My basic complaint is that some staffer at UWM Einstein physics divides 24 by 1 and arrives at a result of 16 (this is basic physics apparently). I have three rigs infected with the 1 hour and change WU's. Why should I be invited to waste CPU cycles doing work anywhere else? Why would any project suggest this if they are on max track?
The purpose of E@H is to help the researchers find gravitational wave signals. We volunteer our CPU time, to be available as needed, according to the priority decided by the investigators. I have yet to see any statement of obligation of the type you are implying. I'm sure other projects would be happy to have some of the horsepower!
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Mike; You are correct. Still, can you explain why 24 divided by 1 equals 16? Thanks for your response. Best wishes to all of you with max rigs humping the 1 hour curse. Let me know how you handle it.
tweakster:
I don’t know what the E@H devs are doing but I believe their uppermost concern is the stability of the project as a whole. I have seen a lot cases where people join E@H grab a bunch of WUs and then do nothing further. Most of those people probably did not know that they were screwing up but I there are some who will make a mess of others efforts out of pure maliciousness. Even a quota of 24 would not resolve the problem entirely, one of my systems is finishing some WUs in less than 40 minutes.
I feel confident that the E@H devs are working on a solution that will gather the best that all of us have to offer. We might wish that they had given more thought to how the shorter WUs were going to impact high performance users. They have admitted being taken by surprise by the ultra-short WUs and perhaps under estimating their impact.
We are both dedicated E@H contributor you obviously more so than I. I for one never believed for a second that opinion expressed by some that we should move on to other projects reflected the opinion of the E@H devs. If I am wrong I hope they will say so.
It wasn’t so long ago that the pinch was on the low performance users they toughed it out. I intend to and hope to continue seeing you out there ahead of me.
Mark; You have a nice 4 rig farm. My compliments on your crunch power. I think an Opty 148 is in my budget for next month. Let us know how you handle the 1 hour nasties. Trying to save my download potential, I will be uploading at 5:30 am CST 3 Jan. Who knew it would get this fun?
Mike; You are correct. Still, can you explain why 24 divided by 1 equals 16?
Regards-tweakster
Tweakster: I gave one possible reason for raising the quota slowly in this posting.
Please do read it - there are issues you have not thought of - or at least have not mentioned here - and I am not going to type out my whole posting again over here.
The whole thread it comes from is also worth reading if you have time. Consensus there after some discussion is that the project do need to raise the limit again, and that they were right to do so slowly, and that future increases should again be small. Aim is to get the quota big enough but only just big enough for the fastest box and the fastest wu. To find out why, click one of the links I've given.
And if you have any new angles on the issues yourself, it would be helpful to post them over there.
It is not the short WUs per say that cause the problem but rather getting too many of them consecutively. If the scheduler could be modified to allocate only a limited number of each speed to any system per day then the current quota may be adequate. My powermac11,2 seems to be operating in this mode, my opteron 275 does not. I do not know if this is the result of some kind of statistical fluke or what.
I messaged these boards on December 29th about the problem of 1 hour Albert work units on fast rigs. As of 8 January, I currently have 4 rigs stuck with a large population of these same nasties. This is not a big deal for my fellow crunchers running a Celeron. I have a Sempy running stable at 2.55 ghz, an A64 at 2.6 ghz, an A64 claw at 2.4 ghz, and another A64 at 2.47 ghz. Why these rigs are getting hit with these nickle and dime WU's I don't understand. No one from the Einstein team has responded regarding this mismatch. Do the math. We have daily quotas of 16. So 24 divided by 1 equals 16? Education please.
I?m not sure whether I?m right here - but I recognized a program called albert_4.37_windows_intelx86.exe on my PC. And the effect of this program is that I get a "no response" message when I try to access my 2. CD ROM Drive. Every thing hangs. I can?t kill this job with the task manager and my CD ROM Drive is dead.
I have to reboot the PC. I don?t think that this is the purpose of the program.
What is albert_4.37_windows_intelx86.exe doing ? How can I force it NOT to kill my CD ROM drive ?
There are three possibilities of why your CD drive is not working. One possibility is that there is something wrong in your hardware like a fried CD drive, loose cable, or fried motherboard. Another possibility is that you have a problem in the software like being infected with the Sony BMG rootkit or having that rootkit removed improperly. A third possibility is that your CD is scratched and needs to be fixed.
In the last 48 hours I have one rig that has gone to ground twice. This box was dead out of work with no relief in sight. My only recourse was detach and reattach. I have three other rigs where this was necessary. This carnival of new host ID's leads me to wonder as to the thinking of the Einstein staff. Since December 29th, I have posted on these boards about the 1 hour nasties and received zip response from the Admins. They have in fact adjusted the downloads to include larger work units to offset the work total. But all too many of the One hour specials are still available to the fast machines. Anybody back from holiday yet?
In the last 48 hours I have one rig that has gone to ground twice. This box was dead out of work with no relief in sight. My only recourse was detach and reattach. I have three other rigs where this was necessary. This carnival of new host ID's leads me to wonder as to the thinking of the Einstein staff. Since December 29th, I have posted on these boards about the 1 hour nasties and received zip response from the Admins. They have in fact adjusted the downloads to include larger work units to offset the work total. But all too many of the One hour specials are still available to the fast machines. Anybody back from holiday yet?
Regards-tweakster
Let me rephrase this post as I see it.
