I'm getting WUs from widely varying dates.


How big is the backlog? Over the past couple of weeks, has it been getting bigger or smaller? Are there any automatically updated graphs or something that has this info?

(Sorry for the rapid fire questions)

Server status page shows what is currently being ran. It looks like they are running Tapes that were previously rejected. You can see that there are large parts that are bad. I still think they should pull out the really old Tapes and run the Astropulse on them. Thats at least 7 or 8 years of work to run through

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

Server status page shows what is currently being ran. It looks like they are running Tapes that were previously rejected. You can see that there are large parts that are bad. I still think they should pull out the really old Tapes and run the Astropulse on them. Thats at least 7 or 8 years of work to run through

I believe that IS the plan, once they get the software Radar-blanker going.

Wouldn't they just reuse the tapes? Why buy more? Why keep them in storage?

I am assuming this is something similar to the medical field. You never destroy anything. How would you recheck something if your original raw data is destroyed? I'm guessing the old tapes are in storage somewhere

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

What kind of tapes are these? (I mean in a physical sense)

---

What kind of tapes are these? (I mean in a physical sense)

Back in 2005, they were DLT - see Sunday at the Lab - December 4, 2005, third picture down.

Cobblestones.... that is your credit, right?

So if cobblestones change based on the CPU that created them, then a powerful machine and a crappy machine should produce the same credit, right? Assuming both of them are not running CUDA and only have 1 logical CPU?

Hi Vistro,

They don't use the DLT tapes anymore. They have Sata enclosures and ship the disks back and forth to Arecibo. So the tapes wouldn't get re-used and as was previously mentioned, until there's no DLT drive working that can read them, why get rid of them or the data on them?

Wiser heads will advise on the cobblestones question you have.

regards, Gizbar.

---

A proud GPU User Server Donor!

Cobblestones.... that is your credit, right?

So if cobblestones change based on the CPU that created them, then a powerful machine and a crappy machine should produce the same credit, right? Assuming both of them are not running CUDA and only have 1 logical CPU?

Cobblestones are cobblestones, as defined here.

The problem is that the basic definition depends on two specific benchmarks, and some machines benchmark better than the crunch, and some less-so.

That introduces another variable. For SETI@Home, actual credit is based on counting flops, which is at least in theory indexed to the benchmark * time credit.

It's not perfect, but the problems are all due to the fact that perfect accounting would take more CPU time, and that would take away from crunching.[/url]

---

Cobblestones.... that is your credit, right?

So if cobblestones change based on the CPU that created them, then a powerful machine and a crappy machine should produce the same credit, right? Assuming both of them are not running CUDA and only have 1 logical CPU?

Cobblestones are cobblestones, as defined here.

The problem is that the basic definition depends on two specific benchmarks, and some machines benchmark better than the crunch, and some less-so.

That introduces another variable. For SETI@Home, actual credit is based on counting flops, which is at least in theory indexed to the benchmark * time credit.

It's not perfect, but the problems are all due to the fact that perfect accounting would take more CPU time, and that would take away from crunching.[/url]

ok to make a concise answer from Neds. Yes a slow Computer running the same WU will get the same points/credits/Cobblestones as a high performance machine. The slow machine may take a week or more to do the work and the high performer might finish the same WU in an hour

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

Cobblestones.... that is your credit, right?

So if cobblestones change based on the CPU that created them, then a powerful machine and a crappy machine should produce the same credit, right? Assuming both of them are not running CUDA and only have 1 logical CPU?

Cobblestones are cobblestones, as defined here.

The problem is that the basic definition depends on two specific benchmarks, and some machines benchmark better than the crunch, and some less-so.

That introduces another variable. For SETI@Home, actual credit is based on counting flops, which is at least in theory indexed to the benchmark * time credit.

It's not perfect, but the problems are all due to the fact that perfect accounting would take more CPU time, and that would take away from crunching.[/url]

ok to make a concise answer from Neds. Yes a slow Computer running the same WU will get the same points/credits/Cobblestones as a high performance machine. The slow machine may take a week or more to do the work and the high performer might finish the same WU in an hour

Or, if you wanted an even more concise answer:

No.

---