I'm wondering how much additional hard disk use I should expect from running BOINC. Any broad guidelines or generalizations will be appreciated.

Thank you!

---

CPDN will take a minimum of 450-500MB whilst working on a model (model sizes are about 350MB each!). CPDN will store it's results for each WU both completed or failed until you manually deleted the directory, so you need to keep this in check or disk space requirements will go through the roof!! What I do is delete failed WU results directories and write a full CD-R when I have enough successful results to fill one.

PP@H will take anything up to about 20-25MB (depending on the size of your WU cache). SETI about 5-10MB (again this depends on your work cache). LHC about 20MB and the same for Einstein.

Hope this helps.

---

<img src="http://boinc.mundayweb.com/one/stats.php?userID=444&amp;trans=off">

Thanks for the information! This helps a good deal.

I have a couple follow-on questions, if you don't mind:

-- How can you tell which directories contain failed results? Does one of the UIs provide a straightforward way?

-- How often does BOINC write to/read from the HD during "normal" use? Should I expect any slowdown in disk access?

Thanks again!

==========
&gt; CPDN will take a minimum of 450-500MB whilst working on a model (model sizes
&gt; are about 350MB each!). CPDN will store it's results for each WU both
&gt; completed or failed until you manually deleted the directory, so you need to
&gt; keep this in check or disk space requirements will go through the roof!! What
&gt; I do is delete failed WU results directories and write a full CD-R when I have
&gt; enough successful results to fill one.
&gt;
&gt; PP@H will take anything up to about 20-25MB (depending on the size of your WU
&gt; cache). SETI about 5-10MB (again this depends on your work cache). LHC about
&gt; 20MB and the same for Einstein.
&gt;
&gt; Hope this helps.
&gt;
&gt;
&gt;

---

&gt; -- How can you tell which directories contain failed results? Does one of the
&gt; UIs provide a straightforward way?

You can tell which WUs you are working on currently either from the BOINC work tab or from the 'My account' page of the website. The name of the WU corresponds with that of the relevant project folder. The easiest way of sorting out the old ones is probably by size - completed WUs condense down to over 300MB of zipped files. Anything with less belongs either to a current WU or a failed one.

&gt; -- How often does BOINC write to/read from the HD during "normal" use? Should
&gt; I expect any slowdown in disk access?

The model is saved every 144 timesteps, or three model days. One timestep takes 4.5 seconds on your machine, so saving will take place every 11 minutes. There is not a great deal of data saved at these intervals, but it could be an issue for old drives with a tiny cache and running at ATA-33 (eg those sharing an IDE cable with a CD drive).

There is more significant activity at the end of each of the three phases, and of course in the post processing phase - sorting out and zipping over 600MB is a significant task - but you won't face that very often!

---