As of today the ABP1 unit has the shortest time among my 5 projects. One CPDN task took me 1961 hours, AQUA@home 128-qubit takes about 200 hours, Astropulse 69 hours, QMC about 50 hours on my Opteron 1210 running Linux at 1.8GHz. So I shall not frightened by any "monster".
Tullio
That´s the problem and I see another one for the faster machines:
You must have a partner for the WU, who is nearly as fast as you. If he is not, I mention, that we will have a huge amount of pending credits for a lot of time. I mean the devs must think about that.
Yeah, that's one aspect.
Anyway, I'm not going to abort anything just yet. I'll hang on and see what Bernd and crew say about it.
I will do the same. Just wait and see what´s coming from the scientists. The only thing I wanted to know was, if the WU is faulty or not. I´m sure now, that the WU isn´t faulty!
@Tullio: Fear? I have no fear, because I´m crunching CPDN since 2005. ;o)
I will do the same. Just wait and see what´s coming from the scientists. The only thing I wanted to know was, if the WU is faulty or not. I´m sure now, that the WU isn´t faulty!
@Tullio: Fear? I have no fear, because I´m crunching CPDN since 2005. ;o)
There's a big difference between CPDN and E@H Arecibo WUs. CPDN gives trickle credits, but with E@H it's all or nothing. If the E@H WU crashes after 1000+hrs, you get nada, zilch, zero!
I have got also a huge WU to crunch. It has done 18.5 hours and will take more than 41 hours to complete. That is however not correct as in hte 18.5 hours only 6.6% was done. There is nothing on the main page of such monster WU's so I think this is an error and I have aborted this WU. I wil see if I was mistaken to do so.
Also AQUA@home has some 1000+ hours units. But they seem to have chosen the GPU/CUDA road rather than trickles, opposite to climateprediction.net.
Tullio
We need a decision one way or the other on canceling these pretty soon.
Otherwise, local circumstances will make the decision for you and that will most likely be an abort. ;-)
Alinator
Here, with my hyper-threaded dual Xeon machine, just having the suspended workunit was keeping me from downloading another workunit, that would have kept the fourth virtual processor active. On my X9650 machine, I fortunately had already downloaded a 10-day cache, so I've had enough to keep all four cores busy.
Of course, on any machines with only one core, one of these bad workunits will prevent the downloading of more good workunits.
As of today the ABP1 unit has
)
As of today the ABP1 unit has the shortest time among my 5 projects. One CPDN task took me 1961 hours, AQUA@home 128-qubit takes about 200 hours, Astropulse 69 hours, QMC about 50 hours on my Opteron 1210 running Linux at 1.8GHz. So I shall not frightened by any "monster".
Tullio
RE: That´s the problem and
)
Yeah, that's one aspect.
Anyway, I'm not going to abort anything just yet. I'll hang on and see what Bernd and crew say about it.
I will do the same. Just wait
)
I will do the same. Just wait and see what´s coming from the scientists. The only thing I wanted to know was, if the WU is faulty or not. I´m sure now, that the WU isn´t faulty!
@Tullio: Fear? I have no fear, because I´m crunching CPDN since 2005. ;o)
RE: I will do the same.
)
There's a big difference between CPDN and E@H Arecibo WUs. CPDN gives trickle credits, but with E@H it's all or nothing. If the E@H WU crashes after 1000+hrs, you get nada, zilch, zero!
Seti Classic Final Total: 11446 WU.
This is a problem that the
)
This is a problem that the project leaders must solve, when they will sent out more of this WU.
I have got also a huge WU to
)
I have got also a huge WU to crunch. It has done 18.5 hours and will take more than 41 hours to complete. That is however not correct as in hte 18.5 hours only 6.6% was done. There is nothing on the main page of such monster WU's so I think this is an error and I have aborted this WU. I wil see if I was mistaken to do so.
Greetings from
TJ
Also AQUA@home has some 1000+
)
Also AQUA@home has some 1000+ hours units. But they seem to have chosen the GPU/CUDA road rather than trickles, opposite to climateprediction.net.
Tullio
Hi guys, Please have a
)
Hi guys,
Please have a look at my reply here...
Cheers,
Oliver
Einstein@Home Project
As I said in my reply in the
)
As I said in my reply in the Q & A:
We need a decision one way or the other on canceling these pretty soon.
Otherwise, local circumstances will make the decision for you and that will most likely be an abort. ;-)
Alinator
RE: As I said in my reply
)
Here, with my hyper-threaded dual Xeon machine, just having the suspended workunit was keeping me from downloading another workunit, that would have kept the fourth virtual processor active. On my X9650 machine, I fortunately had already downloaded a 10-day cache, so I've had enough to keep all four cores busy.
Of course, on any machines with only one core, one of these bad workunits will prevent the downloading of more good workunits.