Hi!

PC:
400 MHz, 64 Mbytes, 256 Mb swap, linux.
Last one where some ISA card work.

Seti worked fine for months, but yesterday it stopped crunching - no work.

2004-12-18 07:52:34 [SETI@home] Requesting 343113 seconds of work
2004-12-18 07:52:34 [SETI@home] Sending request to scheduler: http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
2004-12-18 07:52:40 [SETI@home] Scheduler RPC to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi succeeded
2004-12-18 07:52:40 [SETI@home] Message from server: No work available (there was work but your computer doesn't have enough memory)
2004-12-18 07:52:40 [SETI@home] Message from server: No work available (there was work but your computer doesn't have enough memory)

free
total used free shared buffers cached
Mem: 62276 56372 5904 0 1104 20396
-/+ buffers/cache: 34872 27404
Swap: 249472 3048 246424

Preferences are not to keep it in the memory, anyhow I do not swap, it is only seti@home.

I tried rebooting, played with preferences, but machine is idle. It just connects, and nothing is returned due to lack of memory. Any hint?

BR
Iztok

Look here

http://setiweb.ssl.berkeley.edu/forum_thread.php?id=7353

---

Semper Eadem
So long Paul, it has been a hell of a ride.

Park your ego's, fire up the computers, Science YES, Credits No.

> Look here
>
> http://setiweb.ssl.berkeley.edu/forum_thread.php?id=7353
>
Thanks!

It is MS Windows Q&A, and I keep myself away from there.
Shortly: new WUs have 64 Mb RAM limit, while older works.
Scanning logs, this was first time noticed on November 28th.
PC managed to get some old WUs occasionally, but now they are gone.

Pitty: RAC of 50 is not ignorable. And there were no errors on results.

Setiteam: please lower RAM limit. Linux does job fine wirh 64 Mbytes.

GL
Iztok

Spend 8 bucks and get another 64megs.

---

SETI spends a lot of it's time moving data into and out of memory, this is why it's so dependant on memory bandwidth. When SETI was originally devised back in 1999, typical computers didn't have enough memory to keep all the data in memory, so this was the solution chosen.

However, current efforts to optimize the code are investigating the option of keeping more data in memory so it may be processed faster and negate the need to keep moving data into and out of memory. This should significantly improve the time taken to process a work unit but will mean that systems with insufficient amount of RAM will no longer be able to process work units.

This may be the reason a limit on system RAM has been introduced. Ben or Hans may be able to comment further on this?

Ned

---

*** My Guide to Compiling Optimised BOINC and SETI Clients ***
*** Download Optimised BOINC and SETI Clients for Linux Here ***

Ned,

Um, was that supposed to be cache? Disk drive?

I think about half the times that you used the word memory you should have used cache or disk drive.

Because the phrase " ... didn't have enough memory to keep all the data in memory ... " does not make sense ...

---

> Because the phrase " ... didn't have enough memory to keep all the data in
> memory ... " does not make sense ...
>
umm...it means that the data has to be stored on the drive. cause there isn't enough memory to hold it.....makes perfect sense...

---