Problem is that it's too short, The client cuts the string off before the second ']' It only shows the vendor string on websites... the p_model string is not present due to the missing ']' ....

Here is the URL of a 5.8.17 Linux client with an AMD X2 cpu...

http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=372435

Below is the host_info for the client_state.xml... notice that the second ']' gets cut off in <p_model> .... you will notice that it's not just one or two character too long, but a whole bunch.....

<host_info>
<timezone>-18000</timezone>
<domain_name>shared3</domain_name>
<ip_addr>192.168.5.64</ip_addr>
<host_cpid>2871d8130d4426162ed8c4b9af2548f9</host_cpid>
<p_ncpus>2</p_ncpus>
<p_vendor>AuthenticAMD</p_vendor>
<p_model>AMD Athlon(tm) 64 X2 Dual Core Processor 4200+ [Family 15 Model 75 Stepping 2][fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr</p_model>
<p_fpops>2146728022.455104</p_fpops>
<p_iops>4997074464.835873</p_iops>
<p_membw>1000000000.000000</p_membw>
<p_calculated>1173661624.591153</p_calculated>
<m_nbytes>2124599296.000000</m_nbytes>
<m_cache>524288.000000</m_cache>
<m_swap>2146754560.000000</m_swap>
<d_total>210880741376.000000</d_total>
<d_free>142760882176.000000</d_free>
<os_name>Linux</os_name>
<os_version>2.6.19-1.2911.6.5.fc6</os_version>
<accelerators></accelerators>
</host_info>

The entire flags string from /proc/cpuinfo is below..

fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8legacy ts fid vid ttp tm stc

I'm pretty sure David or Rom checked in a fix for this. I'll have to go dig through CVS check-in notes to be certain.

---

Kathryn :o)

--

Thanks for the efforts..

I must have been thinking of this...

+ David 27 Mar 2007
+ - scheduler: fix bug that cause <p_model> parse error
+ for Linux hosts with long strings
+ (from Andre Kerstens)
+
+ sched/
+ server_types.C

---

Kathryn :o)

I must have been thinking of this...

+ David 27 Mar 2007
+ - scheduler: fix bug that cause <p_model> parse error
+ for Linux hosts with long strings
+ (from Andre Kerstens)
+
+ sched/
+ server_types.C

That just might address the issue... Looks like it was just addressed... Thanks for the help... hopefully the repaired code will be available soon... (I have been posting all over the place trying to get someone's attention for a couple of weeks or so) Thanks again...

It's a server side fix as far as I can tell.

I think I just read something about it over at Docking@Home. Actually the person who submitted that patch to David was one of the admins there.


So, it's quite possible that it isn't the fix I was thinking about. But a search of the CVS check-in notes didn't give anything else applicable.

---

Kathryn :o)

So, it's quite possible that it isn't the fix I was thinking about. But a search of the CVS check-in notes didn't give anything else applicable.

You're probably either thinking of this one:

David 19 Mar 2007
- removed [features] from p_model;
move it to a separate field, p_features,
which is stored on the client and sent to server
but not stored in server DB.
- fix gcc 4.x warnings
- user web: change of app version list XML

client/
hostinfo_unix.C
hostinfo_win.C (new)
main.C
win/
hostinfo_win.cpp (removed)
html/user/
apps.php
lib/
filesys.C
hostinfo.C,h

Or this one:

David 10 Mar 2007
- scheduler: use 1024-char buffer for parsing
(handle large CPU model strings)

sched/
server_types.C

The 2nd one is also server side, although the 1st one is the most likely one to solve the reported problem in the long run.

Been tracking this over at MCDN...

--

OK... this begs a second question... What revision should contain the fix ???

Don't get me wrong, I am willing to wait... I just want to know when I should check back for the new client... (So I can install it and see if it cures my ills...).... ;)

Anything client side should be in the next recommended release. Now whether that will be a 5.8.x release, I don't know. I'm not sure if they're going to freeze the recommended versions as they are now or bump them up and make the latest alpha versions for each platform (currently Win/Mac are on 5.8.16 and Linux is on 5.8.17) the recommended version. I haven't heard anything about when testing for the 5.10.x line will start.


Anything server side is up to the individual projects as to when they want to update server code.

---

Kathryn :o)