> The scheduler RPC mechanism should now be working as it should.
> You can force an update to update stats amongst other things
> and remove those annoying warning messages.
>

Just tried it and everything is fine here now.

---

_________________________________

> The scheduler RPC mechanism should now be working as it should.
> You can force an update to update stats amongst other things
> and remove those annoying warning messages.

Great, thanks Tolu - all mine have now updated without any error messages. :-)

---

<a href="http://www.nmvs.dsl.pipex.com/"><img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=6&amp;team=off&amp;trans=off"></a>

<a href="http://www.nmvs.dsl.pipex.com/">Distributed Mania</a>

Just thought I'd add:

I had this problem on 1 of 5 machines. It still would not contact a scheduler, and gave the 'master file' error. This happened after several forced updates, only minutes ago. I then suspended, exited, and re-launched the GUI. It successfully made contact &amp; updated upon its re-launch.

Strat

---

&gt; only minutes ago. I then suspended, exited, and re-launched the GUI. It
&gt; successfully made contact &amp; updated upon its re-launch.
I restarted GUI on both computers and I still getting the same error ...

---

ALL GLORY TO THE HYPNOTOAD!
Potrebujete pomoc?
My Stats

&gt; The schedulers seem to be up now, new users will want to attach directly to:
&gt;
&gt; http://climateapps2.oucs.ox.ac.uk/cpdnboinc
&gt;
&gt; (instead of the usual climateprediction.net of course)
&gt;

Thanks, Carl. That allowed my new machine (P4 3.4 GHz on 775-pin array) to connect and start CPDNboinc and end hours of frustration. (To make installation of the new Ebox more difficult, my ISP tanked for a few hours. However, at first Trickles, average throughput is 1.25 s/TS -- should be useful for E@H, when called, and the coming CPDN Sulfur Cycle/full ocean Models.)

Thanks, again, and I hope the new job goes well.

Jim

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

&gt; The schedulers seem to be up now, new users will want to attach directly to:
&gt;
&gt; http://climateapps2.oucs.ox.ac.uk/cpdnboinc
&gt;
&gt; (instead of the usual climateprediction.net of course)
&gt;
&gt;
&gt;

how do i go about doing this?

In your boinc client click Settings/Attach to Project.
In the Project URL: field type (or copy-and-paste) http://climateapps2.oucs.ox.ac.uk/cpdnboinc
In the Account ID: field copy-and-paste the account ID from the email that you received when you created an account with CPDN. Normal Windoze hilighting copies a space at the end of the selected string, ensure that there is no space there before clicking OK.
Click OK.
Assuming the URL is working and you supplied a valid Account ID and everything's working at Oxford, your client will then download a client and a WU and start processing it.

---

&gt; In your boinc client click Settings/Attach to Project.
&gt; In the Project URL: field type (or copy-and-paste)
&gt; http://climateapps2.oucs.ox.ac.uk/cpdnboinc
&gt; In the Account ID: field copy-and-paste the account ID from the email that you
&gt; received when you created an account with CPDN. Normal Windoze hilighting
&gt; copies a space at the end of the selected string, ensure that there is no
&gt; space there before clicking OK.
&gt; Click OK.
&gt; Assuming the URL is working and you supplied a valid Account ID and
&gt; everything's working at Oxford, your client will then download a client and a
&gt; WU and start processing it.
&gt;

this just causes it to download, and start on, another project!
and it deletes the previous project from its memory!
how do i get it to continue working on the original data?
or does it do this automatically?

---

So I tried something. And now I know.
I can not connect to original project URL http://climateprediction.net even when I want to attach to project from here. It works at home, but not from this place.
We are using proxy and firewall. At home I'm using only personal firewall.
And yes, when I change URL then I lost my work already done here.

---

ALL GLORY TO THE HYPNOTOAD!
Potrebujete pomoc?
My Stats

&gt; So I tried something. And now I know.
&gt; I can not connect to original project url http://climateprediction.net even
&gt; when I want to attach to project from here. It works at home, but not from
&gt; this place.
&gt; We are using proxy and firewall. At home I'm using only personal firewall.
&gt;

First off, check your proxy/firewall settings. Otherwise if this is a recent problem as a result of the server downtime, try deleting the existing master.html file in the BOINC root directory ( make a back up first ) then do an update on the project.

---

Jim, congrats on arriving your new Ebox.
I only wonder if there is different benchmark measurements among Win and Linux.
My 3.2GHz Northwood shows about 1400/2070 in a BOINC benchmark, yours 3.4 Prescott show only 700/1450. Perhaps just a cosmetic glitch of synthetic measurements - CPDN will prove itself...I just wonder - it's it the opposite to well know M$ penatly (BOINC being slower under Windows)...
Any measurement on CPU temperature?

....my new machine (P4 3.4 GHz on 775-pin array) to
&gt; connect and start CPDNboinc and end hours of frustration. (To make
&gt; installation of the new Ebox more difficult, my ISP tanked for a few hours.
&gt; However, at first Trickles, average throughput is 1.25 s/TS -- should be
&gt; useful for E@H, when called, and the coming CPDN Sulfur Cycle/full ocean Models.)

my original project still refuses to upload its data, even though i've changed the url!

