waitforcopy seems to be on by default

[ariegrossman]

Running the new version in Ubuntu 20.04, and it now wait for copy to finish before start of next plot. It almost acts like I set the -w flag, but it's not set. So how to force it NOT to wait??? -w false? --not-wait. Please add a way to turn this off if you have a flag to turn it on. Thanks!

[ataa]

waitforcopy is false by default:

std::string pool_key_str;
	std::string farmer_key_str;
	std::string tmp_dir;
	std::string tmp_dir2;
	std::string final_dir;
	int num_plots = 1;
	int num_threads = 4;
	int num_buckets = 256;
	int num_buckets_3 = 0;
	bool waitforcopy = false;
	bool tmptoggle = false;

I am using the lastest and it does not wait for copy and start to craft new plot while doing the copy in the background.

[ariegrossman]

I see that it's set to false, but somehow, in Ubuntu 20.04, it's ignored and all my plots now wait for copy. So the bool maybe false, but the execution is contrary to this flag. Now my plots take twice as long.

[davemx]

I am seeing the same problem. I am not passing in the -w option yet the plotter is waiting for the copy process to finish. I’m also using Ubuntu 20.04 Server (Intel Rocket Lake, in case it matters).

Note: this is the latest pooling branch build.

[ariegrossman]

Noticed a slight change in performance. When plotting multi-plots, second plot does not wait for first once to finish copy, but after the second, every plot wait for the previous to finish. This is on ubuntu 20.04 (Ryzen 9 3900x). It's almost as if the -w option has been reset to true.

[ariegrossman]

The best feature of MadMax plotter, was that it did not wait for copy but started the next plot. Now My plots take twice as long because wait for copy was added back in. Please remove this "feature" wait for copy is doubling my plotting time and there is no way to disable it. Just take it out, and revert to the way it was designed. NO WAIT FOR COPY EVER!

[davemx]

I imagine "wait for copy" is a useful feature for some, and it is supposed to be off by default which would meet the previous expectation. So I'm not requesting the feature be removed -- just requesting some bug fixing or troubleshooting as to why it's not working for us. But, as a fellow "waiter", I appreciate your spirited urgency.

[ataa]

I tested it on 20.04 as well and it did not wait for a copy/move. My temp drive is 1TB. How much free space is left on your temp directory/disk while the final plot is still stored there waiting for the copy? Is there enough space to start a new one?

[davemx]

I've confirmed this is fixed in the latest release. It seems to only have affected the puzzles branch. I think we can close this.

[old-square-eyes]

I seem to have this issue using the Chia built in version of MadMax. I'm on Ubuntu, Chia Version 1.2.11. I'm executing without the -w flag form Tmux CLI.

Phase 4 took 120.148 sec, final plot size is 108835767491 bytes
Total plot creation time was 9413.08 sec (156.885 min)
Started copy to /media/computer/plot-k32-2021-12-12-23-04-f9410666e4ce61c2e8ddb5ff5244f7b2673344dfblahblahd263909b64689487.plot
Copy to /media/computer/plot-k32-2021-12-12-23-04-f9410666e4ce61c2e8ddb5ff5244f7b2673blahblah263909b64689487.plot finished, took 32223.6 sec, 3.22105 MB/s avg.
Crafting plot 2 out of -1 (2021/12/13 10:38:21)
Process ID: 17686
Number of Threads: 4
Number of Buckets P1:    2^8 (256)
Number of Buckets P3+P4: 2^8 (256)
Pool Puzzle Hash:  0ac88c539fa03f7c1c986dc1ablahblah347a77eaeblah724
Farmer Public Key: abe33d1481a7ec24a262e99a07119447cad037c5blahblahe8a4ee16e87c0b6098blahblah486efee20
Working Directory:   /media/computer/Chia/
Working Directory 2: /media/computer/Chia/
Plot Name: plot-k32-2021-12-13-10-38-8bb3d1badec065eda8d2eblahblah0aca679edf
[P1] Table 1 took 99.9622 sec
Progress update: 0.01