Unless I am mistaken, there is an error in the display at:

YourAccount/(computer)View/Tasks/WorkUnitID

All tasks seem to have a display shown under \"Errors\"

as \"Too many Total Results\"

Either I\'m misunderstanding what this means, or it\'s incorrectly displayed. This seems to be shown for work units of all tasks for myself and as far as I have seen, for all other crunchers.

I\'m running Climateprediction.net on MacBookPro using BOINC 6.2.4 and OSX 10.5.2.
(Could it be a problem in BOINC 6.2.4, being a test application?)

Keith

\"Too many Total Results\" is a message that only applies to projects that use a quorum of WUs for validation purposes.
In this project, it seems that the large number of trickles, (which I don\'t think any other project uses yet), gets mistaken for the final results of many models in a quorum, hence the message.
Ignore it, and most of the other messages on that page.

The only thing of interest there is the list of work units from other computers running the same data set. And that\'s only of interest for comparing the same \"model\" on different computer setups.

\"Too many Total Results\" is a message that only applies to projects that use a quorum of WUs for validation purposes.
In this project, it seems that the large number of trickles, (which I don\'t think any other project uses yet), gets mistaken for the final results of many models in a quorum, hence the message.
Ignore it, and most of the other messages on that page.

The only thing of interest there is the list of work units from other computers running the same data set. And that\'s only of interest for comparing the same \"model\" on different computer setups.

Les

Your reply is certainly plausible, but wouldn\'t the programmers appreciate that this incorrect display be notified to them? I know that I would appreciate feed back if I were in their position. Would you be able to pass the message on to them?

As you say, ClimatePrediction is in a minority (if not unique) in using trickles, so it is even more relevant that this anomaly be brought to their attention as it will be noticed less frequently.

Thanks for replying to my original post. I do, in fact, like to compare the progress of my computer\'s task with the tasks of others in the same work unit. I

have just revived \"Virtual PC 6\" on my old beige G3 Mac (with G4 upgrade) running OSX 10.2.8, and am astounded at the snail\'s pace it is processing. Strangely the Completion Time is increasing (not decreasing) at almost the same speed as the CPU Time is increasing. So, after 4 days, we have only just touched on 0.001% Progress. No wonder it\'s running at \"High Priority\" to meet a deadline of 17/11/2010!! My calculation on projected CPU time is 15 months, which, allowing 50% of CPU, will only just finish in time (also assuming the Completion Time will start decreasing soon)!!

Keith

Hi

Wow, you\'re running Intel BOINC under Virtual PC. As the G4 is trying to pretend to be an Intel chip using emulation (which is very, very inefficient), this is going to take an age. So slow, I\'m not sure it\'s worth the electricity.

I have a G4 too, but I\'m not running BOINC on it. It\'s not that the computer is slow, it\'s just emulation is way too heavy for number crunching.

Cheers.

e

P.S. I read this in a forum somewhere, they said they were quoting Microsoft but I can\'t be sure:

Virtual PC uses the PowerPC\'s FPU to emulate the Pentium\'s FPU. The PowerPC uses 64-bit numbers, while the Pentium uses 80-bit numbers, so Virtual PC\'s calculations are not quite as accurate. To give you an idea of the accuracy: if you were using Virtual PC to calculate the national debt, AND for some reason you needed exponentiation in your calculations, AND instead of Excel or MS-BASIC, you used a third party accounting software the result will be well off (as yoiu suggest, on doing any kind of rounding at all, you would get the correct answer).

P.P.S. Just read, the Altivec isn\'t as accurate as MMX, but standard G4 FPU work is fully compliant and won\'t cause accuracy errors.

... but wouldn\'t the programmers appreciate that this incorrect display be notified to them? I know that I would appreciate feed back if I were in their position. Would you be able to pass the message on to them?

As you say, ClimatePrediction is in a minority (if not unique) in using trickles, so it is even more relevant that this anomaly be brought to their attention as it will be noticed less frequently. ...

Keith,

The programmers have been made aware of this problem (and other problems) with the work unit page caused by differences between CPDN and other projects. I hope it\'s made its way onto their rather long to-do list ...

Iain