21)
Message boards :
Number crunching :
Best Computer Features for Crunching
(Message 47123)
Posted 28 Sep 2007 by zombie67 [MM] Post: I am getting over 850 RAC on this system and it is only running 75% on R@H. It is also running folding at home and World Computing Grid. Which Folding client? Standard, SMP, or GPU? |
22)
Message boards :
Number crunching :
CPU Comparison question
(Message 47094)
Posted 27 Sep 2007 by zombie67 [MM] Post: Mac os x support 2 quad cores. Yes, that is their DC/cluster implementation. But none of Apple's machines are currently available with more than 8 cores (on two chips). |
23)
Message boards :
Number crunching :
CPU Comparison question
(Message 47064)
Posted 27 Sep 2007 by zombie67 [MM] Post: Mac os x support 2 quad cores. Really? Which servers? I am aware of only the Xserve, which currently comes with only 2 dual core woodcrest Xeons. I assume you could manually replace them with quad core clovertown Xeons (like you can with the Mac Pro). |
24)
Message boards :
Number crunching :
Forum Code
(Message 47060)
Posted 27 Sep 2007 by zombie67 [MM] Post: We'll do an update in the near future. Yes! Thanks! Don't forget to also update the "merge computer" function too. |
25)
Message boards :
Number crunching :
CPU Comparison question
(Message 47059)
Posted 27 Sep 2007 by zombie67 [MM] Post: Mac os x support 2 quad cores. But we don't know that it does not support more. There aren't any chips out yet to test that. I assume that OSX will support any number, as it is UNIX. |
26)
Message boards :
Number crunching :
Ready Boost
(Message 47035)
Posted 26 Sep 2007 by zombie67 [MM] Post: The october 2007 issue of CPU magazine ran some tests with many different USB drives. Ready Boost actually slows down everything. It's worth reading. |
27)
Message boards :
Number crunching :
Please Add Opt-Out Option for Beta Workunits
(Message 46833)
Posted 23 Sep 2007 by zombie67 [MM] Post: The only problem I've had with 5.80 is with the CAPRI WU's, and not with every CAPRI WU at that. For me this means there's hardly a 'butt-load of problems'. The 5.80 are (still) throwing out -161 all over the place on RALPH. |
28)
Message boards :
Number crunching :
Please Add Opt-Out Option for Beta Workunits
(Message 46828)
Posted 23 Sep 2007 by zombie67 [MM] Post: A side note: 5.80 clearly has problems, so why is it already here? ? This makes no sense to me. I understand that testing will never turn up every bug. However, 5.80 turned up a butt-load of problems. Yet it was released here knowing the bugs. What then, is the point of RALPH? |
29)
Message boards :
Number crunching :
Please Add Opt-Out Option for Beta Workunits
(Message 46824)
Posted 23 Sep 2007 by zombie67 [MM] Post: If beta work is done on RALPH, then why do we see beta work here? With RALPH, there should never been the need for "problem with..." threads here. A side note: 5.80 clearly has problems, so why is it already here? |
30)
Message boards :
Number crunching :
Pentium D 925 3.0 GHz for Rosetta?
(Message 46684)
Posted 20 Sep 2007 by zombie67 [MM] Post: Sorry, perhaps I misunderstood OP. That's the way I read it too. |
31)
Message boards :
Number crunching :
Pentium D 925 3.0 GHz for Rosetta?
(Message 46652)
Posted 20 Sep 2007 by zombie67 [MM] Post: I have an opportunity to buy one of these pretty cheap. The downside of these is it's only 800 MHz FSB. The upside is the 4 MB L2 cache. Any thoughts on one of these for a dedicated cruncher? The highest RAC I could find for a PD 925 (3.0ghz) is 472. This sounds like a dedicated cruncher, as the PD 3.4ghz versions hover in the high 500s. If it is a Dell, be aware that you will not be able to OC at all. The larger L2 cache is a plus, if you ever decide you want to crunch for folding@home. The smaller L2 of the 8xx series, and all AMDs, doubles the crunching time for the SMP clients. I'm not saying you should go to folding, but it's nice to have the option. On the other hand, you can buy/build a pretty cheap Core 2 machine too, and get 2-3 times the performance of a PD 925. Only you can tell if it's worth it. |
32)
Message boards :
Cafe Rosetta :
Loss of a Cruncher
(Message 46595)
Posted 19 Sep 2007 by zombie67 [MM] Post: Sad. Respect. I know there is nothing I can say to make things better. |
33)
Message boards :
Cafe Rosetta :
Word link 10
(Message 46594)
Posted 19 Sep 2007 by zombie67 [MM] Post: Futurama (Do folks normally read from the bottom-up? Turn this ship around.) Robot Chicken |
34)
Message boards :
Cafe Rosetta :
Last one to post here wins! #2
(Message 46593)
Posted 19 Sep 2007 by zombie67 [MM] Post: When the resolution of enslaving America was formed in Great Britain, the British Parliament was advised by an artful man, who was governor of Pennsylvania, to disarm the people; that it was the best and most effectual way to enslave them; but that they should not do it openly, but weaken them, and let them sink gradually...I ask, who are the militia? They consist of now of the whole people, except a few public officers. But I cannot say who will be the militia of the future day. If that paper on the table gets no alteration, the militia of the future day may not consist of all classes, high and low, and rich and poor... -- George Mason I win! |
