Due to the confusion surrounding the points/time issue with the new R2 jobs particularly relating to AMD chips, there probably have been some that have left, at least for the time being.
I would think there as been enought time that some heavy duty cruncher who has various versions of AMD Athlons in his/her arsenal could post some ball park numbers that would give the AMD (as well as Intel) users some idea how they can maximize their time.
Personally I would be interested in the performance of the older Athlons (Thunderbird core) which I have several. I should have had a few jobs completed myself, but due to some telephone outages in the area (yes, all I can get is dial-up) I am only beginning to see some results with my faster 64bit machine which were as predicted, down 25-30%. Most posts that I have read seem to say that the 3D Now! technology in the K6s and K-7 Athlons do as well as Pentiums with SSE, but some say they are slower as well.
Does anyone have anything definitive concerning the older Athlons?
F. Prefect
In the beginning the Universe was created. This has made a lot of people very angry and been widely regarded as a bad move.....Douglas Adams
Due to the confusion surrounding the points/time issue with the new R2 jobs particularly relating to AMD chips, there probably have been some that have left, at least for the time being.
I would think there as been enought time that some heavy duty cruncher who has various versions of AMD Athlons in his/her arsenal could post some ball park numbers that would give the AMD (as well as Intel) users some idea how they can maximize their time.
Personally I would be interested in the performance of the older Athlons (Thunderbird core) which I have several. I should have had a few jobs completed myself, but due to some telephone outages in the area (yes, all I can get is dial-up) I am only beginning to see some results with my faster 64bit machine which were as predicted, down 25-30%. Most posts that I have read seem to say that the 3D Now! technology in the K6s and K-7 Athlons do as well as Pentiums with SSE, but some say they are slower as well.
Does anyone have anything definitive concerning the older Athlons?
F. Prefect
Well, the first thing is that unless I have been reading the progress reports incorrectly right now it doesn't matter what kind of SIMD support your processor has, the current app doesn't take advantage of it. IOW, we are all looking at the FPU horsepower only.
Also, Akos posted earlier that 3DNow! is essentially equilvalent to SSE terms of performance, although I don't recall that he characterized how that compared to 3DNow! Pro or the later versions of SSE.
Finally, with the S5R1/I apps, IIRC the optimizations used MMX for the older Intel and AMD processors and SSE for the newer ones which support it, as well as other techniques. I'm not 100% sure about the MMX part, but something had to have been done because the improved app was a lot faster on my K6-x's.
The most likely reason the newer AMD's have been getting beaten up so much is that the Intels have always enjoyed somewhat stronger FPU peformance than AMDs. IIRC, when Athlon came out AMD closed the gap, but apparently Intel still has still has a pretty good advantage there when you have to rely on the FPU alone.
Due to the confusion surrounding the points/time issue with the new R2 jobs particularly relating to AMD chips, there probably have been some that have left, at least for the time being.
I would think there as been enought time that some heavy duty cruncher who has various versions of AMD Athlons in his/her arsenal could post some ball park numbers that would give the AMD (as well as Intel) users some idea how they can maximize their time.
Personally I would be interested in the performance of the older Athlons (Thunderbird core) which I have several. I should have had a few jobs completed myself, but due to some telephone outages in the area (yes, all I can get is dial-up) I am only beginning to see some results with my faster 64bit machine which were as predicted, down 25-30%. Most posts that I have read seem to say that the 3D Now! technology in the K6s and K-7 Athlons do as well as Pentiums with SSE, but some say they are slower as well.
Does anyone have anything definitive concerning the older Athlons?
F. Prefect
Well, the first thing is that unless I have been reading the progress reports incorrectly right now it doesn't matter what kind of SIMD support your processor has, the current app doesn't take advantage of it. IOW, we are all looking at the FPU horsepower only.
Also, Akos posted earlier that 3DNow! is essentially equilvalent to SSE terms of performance, although I don't recall that he characterized how that compared to 3DNow! Pro or the later versions of SSE.
Finally, with the S5R1/I apps, IIRC the optimizations used MMX for the older Intel and AMD processors and SSE for the newer ones which support it, as well as other techniques. I'm not 100% sure about the MMX part, but something had to have been done because the improved app was a lot faster on my K6-x's.
The most likely reason the newer AMD's have been getting beaten up so much is that the Intels have always enjoyed somewhat stronger FPU peformance than AMDs. IIRC, when Athlon came out AMD closed the gap, but apparently Intel still has still has a pretty good advantage there when you have to rely on the FPU alone.
Alinator
Beaten up may be a little too kind, at least when it comes to performing this particular task. Destroyed comes to mind. But AMDs have advantages in other areas. Like I stated in my earlier post the only machine that I have at least 5 completed jobs to do a comparison, the Athlon 64 XP is approx. 25-30% slower running the R2s. But it looks like the newer 64bit Athlons, as well as the older ones may be more on par with comparable Pentiums. A few may move machines with certain chips to different projects, but I suspect those are the ones more interested in the competition than in the science.
