chore: debug mem2reg for load in loop

[asterite]

Description

Problem

Related to #6971

Summary

This just adds an unfinished test with a small SSA snippet that I believe could be optimized out.

Next I'll check the mem2reg code to see if I can understand why it's not optimizing it.

Additional Context

Documentation*

Check one:

• No documentation needed.
• Documentation included in this PR.
• [For Experimental Features] Documentation to be submitted in a separate PR.

PR Checklist*

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[github-actions]

Compilation Report

Program	Compilation Time	%
sha256_regression	1.336s	-2%
regression_4709	0.777s	-2%
ram_blowup_regression	15.040s	-1%
rollup-root	3.570s	-15%
rollup-block-merge	4.310s	4%
rollup-base-public	29.600s	-2%
rollup-base-private	12.200s	-8%
private-kernel-tail	1.034s	0%
private-kernel-reset	6.800s	-1%
private-kernel-inner	2.192s	-1%

[github-actions]

Execution Report

Program	Execution Time	%
sha256_regression	0.101s	1%
regression_4709	0.001s	0%
ram_blowup_regression	0.581s	0%
rollup-root	0.105s	0%
rollup-block-merge	0.106s	0%
rollup-base-public	1.430s	-1%
rollup-base-private	0.648s	1%
private-kernel-tail	0.023s	0%
private-kernel-reset	0.389s	0%
private-kernel-inner	0.117s	0%

[github-actions]

Compilation Memory Report

Program	Peak Memory
keccak256	78.50M
workspace	123.72M
regression_4709	422.91M
ram_blowup_regression	1.58G
rollup-base-public	2.55G
rollup-base-private	1.25G
private-kernel-tail	201.97M
private-kernel-reset	716.57M
private-kernel-inner	292.25M

[github-actions]

Execution Memory Report

Program	Peak Memory
keccak256	74.61M
workspace	123.57M
regression_4709	315.93M
ram_blowup_regression	512.47M
rollup-base-public	773.33M
rollup-base-private	424.88M
private-kernel-tail	182.01M
private-kernel-reset	255.58M
private-kernel-inner	215.19M

[jfecher]

@asterite you may be interested in some of my notes from the last time I looked at this problem: https://hackmd.io/@JakeF/S1FA9rTnh#Better-Loop-Optimizations. I thought I had a solution originally but it ended up being broken (only somewhat implicitly stated by the problems at the bottom being unfinished)

[asterite]

@jfecher Interesting, thanks! I'll take a look. It seems I'm just finding this out when this was already known 😊 . That said, the first snippet:

unconstrained fn main() {
    let mut foo = 7;
    for i in 0 .. 10 {}
    assert_eq(foo, 7);
}

seems to be optimized out now.

[jfecher]

@asterite we sometimes do loop unrolling in unconstrained code now, so you can "fix" that example to get it to not optimize by preventing loop unrolling. I found adding a println does the trick:

unconstrained fn main() {
    let mut foo = 7;
    for i in 0 .. 10 {
        println(i);
    }
    assert_eq(foo, 7);
}

Now in b2 before the constrain you should still see a load v1 that should be able to be optimized out, but isn't.

[asterite]

Thanks! I'll close this PR as the solution seems to be a lot harder than what I thought.