Implement get_node with a get_node_raw

[Desour]

• Goal: Improve performance of get_node.
• How: get_node is implemented in builtin with a get_node_raw(x, y, z) -> content, param1, param2, pos_ok.
• Why is this beneficial: In master, get_node calls lua functions to push nodes, and to read vectors. This is much faster if done in lua.

Flamegraphs:
master:

PR:

Rough benchmark results:
master: 450-850 ms
PR: around 270-300 ms (it's much more persistent, idk why)

Benchmark function adopted from #14225 by @sfence. (Added you as co-author. :))

To do

This PR is a Ready for Review.

How to test

• /bench_bulk_get_node
• TODO: Do we have unit tests for get_node and get_node_or_nil?

Here's some instructions I've noted down at some point to disable cpu scaling, for better testing:

info:
	https://www.kernel.org/doc/html/v5.11/admin-guide/pm/cpufreq.html
	https://wiki.archlinux.org/index.php/CPU_frequency_scaling


check:
	sudo cpupower frequency-info
	watch grep \"cpu MHz\" /proc/cpuinfo


cpu-scaling off:
	sudo sh -c "echo 0 > /sys/devices/system/cpu/cpufreq/boost"
	sudo cpupower frequency-set -g userspace
	sudo cpupower frequency-set -f 1600MHz

cpu-scaling on:
	sudo cpupower frequency-set -g schedutil
	sudo sh -c "echo 1 > /sys/devices/system/cpu/cpufreq/boost"

[Desour]

Added the comments.

[sfan5]

LGTM otherwise

[SmallJoker]

	Time / ms	PR / ms
	314.16	79.64
	290.42	81.25
	254.11	80.21
	193.72	80.7
	184.71	80.32
	329.17	80.05
	251.46	80.42
	301.68	79.2
	197.04	79.23
	197.5	80.82
average	251.4	80.2
MEDIAN	252.8	80.3

I find it strange that there's this much difference. pushnode already uses Lua functions to construct the table from arguments (core.set_push_node). So apparently by moving the table constructor after "returning" from C++ makes such big difference? Why's that?

[Desour]

In master we call lua functions from C++ to read vector and push node values. The C->Lua call overhead is probably quite big. And also we always need to look up the functions to call first, which requires a bunch of Lua C api calls.

[SmallJoker]

Works well. Cannot complain.

[Desour]

Benchmark function adopted from #14225 by @sfence. (Added you as co-author. :))

Somehow that co-authorship didn't make it into the commit. Did github squash it away?

[SmallJoker]

Sorry @Desour (or rather @sfence ). I totally missed that remark and removed it alongside the other commit titles upon merge. GitHub is not to blame.