Just downloaded and started running the new “FAMOUS” models. All I can say is WOW!!! The graphics are really something. I just did something that I didn’t think that I would ever do, I sat there and watched it crunch an entire month. Days take only 24 seconds on my 2.2 GHz (4 GB’s of RAM, Windows7 64 bit) laptop. They must really fly for you guys with the 3.5 GHz desktop machines. These are going to be fun to run. :)

P.S. I just noticed that I am getting 0.27s/TS!!!!! Is that great or what.

---

That is pretty fast. I\'m getting 0.2248 sec/TS with a FAMOUS model. Let\'s hope they become famous for something good, instead of ice worlds like some other types are. :)

I wonder if there are any problems with memory contention running 4 of these models at once (like the am3p models)?

EDIT: Would also like to note, this model has been running for 6 hours and is about 5% complete so far. That\'s definitely a shorter model than what we\'re used to. I really hope that translates into a better percentage of successful models for the whole project.

When our model pages also show the graphs it will be better still because you see the historical level of known volcanic emissions and how they affected the temperatures. I think Tolu may be modifying the graphs before he adds them to our model pages on CPDN.

---

Cpdn news

DJStarfox asked:

I wonder if there are any problems with memory contention running 4 of these models at once (like the am3p models)?

This is something that should ideally have been tested in Beta but it\'s taken so long with so many beta versions to produce a working model that I\'m afraid nobody tested this.

Is there anyone with a quad who\'s been able to get three or better still four models? It would be useful to run just one on its own from trickle to trickle (2 Dec to 2 Dec), then two FAMOUS from trickle to trickle, then three together etc, keeping a record of which models ran when. Then do some arithmetic subtracting CPU times on the models\' web pages to work out whether the CPU times per trickle are adversely affected.

I only have a C2D, not a quad, so I couldn\'t test this properly myself.

I suspect we\'d find that multiple FAMOUS affect each other very little, though when a computer uses 4 cores at 100% I expect there\'ll always be some slowdown whatever is running.

---

Cpdn news

I wonder if there are any problems with memory contention running 4 of these models at once (like the am3p models)?

There is very little memory contention running multiple FAMOUS models. Far less than other models, especially far less than hadam3 or hadam3p. My Q9550 in Windows might run one model at .17 s/TS and 4 at .19 s/TS.

My experience is/was the same as Geophi\'s.
I thought that I\'d mentioned this early in the testing, but I may only have planned to say it.

By all means, load up quads with 4 FAMOUS models. They\'re so fast that a few seconds don\'t matter. Perhaps when/if the 1,000 year models show up.

---

Backups: Here

I\'ve found where I posted about 4 FAMOUS together: it was here, in the Return of the AM3Ps thread.

---

Backups: Here

I\'ve found where I posted about 4 FAMOUS together: it was here, in the Return of the AM3Ps thread.

If you and perhaps many other volunteers are generating zip files every 7 hours, I hope there is a lot of free space on the upload servers. This might be a train wreck otherwise.

Running 4 FAMOUS models too; getting a slight slowdown of +0.02 sec/TS each. I wonder what 8 models would do?

hello,

the answer to that is 0.18 s/ts on a total of 8 models using an intel i7 920

of course faster quad cores will no doubt get this number even lower running 8 models

Ian