climateprediction.net home page
Posts by ML1

Posts by ML1

1) Questions and Answers : Unix/Linux : No work for Linux either now. (Message 58981)
Posted 9 Nov 2018 by ML1
Post:
Really? Are there too few 64-bit Linux systems to be worth a recompile?...

;-)


The greater reliability must be worth something?...

;-) ;-)


Keep searchin',
Martin
2) Message boards : Number crunching : Benefits of hyper-threading on an i7-4790 (Message 58980)
Posted 9 Nov 2018 by ML1
Post:
Thanks for that, interesting.

There's various sources claiming that (Intel) hyperthreading can give at most a 30% increase on total throughput at the expense of greater latency (you don't get anything like x2 CPU cores!). Your +28% seems to bear that out.

The "Joker" for the performance is for what memory bandwidth is required to keep the CPU threads busy and what the CPU cache can support before getting 'poisoned' by too many cache misses...

So... All very task specific!

But in short:

For well behaved sympathetic tasks with sympathetic cache and memory bandwidth needs, then hyperthreading can give you a gained +30% overall throughput. However, that can go negative if the hyperthreading allows (or tries to push) too much activity and the cache is overwhelmed...

Always good to measure what really happens and the 'writes per day' is a good yardstick of progress. Good stuff and interesting for the max number of simultaneous tasks... (IE: Don't max out!)


Thanks,

Happy fast crunchin'!
Martin
3) Questions and Answers : Unix/Linux : Unix ctrl^Z not working (Message 5570)
Posted 22 Oct 2004 by ML1
Post:
> I am able to start up a project fine but when I want to put
> it into the backgroung I press ctrl^z and it keeps running.

Yes, ofcourse it would... (;-))

If by "background running" you really mean 'low priority', then you have no worries. It already runs at minimum priority. See for yourself using the command "top".

There are many ways to run "boinc":

Interactively on the command line as you seem to be doing already. (Just leave it and open another terminal session.)

As a background task "boinc >logfile 2>errorsfile &"

And there are various scripts on the boinc forums for running boinc as a system service. (Try a search or google.)

I just use my own background user script...

Good luck,
Martin

Mandrake 10.0.1

4) Questions and Answers : Unix/Linux : still errors (Message 2602)
Posted 1 Sep 2004 by ML1
Post:
> [ Download problems? ]
> wanting to get it running ..

Pete,

You don't give many clues as to what you are doing or what is happening!

Also, please keep to one thread for your problem.

Where are you trying to download from?
What is your system?
Are you on dialup?

Perhaps try downloading from one of the mirrors? Or are you only having problems for the client models?

Are you running from a shell that is then timing out? Are you logging off and hence killing your process??...


Give us a few clues. Others have this working ok.

Good luck,
Martin

5) Questions and Answers : Unix/Linux : Viz cloud display discontinuities at the meridian (Message 2070)
Posted 29 Aug 2004 by ML1
Post:
> I guess this is hadsm3_4.02, because it's fixed in 4.03, except you can still
> get white cells in very hot areas which will be "not found" in 4.04

Yep. I'm still running the beta stuff:

hadsm3_4.02_i686-pc-linux-gnu
hadsm3se_4.02_i686-pc-linux-gnu
hadsm3um_4.02_i686-pc-linux-gnu
hadsm3viz_4.02_i686-pc-linux-gnu


Aside: Their .zip files are left undeleted after being uncompressed. Is this tidied up in the later versions?

Regards,
Martin
6) Questions and Answers : Unix/Linux : Viz cloud display discontinuities at the meridian (Message 2044)
Posted 29 Aug 2004 by ML1
Post:
Looking at my model with viz and... There is a definite and persistent visual discontinuity in the clouds displayed at 0 deg longitude and 180 deg longitude. Similar effects can be seen less markedly for some of the other longitude segments. There is no similar effect visible for any of the other displays.

