feat(ssa): Add flag to DIE pass to be able to keep store instructions

[aakoshh]

Description

Problem*

Resolves #7104

Summary*

Adds a flattened flag to Function::dead_instruction_elimination which is passed to Instruction::can_eliminate_if_unused to prevent this assumption from removing the store instruction before the inlining, the flattening of the CFG, and mem2reg passes have happened. Thanks @vezenovm for the pointer!

Additional Context

Originally I thought I would have to track &mut block parameters as "use", but the assumptions listed in can_eliminate_in_unused suggested that store must be removed from ACIR, and to stop that from happening it's enough to give it more information, so it doesn't infer whether other passes happened already based on the number of blocks in the function.

It looked like tracking &mut parameters as "use" would be a bigger change, since only instruction results are checked for use at the moment. Maybe there are other cases which would better be handled that way?

For the use case of #7072 I thought we'd just call the function with false and be done with it.

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]

Execution Report

Program	Execution Time	%
sha256_regression	0.051s	0%
regression_4709	0.001s	0%
ram_blowup_regression	0.598s	0%
rollup-root	0.104s	0%
rollup-merge	0.006s	0%
rollup-block-root	36.400s	0%
rollup-block-merge	0.107s	1%
rollup-base-public	1.224s	0%
rollup-base-private	0.451s	0%
private-kernel-tail	0.019s	0%
private-kernel-reset	0.310s	0%
private-kernel-inner	0.068s	0%

[github-actions]

Compilation Report

Program	Compilation Time	%
sha256_regression	1.060s	6%
regression_4709	0.883s	11%
ram_blowup_regression	16.600s	4%
rollup-root	3.650s	-4%
rollup-merge	2.026s	-8%
rollup-block-root-single-tx	143.000s	2%
rollup-block-root-empty	2.118s	-3%
rollup-block-root	136.000s	-3%
rollup-block-merge	3.572s	-3%
rollup-base-public	27.760s	-2%
rollup-base-private	10.280s	3%
private-kernel-tail	1.025s	4%
private-kernel-reset	5.934s	0%
private-kernel-inner	2.106s	0%

[TomAFrench]

LGTM

[github-actions]

Compilation Memory Report

Program	Peak Memory
keccak256	77.65M
workspace	123.87M
regression_4709	424.15M
ram_blowup_regression	1.46G
rollup-root	601.26M
rollup-merge	494.24M
rollup-block-root-single-tx	16.06G
rollup-block-root-empty	488.41M
rollup-block-root	16.07G
rollup-block-merge	601.26M
rollup-base-public	2.38G
rollup-base-private	1.14G
private-kernel-tail	207.43M
private-kernel-reset	584.49M
private-kernel-inner	294.68M

[github-actions]

Execution Memory Report

Program	Peak Memory
keccak256	74.71M
workspace	123.58M
regression_4709	316.02M
ram_blowup_regression	512.62M
rollup-root	498.31M
rollup-merge	473.05M
rollup-block-root	1.22G
rollup-block-merge	498.33M
rollup-base-public	734.18M
rollup-base-private	590.51M
private-kernel-tail	180.91M
private-kernel-reset	245.52M
private-kernel-inner	208.92M