|
Stickman posted:It seems like you could also use redundant computes to weed out bad actors - X% failures and your reinbursement rate takes a hit, Y% and you lose your account (perhaps even banning the GPU device ID, though you'd want some sort amnesty for second-hand purchases). Yeah and you need a way of ensuring that your redundancy is independent. And there's also a matter of padding your hours. Say that provider X takes 4 hours to process some data and provider Y takes 4.5 hours. Did provider Y pad their hours for greater payout? Hard to say. Or if you pay based on availability instead of wall time, how do you detect fake "outages"? These are tough problems but I doubt they're insurmountable.
|
#
¿
Sep 20, 2018 02:18
|
|
|
|
| # ¿ May 21, 2024 16:22 |
|
|
Stickman posted:I assume you'd probably pay by the computation rather than hour in order to promote efficient computation (like *coin mining, but with a more useful outcome). If there are deadlines, you could add a bit extra for priority processing (i.e., the data cruncher would see the deadline for their computation component and receive a bonus if they dedicate enough resources to meet the deadline), or if the problem supports it, simply distribute small enough chunks that you can cut off processing at a certain time. I see your point, though - some problems might not quite be so granular, so there might need to be some additional incentives to meet availability targets. Normally supercomputers charge by computational time, which captures what you described. But it's very easy to spoof that to be whatever you want, if you exclusively own the hardware.
|
|
#
¿
Sep 21, 2018 01:30
|
|
|
Disco De Soto posted:I was reading bitcoin energy consumption and how proof-of-stake algorithms use way less energy compared to proof-of-work algorithms that bitcoin currently uses. Do you think it's likely that bitcoin could switch? no, bitcoin will never ever switch to proof of stake. Other cryptocurrencies may, some day, if someone can ever figure out a design that doesn't just turn into proof of work with extra steps
|
|
#
¿
Sep 21, 2018 01:48
|
|
|
Lockback posted:They pay by CPU hour, which is by computation. You can pretty easily instead attach a computational amount to each piece of work and then pay out based on what you complete. The tricky part won't be tracking what you've done, it'll be ensuring that you actually did it vs just made up something random. "Cpu hour" is computational time, it accounts for the fact that different computations take different lengths of time. "per computation" implies that time per computation is not a factor, which is not done anywhere that I know of I think that attaching a certain amount of computational time (e.g. CPU hours) to a job submission is fine. But say that I want as many digits of pi as possible and I say that I want 100 cpu hours attached to the task. Even with verification of outputs, some systems will produce a different number of digits in the same amount of computational time. How do you identify bad actors in such a situation? "You can't" may be an acceptable answer, but it would be great to solve the problem anyway
|
|
#
¿
Sep 21, 2018 02:25
|
|