| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Retired Forums
Forum: Member-to-Member Support [Read Only]
Thread: AMD Athlon 64 rating |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 34
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Well, I have shot my bolt. Does anybody have any ideas? [Even better, does anybody have any relevant observations?]
Lawrence |
|||
|
||||
|
debrouxl
Advanced Cruncher France Joined: Dec 31, 2004 Post Count: 61 Status: Offline Project Badges: 
|
Yesterday, I returned 5 WUs. During the crunching, the outer temperature was normal, and the processor was used by other a few tasks only a couple hours a day.
----------------------------------------Over the processors ratings I got, there's a 3, a 12 and a 16... ---------------------------------------- [Edit 1 times, last edit by DEBROUX Lionel at Mar 5, 2005 6:18:58 PM] |
||
|
|
Alther
Former World Community Grid Tech United States of America Joined: Sep 30, 2004 Post Count: 414 Status: Offline Project Badges: 
|
My Athlon 64 3000+ , with 1 GB memory, is rated 156, vs a 1.5 Ghz P4. Looks like WCG only takes into account the 1.8 Ghz CPU clock, not the real power of the engine. Is that the overall rating or the Processor rating? The processor rating is the one you want to look at if you are comparing CPUs. The overall rating is a combination of RAM, HD space allocated, and bandwidth. The agent runs a small benchmark before starting a new task. This benchmark determins the processor rating, not your clock speed. It is raw CPU power. Because it's just a benchmark, the processor rating will vary even for the same processor due to other processes running at the time the benchmark is run. e.g. if you have an anti-virus scan or are playing a game when the benchmark is run, you'll get a lower score. Same is true if you have a mobile CPU that is running in a lower clock speed for power/heat purposes.
Rick Alther
Former World Community Grid Developer |
||
|
|
spanky
Cruncher Joined: Feb 25, 2005 Post Count: 6 Status: Offline |
This is a question, not a statement or comment, from a non-technical member.
If you have already allocated "n" gig of HD; RAM is constant; assuming a cable connection; and that most people's homes are temperature regulated within a narrow, or at least consistent daily range; then wouldn't any observed variability HAVE to be the result of competition for processor resources? Wouldn't this also result in greater variability in the ratings depending upon the power of the machine (competition for resources factored in): higher power machines would be less variable than lower end PCs if competition for resources are equal? Now that my concerns about overheating have been dispelled, (thanks again, Lawrence) I am content with whatever reasonable rating the program assigns. Call this an educational question and nothing more. Thanks in advance. |
||
|
|
François
Cruncher Joined: Nov 18, 2004 Post Count: 14 Status: Offline |
Hi Lawrence,
I think that, if the small benchmark was run at normal, or above normal priority, instead of the lowest, its result would be more representative of the computing power offered to Rosetta. Personal computers are idle most of the time, this is the assumption which allows such projects to exist. Every time I open Task Manager, I see Rosetta using more than 95% of my processor. With the "small" benchmark running at low priority, if a background task eats cpu cycles for a few milliseconds while it is running, its results are impacted. Would the project gurus consider raising the priority of the benchmark? After all, if it is really "small", nobody would even notice its existence. François |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Francois,
Noted. This will really become an issue only when we have a Linux Agent. Then we will be able to make changes in this and the several different opinions can fight it out to a decision. Lawrence |
|||
|
||||
|
François
Cruncher Joined: Nov 18, 2004 Post Count: 14 Status: Offline |
Thanks Lawrence ... although I do not understand the relationship with the availability of a Linux agent.
I want to submit another observation. My Athlon 64 is set to run Cool'n Quiet, i.e. something monitors the workload, and adjusts the cpu multiplier dynamically. I just watched Rosetta finishing a work unit and starting another one. While at work, according to MSI Core Center, the cpu was running full speed @ 1.8 ghz. When it sent the result and received another job, it dropped to 1 ghz, then it came back to 1.8 ghz when the primary task started again. AMD says that the cpu can switch speeds about 30 times per second, so if the benchmark runs for a few milliseconds, there is some chance that it has run at the lower speed. This may be the reason why my cpu is rated at 1 ghz only. I realize that this thread is becoming kind of academic, as it relates to the points system, while the only meaningful number is how many results are returned. Cheers, François |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Francois,
Right now we are using essentially the same client Agent that United Devices is running on http://www.grid.org with a few changes to access our server, etc. We do not have the source code, so any additional changes have to be contracted with United Devices. Under license, we can modify Rosetta if that becomes necessary. At least, this is my understanding. Once we have our own client Agent (Linux first), we can make changes and experiment if that seems desirable. I think you might be right about CoolnQuiet dropping the speed of the Athlon 64. One way to check that is to have ThrottleWatch running while a work unit is returned. That will catch any changes in speed. If that is happening, you might want to use Motherboard Monitor to decide whether or not you should look into additional cooling for your computer. I remember one member running a computer of about my speed reporting a temperatue 20-25 degrees C higher than my own computer. So there are many variations. Lawrence |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Just for interest, I am using a Sempron 3000+ and the processor rating is 195
 Millmaag |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
millmaag,
You need to post this on the 'Processor Score' thread at http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=1146 so that RT can add it to the database. He has more than 800 processor scores. Lawrence |
|||
|
||||
|
|