35)
Message boards :
Number crunching :
CPU Comparison question
(Message 46526)
Posted 18 Sep 2007 by zombie67 [MM] Post: I have a Q6600 and I have never exceeded 800 RAC on the system. Granted, that machine also runs the Tivo Server and is used by someone about 8 - 10 hours a day. Here is an 8-way (2x Xeon E5345 @ 2.33ghz), with a RAC of 2377: http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=501863 Xeons are not OC-able (at least not via BIOS), so this machine is probably not OC. So a Q6600 should be able to do a bit more than half, or ~1300. And as much as ~1700 OC. |
36)
Message boards :
Number crunching :
Merging computers
(Message 46488)
Posted 18 Sep 2007 by zombie67 [MM] Post: The current (very old) server software will allow computers to be merged if everything in the CPU description and the OS are exactly the same. The current version of the server software is smarter, and will allow merging when things have changed due to things like OS upgrades, or BOINC versions changing (causing the CPU descriptions change). Yes. The developer uses SETI as the test project, so they always have the latest. However, the version needed is not that new. Many other projects already have it implemented. Any chance the Rosetta project would upgrade the server software? Read here: http://boinc.bakerlab.org/rosetta/forum_thread.php?id=3407 http://boinc.bakerlab.org/rosetta/forum_thread.php?id=3488 Bug the Admins. I tried in the past, with no luck. Perhaps if more did.... http://boinc.bakerlab.org/rosetta/forum_thread.php?id=2150 |
37)
Message boards :
Number crunching :
CPU Comparison question
(Message 46487)
Posted 18 Sep 2007 by zombie67 [MM] Post: I have a pc running win xp with a intel core duo e6600 which has 4mb cache per core and 4 gig ram dedicated to RAH and I only get 800 rac's. Sounds a bit low. The highest RAC I could find for a win/E6600 (2.4ghz) machine is 948. There are several in the low 900s, although, it's conceivable that they could all be OC: http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=119208 For comparison against another project, the highest RAC for the same OS/HW box on Einstein is 1327. Again, no way of telling if it is OC: http://einstein.phys.uwm.edu/show_host_detail.php?hostid=967710 There are no optimized apps currently available for Einstein. |
38)
Message boards :
Number crunching :
CPU Comparison question
(Message 46483)
Posted 17 Sep 2007 by zombie67 [MM] Post: The chip used in the iMac is also the laptop version, and I got a RAC of well over 1k with it on SETI. Let me correct myself. I hit a RAC of 2505 with my iMac (Core 2 Duo T7600, 2.33ghz) on SETI. Of course, this was with an optimized science application, which took advantage of SSE/2/3 functionality. At most, the optimized application doubled the performance. So I would expect this iMac to make at least 1250 RAC here. Probably closer to 1500. So, with a T7200/2ghz, I estimate it should produce a RAC of 1000-1200. The sad reality is the highest computer I could find here has a RAC of 626, and it's a windows machine: http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=514574 The highest Mac has a RAC of 502: http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=556456 We have no way of knowing if they are dedicated crunchers, so maybe their RAC could go higher. But there are couple of issues here: 1) Rosetta awards lower credits. 2) Macs with identical HW do worse than linux/windows machines here, which is a problem. At the very least, they should be equivalent. But it looks like the application for Macs is slower. If anything, intel Macs should be *faster* than windows or linux boxes. Since intel Macs started with the Core processor, there is no need for legacy support of older processors. So SSE functionality can (should) be used without needing to create/support multiple versions. |
39)
Message boards :
Number crunching :
CPU Comparison question
(Message 46457)
Posted 17 Sep 2007 by zombie67 [MM] Post: He said it was a Intel(R) Core(TM)2 CPU T7200 @ 2.00 You have a pretty fast machine. It will provide an impressive RAC here. 2-3 times what your P4 did. Edit: *And* use a lot less power ($) doing it. |
40)
Message boards :
Number crunching :
CPU Comparison question
(Message 46450)
Posted 17 Sep 2007 by zombie67 [MM] Post: Why do people offer optimized BOINC Clients? If the client is optmized for the specific chip, not just the OS, wouldn't that run faster? BOINC is open source. Anyone can compile their own BOINC *client* any way they like. For Rosetta, there is no point in doing so, as the credit method here makes over-claiming clients moot. As for optimized science *applications*, yes, they can be *much* faster using chip abilities such as SSE/2/3/4. Sometimes more than twice as fast. Since the application is not open source for Rosetta, it requires the project to offer such optimized applications. |
©2025 University of Washington
https://www.bakerlab.org