Plotting with Dynamic-Ram ( 280GB ) ram-disk not faster than NVME EVO 980

[room101-dev]

I finally got around to answering this question to myself, what about just 'plotting' in RAM, rather than a high-speed NVME m.2 next to the cpu drive??

Well the results are in, nada, the read/write speeds are +5% on the start, but its actually slows down after a few minutes, I think this is contention between the memory controller, and the NVME fetch controller; where the NVME seems to balance out and help the 'plotter' move along more efficiently.

I was hoping to see a huge performance, I had heard from a few online source that people were seeing 5 hours, I essentially see the same as normal, e.g. 6 hours per plot, I run 4x plots per controller (NVME), I have three NVME on my plot machine 32 core AMD, 512MB RAM.

Lastly the only chia-net software I now run is 1.3 plotter "chia plots create ...", I don't run anything else, I don't need the block-chain, I hate the GUI, and all the third party 'chia tools' are crap and/or malware. I generate plots and mine them on HPOOL, and NO hpool doesn't require you to send them your private-keys that's the BIG LIE propagated by paid influencers on YT, and paid by chia-net in XCH.

Lastly, a few months ago I was burning up NVME's at 100 plots per drive, now I'm doing better, I keep fans on all the NVME drives, I modified the 'plotter' to throttle down if NVME goes over 70c, but now with individual fans for NVME's I never see over 50C, before the fans I would frequently see +90C on the controller, not the memory. Only the random access part was getting killed.

CHIA-DEV are worthless, they should have put in a TEMP switch on the 'chia plots create' -TEMP n, months ago, but they don't give a damn about anything other than their IPO.

...

Why only 4x plots per NVME, IMHO its the controller that dies, no matter if your drive is 512mb, or 8tb, its the same controller, it can't handler more than 4x, and of course I stagger by +30min; so that the plotter doesn't try to write all the plots to the HDD at the same time ( < 20min on average to write 108gb file); so 30min is lots of cushion. Another thing is there are two controllers on the NVME the random, and the sequential, they advertise 5,000MB/sec that' the sequential, the random is 500MB/sec, and its the random that gets destroyed by CHIA.