Hello everybody.
I brought some rigs along to crunch Einstein with my GPUUG buddies whilst SETI is getting fixed.
I see what you mean about download bandwidth, I just trashed half my 50GB allowance in 2 days.
Not sure but i belive, even for 6 hosts in E@H, 25 GB on 2 days is to much, are you sure of that? Something else could be taking a ride at you conection? Did you remember to start with 0 resource share and small number of days in the cache settings? Different from SETI, a 0.5 days cache is a lot here, new WU are allways avaiable in large quantities (someone tells me the E@H limits to 300WU max, must check i´m not sure about this info).
Not sure but i belive, even for 6 hosts in E@H, 25 GB on 2 days is to much, are you sure of that? Something else could be taking a ride at you conection? Did you remember to start with 0 resource share and small number of days in the cache settings? Different from SETI, a 0.5 days cache is a lot here, new WU are allways avaiable in large quantities (someone tells me the E@H limits to 300WU max, must check i´m not sure about this info).
Yes, 25GB is way over for E@H usage, something else perhaps?
Cheers, Mike.
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
Not sure but i belive, even for 6 hosts in E@H, 25 GB on 2 days is to much, are you sure of that? Something else could be taking a ride at you conection? Did you remember to start with 0 resource share and small number of days in the cache settings? Different from SETI, a 0.5 days cache is a lot here, new WU are allways avaiable in large quantities (someone tells me the E@H limits to 300WU max, must check i´m not sure about this info).
Yes, 25GB is way over for E@H usage, something else perhaps?
Cheers, Mike.
Maybe you know, what is the normal complete bandwith DL+UL used by a normal Einstein WU? Just an aproximate figure will help.
Maybe you know, what is the normal complete bandwith DL+UL used by a normal Einstein WU? Just an aproximate figure will help.
Oooh, of the order of a few MB. We use 'locality scheduling' - making use of what you already have - precisely to reduce the ongoing bandwidth loads.
Download : what happens is that you get a 'loaf' of data which is then sliced as the days go by. For first attachment ( or re-attachment after some long break ) to E@H you also get quite a number of smaller data files as a once only ( in addition to executables which are varied occasionally ). The individual WU's are instructions to slice off from a particular part of the loaf that you probably already have - the server keeps track of you and your potential wingmen here - unless that loaf is used up in generality by the project. In which case you get another.
Upload : is a rather more compact results file.
Gary/Oliver/Bernd - what figures would you give?
Cheers, Mike.
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
Is something about 16 MB (8 x2MB) for each 1 1/2 hour WU (the average runing time with 2WU at a time on each GPU on this destroyer with an I5 and 2x670 GPUs) close to 11 MB/hour, about 0.25 GB/day for a normal average host not a top one.
In a battleship with 2x690 running 2 WU on each sub-GPU, with the same average time to run that goes to about 1GB/day. Or more on a big battleship with 3x690 super-crunchers some mates have.
That could be realy a problem for some mates how have a limited usage quota of their conection if they try to put to much hosts to work on that conection.
But they are cleary far from 25GB in just 2 days in a 6 mixed fleet.
Could anyone else verify the calculations, sorry i just in the middle of a few beers drinking task.
Well, I was away from home and it just kept downloading thousands of 2MB tasks.
I got the emails from ISP, 50%
75% then today 100%. I had to buy an extra 5GB. Got 8 hosts on it now.
My cache was set to the max.
Hello everybody. I brought
)
Hello everybody.
I brought some rigs along to crunch Einstein with my GPUUG buddies whilst SETI is getting fixed.
I see what you mean about download bandwidth, I just trashed half my 50GB allowance in 2 days.
Cheers from a VERY wet and stormy Hereford, UK.
Reggie
Not sure but i belive, even
)
Not sure but i belive, even for 6 hosts in E@H, 25 GB on 2 days is to much, are you sure of that? Something else could be taking a ride at you conection? Did you remember to start with 0 resource share and small number of days in the cache settings? Different from SETI, a 0.5 days cache is a lot here, new WU are allways avaiable in large quantities (someone tells me the E@H limits to 300WU max, must check i´m not sure about this info).
RE: Not sure but i belive,
)
Yes, 25GB is way over for E@H usage, something else perhaps?
Cheers, Mike.
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
RE: RE: Not sure but i
)
Maybe you know, what is the normal complete bandwith DL+UL used by a normal Einstein WU? Just an aproximate figure will help.
RE: Maybe you know, what is
)
Oooh, of the order of a few MB. We use 'locality scheduling' - making use of what you already have - precisely to reduce the ongoing bandwidth loads.
Download : what happens is that you get a 'loaf' of data which is then sliced as the days go by. For first attachment ( or re-attachment after some long break ) to E@H you also get quite a number of smaller data files as a once only ( in addition to executables which are varied occasionally ). The individual WU's are instructions to slice off from a particular part of the loaf that you probably already have - the server keeps track of you and your potential wingmen here - unless that loaf is used up in generality by the project. In which case you get another.
Upload : is a rather more compact results file.
Gary/Oliver/Bernd - what figures would you give?
Cheers, Mike.
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
RE: Maybe you know, what is
)
I need to run out the door, but I posted on this topic a year ago in the thread I pointed to earlier.
Here is a link to the relevant point in that thread.
With this numbers lets make
)
With this numbers lets make some calculations:
Is something about 16 MB (8 x2MB) for each 1 1/2 hour WU (the average runing time with 2WU at a time on each GPU on this destroyer with an I5 and 2x670 GPUs) close to 11 MB/hour, about 0.25 GB/day for a normal average host not a top one.
In a battleship with 2x690 running 2 WU on each sub-GPU, with the same average time to run that goes to about 1GB/day. Or more on a big battleship with 3x690 super-crunchers some mates have.
That could be realy a problem for some mates how have a limited usage quota of their conection if they try to put to much hosts to work on that conection.
But they are cleary far from 25GB in just 2 days in a 6 mixed fleet.
Could anyone else verify the calculations, sorry i just in the middle of a few beers drinking task.
Cheers!
RE: RE: Maybe you know,
)
Well done, I stand corrected. Of course it depends on the number of hosts your running .... :-)
Cheers, Mike.
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
Well, I was away from home
)
Well, I was away from home and it just kept downloading thousands of 2MB tasks.
I got the emails from ISP, 50%
75% then today 100%. I had to buy an extra 5GB. Got 8 hosts on it now.
My cache was set to the max.
Reg
Just checking the project
)
Just checking the project folder and I have about 2000 tasks remaining and quite a few are 8MB !