-
|
While I understand that the idea of someone possibly being able to re-plot (phase 1) in under 5 minutes could be an issue I would like to have a better understanding of whether or not this is a realistic scenario. Since plotting does use resources would it really be worth it to plot a k=31 plot between every block (just to get a relatively small increase in effective space)? It seems to be that the k=32 minimum plot size might be prohibitive to smaller farmers (some people have 50-150gb free on their pc and not enough space for a 300gb temp file). And if the goal is to be as decentralized as possible it seems to me that these smaller farmers should be encouraged. Additionally, allowing smaller plot sizes would make it easier to plot on an SSD with limited space and then move the file to your hard drive. In my case I have 100gb free on my SSD but can unfortunately not utilize this to plot the 300gb temp files. This means I have to plot directly onto my 8tb hard drive which further increases the time needed to plot everything. I am relatively new so it is very possible that I have overlooked something, maybe some of you could give me some more insight. Thanks in advance! :) |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 6 replies
-
|
There are two additional issues that have to be kept in mind. The first is that the single fastest way to plot through only phase 1 of a plot - as one can theoretically farm just phase 1 - is to use a very large memory machine and perform all the sorts in memory. That's mostly limited by the CPU to RAM speed actually and it is MUCH faster than you'd expect already today. It's not cost effective and has a very low per day TB of plots written, but it's an attack vector that can turn Proof of Space back into Proof of Work. You can create lots more farmable TB by plotting slower and as parallel as you can. The plotting and farming process is designed to be very parallelize-able. The second is that all k sizes will eventually go up due to a mix of Moore's and Kryder's laws. A k=31 would be likely retired much sooner than a k=32. At k=32 and knowing what we do about RAM and disk prices and technologies today - we think that k=32 will make it certainly 7 years and probably 10+ years. There are always tradeoffs but we think erring on the side of lowering the amount of re-plotting necessary and keeping our security constants pretty high is the right trade. -Gene |
Beta Was this translation helpful? Give feedback.
-
|
Can someone explain to me aft the plots are going into the temp ssd/space then it moves on the Permanent hdd (in my case my nas). |
Beta Was this translation helpful? Give feedback.
-
|
Chiming in to state the probably obvious, but as the typo literally negates the meaning of the answer to the very question that was posed, I think it's worth pointing out that what @dorkmo wanted to say was: Once you fill up your desired HDDs you can stop plotting and simply farm. |
Beta Was this translation helpful? Give feedback.
-
|
WHen my plot are moving to my HDD after 100% ploting, the size is just about 4.2GO.........i really do'nt understand why, somebody have an idea? |
Beta Was this translation helpful? Give feedback.
-
|
Check your block size. |
Beta Was this translation helpful? Give feedback.
-
|
Everybody suggests to plot k32 instead of k34, I did not follow it because I thought that I would look a little bit ahead as k32 will go in a few years. Seven? C'mon, when bitcoin was created they said that the last coin would be mined sometime in 2100, but they counted it without taking into consideration that GPUs will be used, then GPU rigs, pools and finally ... ASICs. Number of bitcoins mined is 90%. So I think that k32 will go much quicker. Who knows what technology brings? I predict that one day home computers' RAM will be in TBs and HDDs will become virtual in RAM. Proof of Space&Time might become Proof of Work again, as @hoffmang9 says. I am not patient to plot one k34 in one month, I better plot 10 k32s in 10 days and start farming. But surely future holds larger k-sizes. Just like with bitcoin, the entry level is not accessible to a simple home computer, Chia might become same type of cryptocurrency, too. |
Beta Was this translation helpful? Give feedback.
-
|
The obsolescence of k32 in less than a decade would require a fundamental - likely discovery in physics - change in the current CPU and RAM roadmaps from all the major players. This isn't like Bitcoin in that way. The ultimate hardware methods are generally understood and the economics strongly force people to just plot k32s instead of trying to grind them. |
Beta Was this translation helpful? Give feedback.
There are two additional issues that have to be kept in mind.
The first is that the single fastest way to plot through only phase 1 of a plot - as one can theoretically farm just phase 1 - is to use a very large memory machine and perform all the sorts in memory. That's mostly limited by the CPU to RAM speed actually and it is MUCH faster than you'd expect already today. It's not cost effective and has a very low per day TB of plots written, but it's an attack vector that can turn Proof of Space back into Proof of Work. You can create lots more farmable TB by plotting slower and as parallel as you can. The plotting and farming process is designed to be very parallelize-able.
The second is…