Let GPU72 decide

[TheMawn]

As someone who doesn't really know what you're talking about but is very interested, would there be a way to set up a donation box or something to let the noobz (me) be more helpful to the smart people (You, Mark Rose) and maybe get these things running a bit longer?

[chalsall]

Originally Posted by TheMawn

As someone who doesn't really know what you're talking about but is very interested, would there be a way to set up a donation box or something to let the noobz (me) be more helpful to the smart people (You, Mark Rose) and maybe get these things running a bit longer?

LOL... Thanks for that.

The most important thing for people like Mark and myself is that others learn.

Sign up for the "free" Amazon EC2 service, and experiment. Be careful what you "spin up", since some services cost "real money".

[Mark Rose]

Originally Posted by chalsall

Be careful what you "spin up", since some services cost "real money".

Indeed. Where I work our monthly AWS bill is in the five digits.

But the restrictions on your account when you first start will stop you from spending that kind of money unless you try really hard.

[Mark Rose]

Originally Posted by chalsall

But the short version is... We've probably now got about a month's worth of DC Cat 4 buffer -- LL Cat 4 is right at the edge.

And now? I dumped another 4THz-days in today haha

Also, is this the first day we found more DCTF factors than LLTF factors?

[kracker]

https://www.gpu72.com/graphs/worker/...be22735054fc5/

Nice... I'm curious, How much is that costing you daily?

[Mark Rose]

Originally Posted by kracker

https://www.gpu72.com/graphs/worker/...be22735054fc5/

Nice... I'm curious, How much is that costing you daily?

Depends which day. I ran 5 cg1.4xlarge for 1.5 days. Then I ran 10 for a day. My cost was about $0.14/hr each, so when running 10 for a day, it was costing me $33.60 per day. The total bill for the work above my normal output is $58.

[Uncwilly]

Originally Posted by Mark Rose

Depends which day. I ran 5 cg1.4xlarge for 1.5 days. Then I ran 10 for a day. My cost was about $0.14/hr each, so when running 10 for a day, it was costing me $33.60 per day. The total bill for the work above my normal output is $58.

I should look into this for my area of the project.

[Mark Rose]

Originally Posted by Uncwilly

I should look into this for my area of the project.

Do you normally use mfaktx and GPU72 for your work?

[kladner]

Originally Posted by Uncwilly

I should look into this for my area of the project.

Can anyone estimate how long a 332M LL would take on one of these Amazon units? What would be your level of confidence in the result compared to that of a high end, conservatively run, ECC PC?

Could you be sure of unbroken, or at least resumable computing resources (at least for the right price?)

[Mark Rose]

Originally Posted by kladner

Can anyone estimate how long a 332M LL would take on one of these Amazon units? What would be your level of confidence in the result compared to that of a high end, conservatively run, ECC PC?

Could you be sure of unbroken, or at least resumable computing resources (at least for the right price?)

You'll want to use new hardware that has more CPU instructions exposed to the virtual machine (i.e. C3.large, Xeon E5-2680 v2, 2 cores, about $0.033/hr spot). Spot instances might last several days, but you're almost guaranteed to be interrupted at some point. When they die, you'll have to manually recover the work (you can do a resumable spot request, but it will launch a new instance with a new filesystem). So there is that to deal with.

[Uncwilly]

Originally Posted by Mark Rose

Do you normally use mfaktx and GPU72 for your work?

I am running Prime95 on cpu's, but I feed them manually from GPU72. If I ran 1 GPU, I would use it to move exponents from 77 upward. I have 16 cores currently working on 76->77 and 2 slower cores doing 73->74.