Is this a quirk of just my model, viz, or something else?

The temperature view also shows good examples of out of range (white) temperatures (:-))

Here is one example:

Regards,
Martin


The clouds:
<img src="http://homepage.ntlworld.com/pazml1/CPDN/CPDNclud6b.png">

Temperature:
<img src="http://homepage.ntlworld.com/pazml1/CPDN/CPDNtemp3b.png">

Pressure:
<img src="http://homepage.ntlworld.com/pazml1/CPDN/CPDNpres4b.png">




7) Message boards : Number crunching : Succesful completion & upload... :-) (Message 1586)
Posted 24 Aug 2004 by ML1
Post:
[...]
&gt; But seriously, here is what has been fixed off of the top of my head:
[...]

Carl,

Do you want the present Beta runs to continue with their present models, or should we all reload with the updated 'official' release?

(Are the Beta model results useful?)


Aside: Aswel as the global temperature charts, can charts be plotted for global min/max temps, and similarly for global min/ave/max windspeed and pressure?

Regards,
Martin

8) Message boards : Number crunching : Succesful completion & upload... :-) (Message 1570)
Posted 24 Aug 2004 by ML1
Post:
&gt; Uploaded automaticaly without a bobble

Nick,

Good to see your result uploaded ok. I'm still on phase 1, so you've blazed past me for that WU (:-))

Happy crunchin',
Martin

9) Message boards : Number crunching : Crashing CPDN client (Message 1568)
Posted 24 Aug 2004 by ML1
Post:
[...]
&gt; Anyway, my computer locked up again last night. When I first upgraded my CPU
&gt; I had some stability issues like this but I tweaked some settings and got it
&gt; stable running seti@home. I'm wondering if CPDN uses some secret circuit in
&gt; my CPU that seti@home never touched which is still not stable. This time,
[...]

Test your system with:
memtest86
GIMPS mprime 'torture test'
and just to be thorough, check your HDD with the manufacturer's disk check utility.

You get some reassurance that the machine is ok for whatever ambient conditions you did the tests under. If your machine later gets hotter or just more dirty, then it can still fail.

A UPS might be a good idea if your mains is flakey.

Good luck,
Martin

10) Message boards : climateprediction.net Science : Outcome- Client Error (Message 974)
Posted 14 Aug 2004 by ML1
Post:
[...]
&gt; I still got some errors, and found my memory tweaked a little too high, which
&gt; under normal circumstances, work fine, but not w/BOINC apparently.
[...]

Check out your system with Memtest86 and also the GIMPS prime torture test. They will pick up errors that are a problem but that often go unnoticed...

Do the checks at the max operating temperature!

Good luck,
Martin

11) Questions and Answers : Unix/Linux : error 251, now not segfaulting, just keeps downloading more.... (Message 830)
Posted 12 Aug 2004 by ML1
Post:
Try downloading boinc oncemore from cpdn. Install into a new directory and try running again.

Also check your system logs for clues.

If you get repeat errors, then time to check further...


Wild guesses:

File and directory permissions correct?

Are you overclocking or running hot? Does your system check out ok with Memtest86 and the GIMPS mprime 'torture test'? Have you compiled the source with over-agressive or experimental compiler options?


Good luck,
Martin

12) Questions and Answers : Unix/Linux : kernel panic - preempted kernel (Message 710)
Posted 11 Aug 2004 by ML1
Post:
cpdn and boinc are not at fault for your kernel panic other than just happening to be running and loading your system.

They are "user space" programs and as such they are completely at the mercy of the kernel.

Note that the preempt features in 2.6 are still being developed. There's lots about this on linux.kernel.

See if you get better stability with preempt turned off or with a lower compiler optimisation for your system.

Check your system logs for what else was happening!


