Use parallel zstd by default

[curtis]

Using pzstd speeds up compression considerably on multi-core systems. And incurs only a minor size penalty.

I threw together this patch to try it out. ​https://github.com/dotysan/mc/commit/589227e13abb99fbe23667203eace3c37b4d12a9

It does pass make check but I haven't hand-tested all scenarios yet.

Anyone have strong thoughts?

Maybe it should be a separate menu item, in case there are really old systems without it? But it's been around since at least 2017.

Related to: #4571

[zaytsev]

Relates to #4571.

[zaytsev]

I don't care much either way, because I don't use it. I think the size difference is rather negligible, but I'm not sure whether the users would want all their cores utilized by default, or it's rather the opposite. Also, one could do a fallback to zstd if pzstd is not found with command -v - this is regarding the possible lack of pzstd on some systems.

Andrew, do you have any opinion on that?

[andrew_b]

Replying to zaytsev:

I don't care much either way, because I don't use it.

Me too.

I think the size difference is rather negligible, but I'm not sure whether the users would want all their cores utilized by default, or it's rather the opposite.

Agree.

Also, one could do a fallback to zstd if pzstd is not found with command -v - this is regarding the possible lack of pzstd on some systems.

Agree again:

$ pzstd
bash: pzstd: command not found

[and]

Parallel compression will reduce overall compression rate.

Parallel compression has advantages only on BIG files to reduce long-running compress time.

For mc: default should be single stream compress default command.

[zaytsev]

Alright, so let's leave it as is. Thanks for your opinions!