I think the difference is you can produce more farmable full plots with k32s.
For example say you can create 2 k32 plots per day or 1 k33.
With k32s, at the end of Day 1 you would be able start farming those 2 plots. With k33s, you would be halfway thru your 1st plot which wouldn’t be a farmable valid plot at that point. So while you can spend Day 2 farming k32s all day, you would be doing nothing but wait with k33s.
In the long run, meaning after you’ve run out of farming space, it wouldn’t matter but during that time you would have chance to respond to more challenges with k32s.
I think I understand why topic starter so mad, really all around and nobody can answer.
What I know from the code and discussions at the moment:
If you have a bigger amount of plots - you have more chances to pass the filters. I mean if you have 2 k32 - they have two attempts to pass filter, and for this space you’ll have just 1 k33 - so only one chance to pass the filter. That’s the reason Why k32 better.
The thing I’m trying to figure out right now, when a bunch of plots around the world passed the filter, for example k32, k32, k33, k32. And all of them trying to get reward. Does it mean that k33 will have 40% chance to get reward and any k32 from this list will have 20% chance to get reward?
Yes, it’s true that only k32 plots, you start farming sooner compared to k33. But personally I think the probability is already too low that the time difference is just not that relevant.
That is why I’m betting on HDD optimization.
I’m not saying this is the best possible and right way to do it. It’s just the way I’m doing it.
Hmm, i was thinking the same, that k33 and bigger will have more chances for proof checks, directly proportional with the size, but most of the people here will disagree, so… i don’t really know honestly.
Still don’t see why we all go for 100% K32 plots. I suggest considering this
The unused space (let’s say 90GB) on the drive with 100% K32 / total usable space on that drive
(Size of K33 - Size of K32) / total plotted space in all your drive
Let go through some real numbers.
For 1, let’s say it’s 90/4000 which is around 2%
For 2, (512-239) / 100,000 which is 0.27% ; assume 100TiB already plotted
If plotting time is linear, your on-going plot will add only 0.27% chance to the winning (plot space and winning chance are linearly correlated). On the other hand, optimizing space is much more important which can add 2% of winning chance.
this is a bit oversimplified because you need a combination of K32 - K35 to achieve maximum winning chance… but you see the point - it can be 10x difference
This is a better answer, there are differences a bigger plot has more chances to win, but proportional to its size, we maybe can conclude is equivalent. We need to study details of implementation of the plot and be good at math to reasoning properly on this question, not anyone is competent to answer this kind of questions.
Hello fellow farmers, first off, stay positive, never get discouraged, and try to enjoy the grind…
Of what I understand about the plots - k32s are 256gbs of code compressed into 108.8gb.plot files… They are presently recommending to the extent of pushing making k32s. Apparently not all.plot files get made with eligible code, they can have errors so making two k32s as opposed to one k33 improves your chances of saving half the effort just incase a plot gets written with erroneous code.
Windows 10 being the “bs” I mean the OS that it is can randomly restart your pc after certain time having elapsed post update (delay updates on Windows manually to prevent a mishap mid plotting) - if that kind of a thing happens to you at 98% on a k33 it is a waste of your time, ssd read write, electricity. A surge, power outage, bsod, any number of things can result in wasted effort and resources (not to forget any rage damage keyboards, mice and computer screens are likely to experience as a result).
When they say five to ten years they are expecting adoption of faster hardware making k32 challenge too easy thereby making it obsolete for generation. I don’t think existing k32s will be discarded (they are conserving the effort and resource expenditure of proof of work model in the future).
So fundamentally there seems to be no apparent edge to generating k33s for the masses, those with cutting edge hardware can go ahead and make k33s as future proof plots while putting fail-safes in place.
I may be wrong, will appreciate more input.
Day 1 of grind: 10th May 2021
For me, it is better disk usage - i have predominantly 3TB SAS drives that fit 27xK32 and 50GB space left - that is REALLY annoying even though its only 2-3%. I have on my todo list list an item to work out a better fit and look at K33 to see how much they can help
I COULD set up lots of little LVM volumes straddling drive pairs, move to ext4 and that SHOULD make space for 1 plot per disk pair - is that being OCD ? lol
One point is that I heard that we should not max (99.xx%+) out the HD because some reason… which makes the life of HD shorter…? cannot remember exactly
I think this refers to the endurance of your drives. When farming for a long time, there will be wear on your drives. There is a method to extend the lifetime when it fails, which requires you to leave unused space on the drive equal to the largest file on it. I will try to find the reddit post I think you are referring to and add a link if I find it.
One K33 provides the same chance to win as two K32s.
K32s are destined to become useless. I have read guesses of from 3 to 15 years. K33s will also become useless even further down the road if Chia grows that much.
I am plotting K33s. My 2TB NVMe plotting drive can comfortably handle 3 parallel K33 plots. Yes, I could do 7 K32s, but I do not consider this a loss.
Contrary to comments I have seen, I have found that 3 parallel K33s take 12% less time than 6 parallel K32s. This is the the second best plotting speed increase I have found. The first was NVMe plotting.
With my plotting done on a 2TB NVMe as the main and a 2TB SSD as the secondary plotting drives I now complete 3 K33s every 18 hours.
I did first performance checks before the NVMe and 3 K33s were produced in 28.5 hours. At that time 6 K32s took my machine 32 hours.
So I got a performance increase when I switched to plotting K33s and I believe they hold more long term value.
Lastly, 32 K33s fit on my 8TB (yeah right) HDDs with space left for a K32, but I don’t bother.
I would plot 2 K34s, but only 15 fit on an 8TB and I would need another NVMe to have enough plotting space.
If I am still enthused when I run out of farm space I will buy HDDs appropriate to fitting K34s and upgrade my machine to start plotting them.
There is a lot of investment in K32 thinking. In my opinion, if you are in for the long haul it makes little sense to plot K32s. One K33 is produced slightly faster than two K32s and will hold its value longer.
Good points here! I didn’t know plotting was marginally faster with K33s.
I’ve been worried about this recently actually: we keep hearing K32s will be good for another few years at least. But this is from the same people who estimated netspace growth!
How mad would you be if Bram came out next week and said “uh guys, we are moving to K33 next month. Sorry!”
That is interesting - once i have plotted my disk collection and start to replot (poolable) i am moving towards SAS HDD plotting and retiring my SSD to an easier life (an enhanced desktop experience) - SAS HDD tend to be larger drives (600GB, 900GB). When you plot to SAS HDD you usually run 1 plot per drive - so increasing to a K33 has no additional cost