Good luck,
Martin

Mandrake 10.0.1, kernel 2.6.3-7mdk

13) Questions and Answers : Unix/Linux : SCHEDULER_REPLY::parse(): bad first tag (Message 415)
Posted 8 Aug 2004 by ML1
Post:
&gt; this sound like a problem created by the server or the line, not by the
&gt; client. So nothing you can do with the client can help there.

Looks like it was server or comms problems. A few attempts later and it got through ok.

The simulation continues.

Regards,
Martin

(Yes, most linux people can use a keyboard (;-))


FYI:

2004-08-08 00:32:16 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
2004-08-08 00:32:16 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded

14) Questions and Answers : Unix/Linux : SCHEDULER_REPLY::parse(): bad first tag (Message 368)
Posted 8 Aug 2004 by ML1
Post:
This error has appeared in the log:

0008_000024993 - PH 1 TS 032503 - 18/10/1812 03:30 - H:M:S=0042:09:09 AVG= 4.67 DLT= 2.40
2004-08-08 00:19:10 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
2004-08-08 00:19:10 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded
2004-08-08 00:19:10 [climateprediction.net] SCHEDULER_REPLY::parse(): bad first tag Content-type: text/plain
2004-08-08 00:19:10 [climateprediction.net] SCHEDULER_REPLY::parse(): bad first tag Content-type: text/plain
2004-08-08 00:19:10 [climateprediction.net] Can\'t parse scheduler reply
2004-08-08 00:19:10 [climateprediction.net] Can\'t parse scheduler reply
2004-08-08 00:19:10 [climateprediction.net] Deferring communication with project for 13 minutes and 6 seconds
2004-08-08 00:19:10 [climateprediction.net] Deferring communication with project for 13 minutes and 6 seconds
0008_000024993 - PH 1 TS 032504 - 18/10/1812 04:00 - H:M:S=0042:09:26 AVG= 4.67 DLT=16.77


Kill and restart?
New boinc client required??
Related to the boincstat note for bad xml tags?
Connected with the \'failing to trickle\' thread?

Regards,
Martin

15) Questions and Answers : Unix/Linux : Client installation instructions (Message 309)
Posted 7 Aug 2004 by ML1
Post:
Here are few brief notes to get you going. Sorry, you're going to have to do a little man page reading...

Don't run as root. There is no need to. If you are running as root then you don't know what you're doing and you're bound to make extra work for yourself (:-))

Root is there to let you play god for 'system' tweaks only.


To have boinc as an always running background process, you're going to have to discover the command line (terminal) interface...

When you have set up boinc as you want and attached, stop it. Then one way (of many) is to run it redirecting all output to a log file:

(cd into the directory you have for boinc)
$ ./boinc_4.02_i686-pc-linux-gnu &gt;boinc.log 2&gt;&amp;1 &amp;

This runs as a 'background job'. You can see it running with "top" or "ps".

To keep it running even when you log out, then you need to 'disown' it:

Type:
$ jobs
to get its job number. Then:
$ disown job_number

To continuously see the log output, use:

$ tail -f boinc.log
Use ctl-c to kill tail and get your command prompt back.

To see the man pages for any command, use:
$ man tail
for example, or in KDE's konqueror on the address line:
#tail


I've got something similar to that lot scripted to make it automatic. If you're really interested, you can make it into a 'service'.

Good luck,
Martin

16) Questions and Answers : Unix/Linux : boinc client zip files not deleted (Message 307)
Posted 7 Aug 2004 by ML1
Post:
Looking alongside viz and in that directory there\'s various \'zip\' files that have been uncompressed.

The original zip files are left undeleted. Deleting these will recover a good few MBytes and keep things more tidy.

Regards,
Martin

17) Questions and Answers : Unix/Linux : viz useage/features on linux (Message 244)
Posted 6 Aug 2004 by ML1
Post:
Tried ./viz and... very nice if only a little slow.

1: Can you zoom in to get a closer view of the planet or just one continent even?

2: Can you set an update time step rather than have it continuously (CPU hogging) update itself?