I don't care about any back end processes, further working being done on the project nor any negative ramifications that might occur by increasing the quota, as long as I get want I want. I will also continue to get what I want by any means necessary regardless of any problems that will be created for the servers and project. I also realize that I am the most important member of this project, therefore I deserve... No... I demand an immediate and personal reply to any wish, whim, or gripe I might have.
Tweakster, take a Xanax and relax. Did you ever think that those "1 hour curse" WU's might be just as valuable as the other WU's?
RE: My basic complaint is
)
???? 'complaint' 'some staffer' 'infected' 'waste' ????
Chill out! :-)
The purpose of E@H is to help the researchers find gravitational wave signals. We volunteer our CPU time, to be available as needed, according to the priority decided by the investigators. I have yet to see any statement of obligation of the type you are implying. I'm sure other projects would be happy to have some of the horsepower!
I have made this letter longer than usual because I lack the time to make it shorter ...
... and my other CPU is a Ryzen 5950X :-) Blaise Pascal
Mike; You are correct.
)
Mike; You are correct. Still, can you explain why 24 divided by 1 equals 16? Thanks for your response. Best wishes to all of you with max rigs humping the 1 hour curse. Let me know how you handle it.
Regards-tweakster
tweakster: I don’t know
)
tweakster:
I don’t know what the E@H devs are doing but I believe their uppermost concern is the stability of the project as a whole. I have seen a lot cases where people join E@H grab a bunch of WUs and then do nothing further. Most of those people probably did not know that they were screwing up but I there are some who will make a mess of others efforts out of pure maliciousness. Even a quota of 24 would not resolve the problem entirely, one of my systems is finishing some WUs in less than 40 minutes.
I feel confident that the E@H devs are working on a solution that will gather the best that all of us have to offer. We might wish that they had given more thought to how the shorter WUs were going to impact high performance users. They have admitted being taken by surprise by the ultra-short WUs and perhaps under estimating their impact.
We are both dedicated E@H contributor you obviously more so than I. I for one never believed for a second that opinion expressed by some that we should move on to other projects reflected the opinion of the E@H devs. If I am wrong I hope they will say so.
It wasn’t so long ago that the pinch was on the low performance users they toughed it out. I intend to and hope to continue seeing you out there ahead of me.
Mark; You have a nice 4 rig
)
Mark; You have a nice 4 rig farm. My compliments on your crunch power. I think an Opty 148 is in my budget for next month. Let us know how you handle the 1 hour nasties. Trying to save my download potential, I will be uploading at 5:30 am CST 3 Jan. Who knew it would get this fun?
Regards-tweakster
RE: Mike; You are correct.
)
Tweakster: I gave one possible reason for raising the quota slowly in this posting.
Please do read it - there are issues you have not thought of - or at least have not mentioned here - and I am not going to type out my whole posting again over here.
The whole thread it comes from is also worth reading if you have time. Consensus there after some discussion is that the project do need to raise the limit again, and that they were right to do so slowly, and that future increases should again be small. Aim is to get the quota big enough but only just big enough for the fastest box and the fastest wu. To find out why, click one of the links I've given.
And if you have any new angles on the issues yourself, it would be helpful to post them over there.
R~~
~~gravywavy
It is not the short WUs per
)
It is not the short WUs per say that cause the problem but rather getting too many of them consecutively. If the scheduler could be modified to allocate only a limited number of each speed to any system per day then the current quota may be adequate. My powermac11,2 seems to be operating in this mode, my opteron 275 does not. I do not know if this is the result of some kind of statistical fluke or what.
I messaged these boards on
)
I messaged these boards on December 29th about the problem of 1 hour Albert work units on fast rigs. As of 8 January, I currently have 4 rigs stuck with a large population of these same nasties. This is not a big deal for my fellow crunchers running a Celeron. I have a Sempy running stable at 2.55 ghz, an A64 at 2.6 ghz, an A64 claw at 2.4 ghz, and another A64 at 2.47 ghz. Why these rigs are getting hit with these nickle and dime WU's I don't understand. No one from the Einstein team has responded regarding this mismatch. Do the math. We have daily quotas of 16. So 24 divided by 1 equals 16? Education please.
Regards-tweakster
RE: I?m not sure whether
)
There are three possibilities of why your CD drive is not working. One possibility is that there is something wrong in your hardware like a fried CD drive, loose cable, or fried motherboard. Another possibility is that you have a problem in the software like being infected with the Sony BMG rootkit or having that rootkit removed improperly. A third possibility is that your CD is scratched and needs to be fixed.
In the last 48 hours I have
)
In the last 48 hours I have one rig that has gone to ground twice. This box was dead out of work with no relief in sight. My only recourse was detach and reattach. I have three other rigs where this was necessary. This carnival of new host ID's leads me to wonder as to the thinking of the Einstein staff. Since December 29th, I have posted on these boards about the 1 hour nasties and received zip response from the Admins. They have in fact adjusted the downloads to include larger work units to offset the work total. But all too many of the One hour specials are still available to the fast machines. Anybody back from holiday yet?
Regards-tweakster
RE: In the last 48 hours I
)
Let me rephrase this post as I see it.
I don't care about any back end processes, further working being done on the project nor any negative ramifications that might occur by increasing the quota, as long as I get want I want. I will also continue to get what I want by any means necessary regardless of any problems that will be created for the servers and project. I also realize that I am the most important member of this project, therefore I deserve... No... I demand an immediate and personal reply to any wish, whim, or gripe I might have.
Tweakster, take a Xanax and relax. Did you ever think that those "1 hour curse" WU's might be just as valuable as the other WU's?
Jim