Anyone know what is wrong?
What can i do to solve the problem?

---

&gt; Jim, congrats on arriving your new Ebox.
&gt; I only wonder if there is different benchmark measurements among Win and
&gt; Linux.
&gt; My 3.2GHz Northwood shows about 1400/2070 in a BOINC benchmark, yours 3.4
&gt; Prescott show only 700/1450. Perhaps just a cosmetic glitch of synthetic
&gt; measurements - CPDN will prove itself...I just wonder - it's it the opposite
&gt; to well know M$ penatly (BOINC being slower under Windows)...
&gt; Any measurement on CPU temperature?
&gt;
&gt; ....my new machine (P4 3.4 GHz on 775-pin array) to
&gt; &gt; connect and start CPDNboinc and end hours of frustration. (To make
&gt; &gt; installation of the new Ebox more difficult, my ISP tanked for a few
&gt; hours.
&gt; &gt; However, at first Trickles, average throughput is 1.25 s/TS -- should be
&gt; &gt; useful for E@H, when called, and the coming CPDN Sulfur Cycle/full ocean
&gt; Models.)
&gt;
&gt;


i'm assuming that Jim has hyperthreading turned on, and you don't.....that would explain the benchmarks

---

<img src="http://tinyurl.com/56gex"> <img src="http://tinyurl.com/5lcmd"> <img src="http://tinyurl.com/48rdp">

Hi virex,

Well, good idea but doesn't apply - we both are running all P4s with HT on and two models simultaneously (which does improove final performance).
Let's date put another assumptions:
- BOINC benchmark tests are CPU-numbers blind. e.g. having HT, dual of even multi-cpu machine doesn't run more benchamrks threads hence donesn't affect final score. Take a look at faster quad-Xeons with HT on.
- Prescotts with larger but slower cache/memory access then Northwoods tends to do poor in such tests; AMD64 or FX can easily outperform all Intel CPUs because of their build-in memory controller.

My final assumption is that Prescotts will do a bit slower than Northwoods in BOINC/CPND. They might benefit from larger cache in distributed.net effort or other apps where apps itself fits in cache.


&gt; i'm assuming that Jim has hyperthreading turned on, and you don't.....that would explain the benchmarks

Prescotts are slower then Northwoods and heating more. That's clear.
(In any ocasions)
&gt; My final assumption is that Prescotts will do a bit slower than Northwoods in
&gt; BOINC/CPND. They might benefit from larger cache in distributed.net effort or
&gt; other apps where apps itself fits in cache.

---

ALL GLORY TO THE HYPNOTOAD!
Potrebujete pomoc?
My Stats

Tolu:
Other projects on one of those boxes is running (SETI@home), so should not be FW/Proxy problem (nothing changed in this area anyway on my side). But I will try to remove the file, seems like something went wrong during the last outage in my files. Thanks for now, I will try it in 8 hours on morning at work. ;o)
I have 1 WU at 35% and 2 at more then 40%, to bad to start new one.

&gt; First off, check your proxy/firewall settings. Otherwise if this is a recent
&gt; problem as a result of the server downtime, try deleting the existing
&gt; master.html file in the BOINC root directory ( make a back up first ) then do
&gt; an update on the project.
&gt;

---

ALL GLORY TO THE HYPNOTOAD!
Potrebujete pomoc?
My Stats

&gt; Jim, congrats on arriving your new Ebox.

Hi, Honza; thanks.


&gt; I only wonder if there is different benchmark measurements among Win and
&gt; Linux.
&gt; My 3.2GHz Northwood shows about 1400/2070 in a BOINC benchmark, yours 3.4
&gt; Prescott show only 700/1450. Perhaps just a cosmetic glitch of synthetic
&gt; measurements - CPDN will prove itself...I just wonder - it's it the opposite
&gt; to well know M$ penatly (BOINC being slower under Windows)...
&gt; Any measurement on CPU temperature?
&gt;

Since loading boinc 4,13, the Linux machines' numbers dropped. Ebox is Linux-only, so I can't make a M$-Linux comparison. Cbox, P4 3,0 Northwood, XP-SP2, shows 1263 Whet... per CPU, and 1969 Dhry..., much higher than Bbox or Dbox (similarly configured machines, but running Linux).

After five Trickles, the s/TS numbers on Ebox are 2,4855 &amp; 2,4899.

I haven't looked into the hardware-monitor program, supposedly fixed in SuSE 9.1, so don't know current temps. When playing in the BIOS prior to OS load, temps were low -- but that isn't a useful reading, as there was no load on the CPU. I'll look into it when I get caught-up. (Still no screaming CPU fan noise; I'm pleased about that.)


virex: Yes, HT is turned on, as Honza noted. I consider the Benchmarks meaningless because there is no interference with actual computation and, as I run only CPDN, there is no consequence for Cobbles.

Jim

---

"We have met the enemy and he is us." -- Pogo
Greetings from coastal Washington state, the scenic US Pacific Northwest.

Tolu: I removed master.html file. It works! Great! Thank you very much!

---

ALL GLORY TO THE HYPNOTOAD!
Potrebujete pomoc?
My Stats