can’t make a mistake.

If you answered A to both questions, congratulations, you are at least reasonably sane when it comes to this human endeavor.

If you answered B to either or both questions:

PC Version: You face moderate to severe mental, motivational and/or emotional challenges.
or
Non-PC Version: You are an asshole (a complete flaming one if you answered both questions “B”).

I Don’t Want To Be . . . uhh . . . Challenged. How Can I Tell What My Machine Is Doing

I received two emails from David Maza from Ars Technica Team Lamb Chop.
He wanted me to say upfront that the method described in his email was devised by Roelof and the Team Lamb Chop SETI team, and that they welcome all SETI users, including non-LambChop members, to submit benchmarks.

From: “Maza, David” [email protected]
To: [email protected]
Sent: Thursday, July 12, 2001 1:42 PM
Subject: Seti and overclocking

. . . .

I think it’s our responsibility to make sure our machines are generating
good results. One method implemented by the folks at Teamlambchop.com can
be modified slightly and used by almost anyone. Here’s what I propose:

1. Download a workunit and save a copy to your hard disk, somewhere other
   than your SETI directory.

2. Complete the workunit at a known-stable, preferably non-overclocked
   speed.

3. Save a copy of your result.sah file to a non-seti directory.
4. Re-process the saved work unit at your desired overclocked speed.

5. Compare your overclocked result.sah to the non-overclocked copy using a
   file-compare utility like Beyond Compare.

If there are any differences between the two, most likely the overclocked
config is not stable enough and you are generating “garbage”. Slow it down a
bit!

If you are a Serious SETI man, you really ought to do this right away and give your machine a clean bill of health. You should also
rechack every once in a while, after all, you don’t go to the doctor’s once in your life for a checkup, either. Some good times to
check might be if your work area gets hot as summer approaches, or if your machine starts acting funny doing other things.

Don’t be challenged. Your machine works hard enough. Make it count.

Thanks to TeamLambChop and the SETI folks for setting everything straight.

Email Ed