(Or where have I missed the instructions for all this? (:-))


Regards,
Martin

Mandrake 10.0.1 GNU/Linux

18) Questions and Answers : Unix/Linux : Bad XML (Message 237)
Posted 6 Aug 2004 by ML1
Post:
I have a similar error here. boincstat worked fine for when I was running s@h under boinc 3.20.


More useful for debug is (boinc 4.02):

$ ./boincstat
/home/bgusr/.boinc/client_state.xml:130: parser error : Opening and ending tag mismatch: file_signature line 128 and signed_xml

^
/home/bgusr/.boinc/client_state.xml:131: parser error : Opening and ending tag mismatch: signed_xml line 124 and file_info

^
/home/bgusr/.boinc/client_state.xml:412: parser error : Opening and ending tag mismatch: file_info line 109 and client_state

^
/home/bgusr/.boinc/client_state.xml:413: parser error : Premature end of data in tag client_state line 1

^
Document not parsed successfully.


Looking up to line 131:
{app}
{name}hadsm3{/name}
{/app}
{file_info}
{name}hadsm3_4.02_i686-pc-linux-gnu{/name}
{nbytes}1202259.000000{/nbytes}
{max_nbytes}0.000000{/max_nbytes}
{status}1{/status}
{executable/}
{signature_required/}
{file_signature}
5711334c4eb0af326bc28cad76cdfa5c15480545d68c2311b4ff50c3efd0197b
e23f230212b7035533a7c0a7eebf373943564865a91193e8ac4490c2562b469d
b706ff027eca98016f6eaba45b1a1ea5528ea9847809528fda0fc0622ba42f54
a82209bc756caa8a19cf26a4fb771747b3c2ad1f2962eecd67cef49a85087a2b
.
{/file_signature}
{url}http://climateapps2.oucs.ox.ac.uk/cpdnboinc/download/hadsm3_4.02_i686-pc-linux-gnu{/url}
{signed_xml}
{name}hadsm3_4.02_i686-pc-linux-gnu{/name}
{url}http://climateapps2.oucs.ox.ac.uk/cpdnboinc/download/hadsm3_4.02_i686-pc-linux-gnu{/url}
{executable/}
{file_signature}
{nbytes}1202259.000000{/nbytes}
{/signed_xml}
{/file_info}

(OK, so how do you get tags _not_ to be interpreted by the message board?!)

And yep... The xml tags are not bracketing properly...

Is this a boinc 4.02 bug?
Also, why is 4.02 so much smaller than the previous 3.20 version?


(Meanwhile, the simulation is burning through phase 1)

Regards,
Martin

19) Message boards : Number crunching : Server time? (Message 95)
Posted 5 Aug 2004 by ML1
Post:
&gt; I'm going to look into just "ntp'ing" all CPDN servers tomorrow to make sure
&gt; we're "in sync":
&gt;
&gt; http://www.ntp.org/

If you're on linux, if you're interested and to save you some time, I can email a configured ntpd.conf file to get you going quickly.

I'm sure you can find my email (;-))

Regards,
Martin

20) Message boards : Number crunching : BOINC user preferences (Message 64)
Posted 5 Aug 2004 by ML1
Post:
\" These apply to all BOINC projects in which you participate.
On computers attached to multiple projects, the most recently modified preferences will be used. \"


I\'m running both s@h and CP beta with a 100:950 split... (:-))

CP BOINC v4.02 shows slightly different default prefs to those for s@h. Also, in those defaults are the cache limits which will apply to \'all projects\'.

CPDN &amp; s@h have very different WU timescales. I certainly don\'t want to hog 60 day\'s worth of s@h WUs!


Can BOINC be updated so that the cache defaults are overridden in the project specific prefs?

Or how will CP react if I knock the cache back down to the s@h defaults of 0.1 to 1 day??

Regards,
Martin



Next 20

©2024 climateprediction.net