I tried many times in the last 3-4 hours without success so there must be different regions/servers if you just got a GPU on 2 accounts, or you are just very very lucky.

[QUOTE=ATH;534294]I tried many times in the last 3-4 hours without success so there must be different regions/servers if you just got a GPU on 2 accounts, or you are just very very lucky.[/QUOTE]

If they have anyone on staff that is as capable as Chris is, then I'm sure they've already started implementing automated processes to whack us as soon as we're detected.

[QUOTE=PhilF;534298]If they have anyone on staff that is as capable as Chris is, then I'm sure they've already started implementing automated processes to whack us as soon as we're detected.[/QUOTE]

Thank you for the kind words, but if they really wanted to stop us they could:

1. Contact me. I've made every effort to ensure they know what's going on.

2. Simply block the GPU72 server's DNS / IP.

2.1. The payload is downloaded from there, as is all work fetching/reporting.

I'm hoping this is simply a matter of high demand.

Just got an instance running, so let's see how long for ...

... just 48 minutes, sigh

Just obtained another instance

[QUOTE=bayanne;534375]Just got an instance running, so let's see how long for ...

... just 48 minutes, sigh

Just obtained another instance[/QUOTE]

Just lasted 20 minutes

Got 2 that have been up for about 30 minutes.

My latest error.
 

Ran for 15 minutes then on restart this...

[QUOTE=petrw1;534444]Ran for 15 minutes then on restart this...[/QUOTE]

IIRC (a bit sad to say considering it's my code) that happens if the bootstrap is killed very soon after launch.

It looks like Google doesn't like us anymore. Would have been nice if they'd just said so... :sad:

Edit: And, yet, I just now spun up two BB instances (out of three), getting P100s. Sometimes I don't even think Google understands what their AIs do. Maybe that's the plan.

roll the dice
 

Had a session run with gpu 10:47, well beyond the new nominal 10 hours, yay!
Same account got a gpu on the first try, seconds after another account could not get one. Retry that other account afterward, still can't get one in repeated tries.

Oh well, remember that it's very good for what we pay for it, which is only a freely chosen portion of our recreational time. And a cpu-only session, is up to 12+ hours of free mprime or mlucas, better than nothing, and also free. Can't really fault them for kicking us freeloaders off when someone comes along willing to pay. Did I mention they're giving this underutilized cpu and gpu time away for FREE?

I am still getting the full 12 hours on each of the four CPU instances I run (4 different Google accounts). Have already found 14 ECM factors. Not bad, considering tthe price... :cool:
I have given up GPUs completely, tho.

Colab follies
 

I wish we could get a definitive answer about what is happening here. For two days I was unable to get a session that lasted more than half an hour, and then for another day I got the "no GPU backend available" message. I thought they might be disallowing GPU72 use of colab.

However, tonight I have a P100 session that has been running for 3.5 hours as I write this.

Is there some sort of weekly quota that has come into being? We can't be completely banned since I have this session running. I wish they would respond to Chris' requests for information.