F. Prefect
In the beginning the Universe was created. This has made a lot of people very angry and been widely regarded as a bad move.....Douglas Adams
Beaten up may be a little too kind, at least when it comes to performing this particular task. Destroyed comes to mind. But AMDs have advantages in other areas. Like I stated in my earlier post the only machine that I have at least 5 completed jobs to do a comparison, the Athlon 64 XP is approx. 25-30% slower running the R2s. But it looks like the newer 64bit Athlons, as well as the older ones may be more on par with comparable Pentiums. A few may move machines with certain chips to different projects, but I suspect those are the ones more interested in the competition than in the science.
F. Prefect
LOL...
Agreed, I imagine going back to a plain vanilla app has been an eye-popping experience for a lot folks! ;-)
That being said, when serious 'Akosification' is done to S5R2 things will improve tremendously on the runtime front. Unfortunately, given the Cross Project Parity issue, we're all going to take a bite from the you-know-what sandwich on the rates compared to S5R1/I.
For anyone who's intersted, here are some updated results for S5R2 workunits on my machines. (These represent the largest 'R2's that each machine has completed.)
Pentium II 333, running Xubuntu 7.04
Total Time: 79.7 hours
Cobblestones awarded: 166.68
Pentium III 450 (Katmai), running Ubuntu 6.06
Total Time: 89.35 hours
Cobblestones awarded: 255.83
The largest workunit I've received so far is still in progress. It's running on the slowest machine I have going, an iMac with a G3 400. My wingman already received 422.58 Cobblestones for this one, and I'm on track to finish it after a total of 160 hours. (Yeah, a real speed-demon, that iMac.)
For anyone who's intersted, here are some updated results for S5R2 workunits on my machines. (These represent the largest 'R2's that each machine has completed.)
Pentium II 333, running Xubuntu 7.04
Total Time: 79.7 hours
Cobblestones awarded: 166.68
Pentium III 450 (Katmai), running Ubuntu 6.06
Total Time: 89.35 hours
Cobblestones awarded: 255.83
The largest workunit I've received so far is still in progress. It's running on the slowest machine I have going, an iMac with a G3 400. My wingman already received 422.58 Cobblestones for this one, and I'm on track to finish it after a total of 160 hours. (Yeah, a real speed-demon, that iMac.)
Thanks to all you guys. That will help me decide whether to only run E@H, Rosetta, or a combination of both.
But the way a combination of both is working, that alternative has just about seen it's days. For some reason on my Athlon 64XP 3200, I can only recieve
1 (!) E@H job at a time and only after the previous job is completed. It's set up just like a couple of other machines running both and I get plenty of work for both. All I'm using is the default settings, (no school,work, etc)but to get a new E@H job when the previous one if finished I have to suspend Rosetta and wait a few minutes before anything happens. Crazy thing.
thanks,
F. Prefect
In the beginning the Universe was created. This has made a lot of people very angry and been widely regarded as a bad move.....Douglas Adams
Thanks to all you guys. That will help me decide whether to only run E@H, Rosetta, or a combination of both.
But the way a combination of both is working, that alternative has just about seen it's days. For some reason on my Athlon 64XP 3200, I can only recieve
1 (!) E@H job at a time and only after the previous job is completed. It's set up just like a couple of other machines running both and I get plenty of work for both. All I'm using is the default settings, (no school,work, etc)but to get a new E@H job when the previous one if finished I have to suspend Rosetta and wait a few minutes before anything happens. Crazy thing.
thanks,
F. Prefect
Check your Active Fraction value for that machine. I had one system where it went down near zero and wouldn't download anything. I edited it back up near .9999 and it started downloading again. It's in file client_state.xml under parm .
Always be sure to shutdown BOINC and save your files before making changes.
RE: I believe Account
)
Hmmm, http://www.boincstats.com/stats/project_graph.php?pr=mcp&view=users shows a handful of new members in the past few days, maybe they allow only as many in as there are people leaving ?? I'll try .
CU
BRM
RE: RE: I believe
)
Due to the confusion surrounding the points/time issue with the new R2 jobs particularly relating to AMD chips, there probably have been some that have left, at least for the time being.
I would think there as been enought time that some heavy duty cruncher who has various versions of AMD Athlons in his/her arsenal could post some ball park numbers that would give the AMD (as well as Intel) users some idea how they can maximize their time.
Personally I would be interested in the performance of the older Athlons (Thunderbird core) which I have several. I should have had a few jobs completed myself, but due to some telephone outages in the area (yes, all I can get is dial-up) I am only beginning to see some results with my faster 64bit machine which were as predicted, down 25-30%. Most posts that I have read seem to say that the 3D Now! technology in the K6s and K-7 Athlons do as well as Pentiums with SSE, but some say they are slower as well.
Does anyone have anything definitive concerning the older Athlons?
