A few questions about ploting and mining
I am new to burst mining and I have spend about a week on studying everything of burst, but there are still some questions I can't find a answer.
1.How does pool mining of burst work?
Is is true that the possibility of found a block by myself in pool mining the same as solo mining?
2.If I make 5 optimized plots in 1TB size, will it be much slower than make 1 optimized plot in 5TB?
My guess is 5 plots will only cost 4 more seek than 1 plot and 4 more seek time shouldn't cost much. Is that true?
3.How does plot reading speed effect mining? what should be a reasonable read time for total 5TB plots?50TB plots?100TB plots?or even 200TB plots?
I am using nfs for reading plots.It cost about 5sec to read a 1TB plot.
Pilabius last edited by Pilabius
@shadowlin Hey, I'm not much experienced but I think I could answer some of your questions. When you mine at a pool, the chances for you to find a block is more or less the same of you mining alone, the difference is that everybody mines like they are a single person, so the plots of everyone kind of add together and the chances of finding a block boosts a lot.
About the plots, in the beginning I had 3 plots of 1TB each, and made 15s to read them all (not so good CPU ):), after changing some things, I ended up with 9 plots, and the speed is the same. I think if you make 10 plots of 100GB per TB, you could see a difference in speed, but with that much I don't think so.
The speed of reading don't affect directly in mining, the thing is, if you are in a pool and your plots take 60s to read, the pool may find another block while you are still reading your plots, so the reading stop and begin from 0 for the new block, making you lose the remaining chances that you had from the total plots.
A reasonable time I think would be like 30s most.
Hope this could be of some help :)
If you find a block, you would have found it solo or pool mining. With pool mining you share your rewards with the pool, and the rest of the pool share with you, to give you a more consistent payout.
It'll be slower, and it's a lot more than 4 seeks. With a 1TB plot your have to read 1/4096th of the drive, so about 244MB. The miner will read a buffer worth of that plot, then seek to the next plot, then seek to the next, then seek to the next, then seek back to plot 1, where you read the second buffer, then seek .... If your read buffer is 1MB, then it's around 1,000 seeks to mine the drives. And seeks are expensive. With a single 4TB you seek to the first nonce, read a buffer, wait for a drive rotation read the next buffer - no more seeks required.
The faster you mine the better you are - a 31 second deadline will lose to a 32 second deadline if it takes you 33 seconds to find yours. There are so many variables tht it's hard to say what is a "reasonable" time, drive type, speed, how it's connected to the CPU, ow many cores/threads your CPU has, or the type of GPU you're using.
As an example, I have a GPU miner that does 24 drives in 34 seconds, and another miner that does 80TB in 35 seconds, even though most of the CPU cores are plotting more drives.
Thank you for your reply:)
@haitch Thank you for your reply.
About multiple plots against single plot question.I am still confused.Why wouldn't the miner read 1/4096th of the first plot try to find the target nouce then read the next plot?Why does the miner need to read part the 1/4096th of plot1 then read the next plots?
@shadowlin If there are 4 plot files on the drive, it'll rotate through the files reading and processing a buffer worth of nonces at a time. Each switch to another plot requires another seek. On Blago you can set it up so the plots are read sequentially which would cut down seeks, but it's still going to be less efficient, I believe there is no equivalent option injMiner.