getting 10-days cache inconsistency

[4GHZ_or_bust]

I did a clean install of Windows on my main after experiencing some problem the last few days. Not long after I installed a fresh copy of BOINC and set it up with SETI, i got 10 days cache.

My newer rig that has been running for almost 2 months still haven't gotten more than a couple hours worth. Same setting, tried 2 different versions of BOINC. :confused:

Do BOINC really hate Intel CPU or is there something else? Both computer used the same default profile (Home) and are both showing 10 days cache setting.

[eaglescouter]

Does your log show that the machine has reached the limit on work in progress?

Some of the newer machines (and most of those with strong GPU's) will hit a mathematical limit before they achieve the desired 10 day work cache.

[4GHZ_or_bust]

It just say no work is available. I don't know if it's because my new rig is that powerful but my main is not that much more powerful. On the average I get 22k rac on my AMD rig and 10k rac on my new i7 rig.

[eaglescouter]

No work available: Indicates that the one hour work distribution cache at Berkeley is empty until it is refilled again from the splitter servers. This middleman system is supposed to ensure that no single super user drains the available work because it is metered out more carefully when combined with the increasing time limit on new connections from a host.

No work available can also be caused by Berkeley hardware problems in addition to the weekly multi day project outage.

[4GHZ_or_bust]

Yeah but you'd think one that has BOINC running for a couple months would get 10 days cache at least once while my freshly installed Windows and BOINC not get 10 days cache right away.

[eaglescouter]

Additionally: the 10 day cache is a thing of the past unless you are on an old system (perhaps a non ht pentium). SETI currently has a work limit of 200 units per cpu and and additional xx (don't recall the exact number) per gpu. (I vaguely referred to this as a mathematical limit previously in this thread and should have explained that more completely.)

This hard cap is expected to be lifted at some point, but I have not seen any schedule yet.

Today I am not seeing any new work on my machines, so it is possible that the splitters are out of work until staff gets onsite Monday morning.

[eaglescouter]

Work units are available again!

I just picked up 30 units on one machine.

[4GHZ_or_bust]

I think I got it figured out. I told my computer I would take it apart, send the CPU and mobo to scrap yard, and get AMD CPU instead.

It got 10 days worth as of today. :) Took it 2 months to start behaving :P

I do prefer to have some cached WU in case SETI goes down for more than a few days. I wouldn't want my rig to idle and my room to get cold.

[4GHZ_or_bust]

In just a few days my daily RAC jumped to 40k (and climbing) from the high of 25k before my i7 started behaving. And I still have one more machine to add when it comes in: dual Xeon 5650 (24 cores total)
:comp::comp::comp::comp:

Not long before I get to replace my top 30 avatar with a top 20 avatar :burn: