| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Retired Forums
Forum: UD Windows Agent Support [Read Only]
Thread: BOINC |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 6
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi all,
Can anyone tell me the following. If so I would be very greatful. 1. Is it definately ok to use the BOINC client to contribute to the WCG? There aren't any issues on corrupt results or not getting points or anything like that? 2. When using BOINC, I seem to complete work units quite a bit faster. I have a dual core AMD machine and it runs two work units at a time. But each work unit seems to complete quite a bit faster than when running the UD client. Is there any reason for this? Does the BOINC client work on different work units or something? 3. I have an AMD X2 3600 system and a Pentium D dual core 3.0Ghz system. The AMD system seems to work noticably faster than the intel machine although both are dual core and the machines have otherwise similar specs (2 Gig ram and sata drives). Can anyone tell me if it is the case that AMD dual core chips typicaly perfrom better than their intel counterparts or is this not the case? Thanks to anyone who can advise Kindest Regards Simon |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello sh856531,
1. Is it definately ok to use the BOINC client to contribute to the WCG? There aren't any issues on corrupt results or not getting points or anything like that? Yes, BOINC is really working. We still have to do some programming on the server in order to show you the BOINC results page and to export the XML Team Statistics, but every part needed to run the project is working.2. When using BOINC, I seem to complete work units quite a bit faster. I have a dual core AMD machine and it runs two work units at a time. But each work unit seems to complete quite a bit faster than when running the UD client. Is there any reason for this? Does the BOINC client work on different work units or something? The speed is just the luck of the draw but - - Each batch of work units seems to have its own characteristics; some taking longer times, some shorter. But a work unit has to be specialized to run with a particular client and OS. So BOINC Linux is working with one batch, BOINC Windows with a second batch and UD Windows with yet another batch. The differences you observe are not caused by the client or the OS.3. I have an AMD X2 3600 system and a Pentium D dual core 3.0Ghz system. The AMD system seems to work noticably faster than the intel machine although both are dual core and the machines have otherwise similar specs (2 Gig ram and sata drives). Can anyone tell me if it is the case that AMD dual core chips typicaly perfrom better than their intel counterparts or is this not the case? First off, AMD is claiming that a 3600+ system is as fast or faster than a 3.6 Ghz Pentium, so it definitely should be faster than a 3.0 Ghz Pentium. Historically, AMD started off by lowballing their ratings so that their processors were even faster than the Intel CPU they supposedly equalled. This built up a rabid community of fans. Today their ratings occasionally overstate their performance but normally are about right. So assume that a 3600+ really is about 20% faster than a 3.0 Ghz Pentium D, unless experience shows otherwise. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
3. I have an AMD X2 3600 system and a Pentium D dual core 3.0Ghz system. The AMD system seems to work noticably faster than the intel machine although both are dual core and the machines have otherwise similar specs (2 Gig ram and sata drives). Can anyone tell me if it is the case that AMD dual core chips typicaly perfrom better than their intel counterparts or is this not the case? Yes the AMD Dual Cores *ARE* faster processors than the Intel Dual Cores. Go read www.tomshardware.com and they will tell you why. The AMD's are simply designed better, Intel kind of rushed theirs out there. If I remember right it's due mainly to the fact that the Memory Controller is onboard the CPU so the CPU's don't have to go offboard to talk to one another. Someone correct me if I'm wrong, I'm just going off my memory and it's bad sometimes :) |
|||
|
||||
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
2. When using BOINC, I seem to complete work units quite a bit faster. I have a dual core AMD machine and it runs two work units at a time. But each work unit seems to complete quite a bit faster than when running the UD client. Is there any reason for this? Does the BOINC client work on different work units or something? The above answers are good. I just wanted to add to this answer. We just finished our first batch on BOINC (good work everyone!) and we started handing out work from our second batch on Saturday. The first batch on BOINC consisted of what appear to have been on average short workunits. This second batch appear to have workunits that run much longer. As mycroft said - it is a bit of the luck of the draw. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
OK that makes sense. Thanks. I was just a bit surprised that the AMD machine was so much faster than the intel equivelant. It was a bit more than expected but I'll watch it over the coming weeks and see how they both get on.
Yes the AMD Dual Cores *ARE* faster processors than the Intel Dual Cores. Go read www.tomshardware.com and they will tell you why. The AMD's are simply designed better, Intel kind of rushed theirs out there. If I remember right it's due mainly to the fact that the Memory Controller is onboard the CPU so the CPU's don't have to go offboard to talk to one another. Someone correct me if I'm wrong, I'm just going off my memory and it's bad sometimes :) This is quite right, the architecture of the AMD processor allows for very rapid access to memory because the controller is on the proccessor die. I also seem to recall that communication between the cores is more efficient because the cores can communicate directly. In the intel chips I think communication has to go out onto the hyper transport bus. I think thats what happens. AMD did definately get the edge over intel in the most recent round of chips, and arguably the last. What will be really interesting for me though is how intel will respond. I don't think they're going to sit around too long letting AMD rule the performance stakes. With any luck Intel will pull out all the stops with their desktop chips and AMD will do the same with their mobile chips. Even better would be if intel prices were reduced a bit as well but if anything, AMD chips have been creaping up towards their intel counterparts. Oh well.... Thanks to everyone who provided answers. Much appreciated Kindest Regards Simon |
||
|
|
nealschim
Cruncher Joined: Nov 24, 2005 Post Count: 14 Status: Offline |
sh856531, just one precision: HyperTransport is for AMD ( a packet-oriented high speed low latency protocol), and it is the direct link between the cores ,thus the perfs + the mem controller inside.
----------------------------------------while intel's HT means HyperThreading. a good explanation : Alther 's post http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=4190#33710 happy crunching   |
||
|
|
|