When crunching Classic CPDN THC-runs, my finished experiments were neatly packed up. Then I should be able to look at the experiments later using Martin\'s CPWiew or the Advanced Visualisation packet.

I am now running BOINC, crunshing two Sulphur-runs.
Reading the posts I get to understand that BOINC erases the data, once the completed run has been uploaded.
Has anybody found a way to \'snip\' a copy, for the records (without trashing the results, that is :)).
I have an idea of suspending network activity, before the Run is complete. When the run reports ready for upload, I will take a backup, and then alow the result to be uploaded.
Is this a way that will work, or is there a better way?

Another silly q.
In order to look at one of my previous THC runs, I need to extract the whole result-dir.
Doing this one file at a time takes forever.
I know someone has made a Batch-file for this task.
Could I be directed to it?

Thanks Folks, sorry if these q. has been asked before, but I just ran out of luck searching through years of posts.

One last Q. My Athlon XP3000+ takes 4,03 S/TS and my Athlon64 3500+ takes 2.67 S/TS.
How does this compare?

Happy Crunching.

Chris D

---

Not all of the data from BOINC\'s slab and sulphur runs are uploaded.
Some is left on your computer, and you can use this to look at what you have created.
About 660Megs for slab, and 800-900Megs for sulphur.
This is about to change for the next lot, the coupled ocean, where more data is uploaded, and more frequently, and very little is left on your computer.
There\'s no info yet on this model type.

In order to look at one of my previous THC runs, I need to extract the whole result-dir.
Doing this one file at a time takes forever.
I know someone has made a Batch-file for this task.
Could I be directed to it?

Do a search for \"FilZip\" on google. That application will allow you to extract all files.

One last Q. My Athlon XP3000+ takes 4,03 S/TS and my Athlon64 3500+ takes 2.67 S/TS.
How does this compare?

My XP3200+ does about 3.2 s/TS. But how fast the 3000+ goes depends on whether its front side bus is 166 or 200 MHz, what type of motherboard you have, and what timings your memory is running at. If it\'s a 200 MHz front side bus 3000+ running on an NForce2 motherboard with dual channel memory, it\'s probably a little slow. If it\'s otherwise, it might be about right.

Author said:

My XP3200+ does about 3.2 s/TS. But how fast the 3000+ goes depends on whether its front side bus is 166 or 200 MHz, what type of motherboard you have, and what timings your memory is running at. If it\'s a 200 MHz front side bus 3000+ running on an NForce2 motherboard with dual channel memory, it\'s probably a little slow. If it\'s otherwise, it might be about right.

My XP3000+ is a BARTON Core running 166 MHz FSB with one 256 Mb DDR Module.
MB is a GigaByte GA7-VT600. Memory runs FSB+33MHz at 2.5-3-3-7.
Memory is Corsair XMS3200LL and should run 2-2-2-5 at 200 MHz, but there is no BIOS options for memory timings other than I can make it run at 200 MHz. Setting memory speed to 166 MHz gives me the same timings. Seems that this MainBoard is no good when using LL RAM :(

Thank for the FilZip hint.

ChrisD

---

My XP3000+ is a BARTON Core running 166 MHz FSB with one 256 Mb DDR Module.
MB is a GigaByte GA7-VT600. Memory runs FSB+33MHz at 2.5-3-3-7.
Memory is Corsair XMS3200LL and should run 2-2-2-5 at 200 MHz, but there is no BIOS options for memory timings other than I can make it run at 200 MHz. Setting memory speed to 166 MHz gives me the same timings. Seems that this MainBoard is no good when using LL RAM :(

I used to have that motherboard, but it was quite awhile ago. When in the BIOS main screen, hitting Ctrl and F1 at the same time should give you advanced options in the submenus. Does this give you some option for setting the memory timings?

I thought the XMS 3200LL was 2-3-2-6 and that the 3200XL was 2-2-2-5.

I used to have that motherboard, but it was quite awhile ago. When in the BIOS main screen, hitting Ctrl and F1 at the same time should give you advanced options in the submenus. Does this give you some option for setting the memory timings?

Ctrl + F1 option was unknown to me. I better study the manual one more time :)

And: Sure, I am going to check that one out. :)

I thought the XMS 3200LL was 2-3-2-6 and that the 3200XL was 2-2-2-5.

Sorry, just checked. You are right and my menory is the 3200XL. Seems that my memory failed me.

ChrisD

---

Ctrl + F1 option was unknown to me. I better study the manual one more time :)

The manual doesn\'t say much about it, other than it exists, if I recall correctly. Gigabyte seems to like to hide performance features, in the BIOS, and in the manual. Anandtech talks about this in the BIOS section of their review of the slightly different 7VT600 1394.

Sorry, just checked. You are right and my menory is the 3200XL. Seems that my memory failed me.

Yes, very good memory. I have some of that myself. I would think with the timings at 2-2-2-5 and 1T command rate, you could get about 0.2 s/TS better than it is currently doing. It would also be interesting to see if you get better performance at with the memory at 166 (333) or 200 (400). I know with the NForce2, the memory does better at the same frequency as the front side bus but I can\'t recall if that is the same on the KT600 chipset motherboards.

ChrisD,

Looks like your PC has speeded up (averaging about 3.88 s/TS for the last few trickles). Not quite the 0.2 s/TS improvement I predicted, but close.

ChrisD,

Looks like your PC has speeded up (averaging about 3.88 s/TS for the last few trickles). Not quite the 0.2 s/TS improvement I predicted, but close.

I run the FSB at 166 MHz only, and so does the memory. I am not sure the processor is stable at 200 MHz FSB. If this is the case, then I have to choose a x 10.5 multiplier which is slower than the normal 2167 MHz or OC at x 11.
As CPDN is a long experiment I am not keen of risking OC to gain the last ounce. I might loose the whole basket..

Happy crunching

ChrisD

---

I run the FSB at 166 MHz only, and so does the memory. I am not sure the processor is stable at 200 MHz FSB. If this is the case, then I have to choose a x 10.5 multiplier which is slower than the normal 2167 MHz or OC at x 11.
As CPDN is a long experiment I am not keen of risking OC to gain the last ounce. I might loose the whole basket..

Right. I wasn\'t suggesting overclocking, but you must have changed the memory timings to have speeded up that much.

Right. I wasn\'t suggesting overclocking, but you must have changed the memory timings to have speeded up that much.

Sure :) Thanks to Your info about this hidden ctrl + F1, I was able to change the timing for the memory to the 2-2-2-5, instead of the automatically selected 2.5-3-3-6.

ChrisD

---