F. Prefect
In the beginning the Universe was created. This has made a lot of people very angry and been widely regarded as a bad move.....Douglas Adams
RE: Due to the confusion
)
Well, the first thing is that unless I have been reading the progress reports incorrectly right now it doesn't matter what kind of SIMD support your processor has, the current app doesn't take advantage of it. IOW, we are all looking at the FPU horsepower only.
Also, Akos posted earlier that 3DNow! is essentially equilvalent to SSE terms of performance, although I don't recall that he characterized how that compared to 3DNow! Pro or the later versions of SSE.
Finally, with the S5R1/I apps, IIRC the optimizations used MMX for the older Intel and AMD processors and SSE for the newer ones which support it, as well as other techniques. I'm not 100% sure about the MMX part, but something had to have been done because the improved app was a lot faster on my K6-x's.
The most likely reason the newer AMD's have been getting beaten up so much is that the Intels have always enjoyed somewhat stronger FPU peformance than AMDs. IIRC, when Athlon came out AMD closed the gap, but apparently Intel still has still has a pretty good advantage there when you have to rely on the FPU alone.
Alinator
RE: RE: Due to the
)
Beaten up may be a little too kind, at least when it comes to performing this particular task. Destroyed comes to mind. But AMDs have advantages in other areas. Like I stated in my earlier post the only machine that I have at least 5 completed jobs to do a comparison, the Athlon 64 XP is approx. 25-30% slower running the R2s. But it looks like the newer 64bit Athlons, as well as the older ones may be more on par with comparable Pentiums. A few may move machines with certain chips to different projects, but I suspect those are the ones more interested in the competition than in the science.
F. Prefect
In the beginning the Universe was created. This has made a lot of people very angry and been widely regarded as a bad move.....Douglas Adams
RE: Beaten up may be a
)
LOL...
Agreed, I imagine going back to a plain vanilla app has been an eye-popping experience for a lot folks! ;-)
That being said, when serious 'Akosification' is done to S5R2 things will improve tremendously on the runtime front. Unfortunately, given the Cross Project Parity issue, we're all going to take a bite from the you-know-what sandwich on the rates compared to S5R1/I.
Alinator
Here are my teams latest S5R2
)
Here are my teams latest S5R2 results if they are any use to anyone.
AMD 850, 448MB RAM, Win XP, 37.64H for 162 credit
P4 3Ghz, +HT, 256MB RAM, Win XP, 24.16H for 173 credit
P4 3Ghz, +HT, 512MB RAM, Win XP, 60.23H for 430 credit
P4 1.4Ghz, 256MB RAM, Win XP, 70.23H for 412 credit
Mrs Miggins - A lady of uncommon refinement
For anyone who's intersted,
)
For anyone who's intersted, here are some updated results for S5R2 workunits on my machines. (These represent the largest 'R2's that each machine has completed.)
Athlon64 x2 6000+, running Kubuntu 7.04
Total Time: 12.5 hours
Cobblestones awarded: 332.57
Athlon64 3500+, running WindowsXP
Total Time: 19 hours
Cobblestones awarded: 204.68
Duron 950, running Startcom 4 Linux
Total Time: 47.6 hours
Cobblestones awarded: 319.38
Athlon 700, running Edubuntu 7.04
Total Time: 48.4 hours
Cobblestones awarded: 246.66
Pentium II 333, running Xubuntu 7.04
Total Time: 79.7 hours
Cobblestones awarded: 166.68
Pentium III 450 (Katmai), running Ubuntu 6.06
Total Time: 89.35 hours
Cobblestones awarded: 255.83
The largest workunit I've received so far is still in progress. It's running on the slowest machine I have going, an iMac with a G3 400. My wingman already received 422.58 Cobblestones for this one, and I'm on track to finish it after a total of 160 hours. (Yeah, a real speed-demon, that iMac.)
Recent results: Barton 1.8ghz
)
Recent results:
Barton 1.8ghz on Win2k3
Running time: 189,666 sec (~52h 41m)
Credit: 438.18
Running time: 183,575 sec (~50h 59m)
Credit: 438.16 (still pending)
RE: For anyone who's
)
Thanks to all you guys. That will help me decide whether to only run E@H, Rosetta, or a combination of both.
But the way a combination of both is working, that alternative has just about seen it's days. For some reason on my Athlon 64XP 3200, I can only recieve
1 (!) E@H job at a time and only after the previous job is completed. It's set up just like a couple of other machines running both and I get plenty of work for both. All I'm using is the default settings, (no school,work, etc)but to get a new E@H job when the previous one if finished I have to suspend Rosetta and wait a few minutes before anything happens. Crazy thing.
thanks,
F. Prefect
In the beginning the Universe was created. This has made a lot of people very angry and been widely regarded as a bad move.....Douglas Adams
RE: Thanks to all you
)
Check your Active Fraction value for that machine. I had one system where it went down near zero and wouldn't download anything. I edited it back up near .9999 and it started downloading again. It's in file client_state.xml under parm .
Always be sure to shutdown BOINC and save your files before making changes.
Seti Classic Final Total: 11446 WU.