perf(levm): add more benchmarks to the suite

[dsocolobsky]

Motivation

We need more complex contracts to benchmark our implementation against other VMs.

Description

• Added 5 new benchmarks: FactorialRecursive, ManyHashes, ERC20Mint, ERC20Approval and ERC20Transfer.
• Benchmarks contracts are now compiled from Solidity code in our codebase, instead of being hardcoded bytecode.
• Simplified crates/vm/levm/Makefile since it had a lot of repetition and hardcoded values, mainly avoid copy-paste mistakes and make it easier to modify.
• Benchmarks are now run from a single crates/vm/levm/bench/revm_comparison/src/benchmark.rs file with parameters, instead of two files per benchmark case.
• In crates/vm/levm/bench/revm_comparison/lib.rs:run_with_levmwe now use VM::transact() instead of directly VM::execute(callframe). This might modify slightly our benchmarks because we now take into account prepare_execution, however this fixes the benchmarks not working with contracts that call themselves recursively (like FibonacciRecursive)
• I lowered the repetitions for each benchmark from 100k to 10k, from what I've seen the results are the same and it runs much faster, we can change later if we need more precise measurements.

Further Steps

• We can think of new benchmarks.
• Compare if the results from LEVM and REVM are the same and fail if not.
• Analyze the flamegraphs to see possible bottlenecks and add to Notion the findings.

Notes
Right now ci_bench_levm.yaml for the job benchmark-main is unchanged, because else it wouldn't pass the CI since the new benchmarks aren't in main. I think the best is to merge like this and then change the CI again in main in another small PR?

Closes #1623

[github-actions]

| File                                                                                  | Lines | Diff |
+---------------------------------------------------------------------------------------+-------+------+
| /home/runner/work/ethrex/ethrex/crates/vm/levm/bench/revm_comparison/src/benchmark.rs | 34    | +34  |
+---------------------------------------------------------------------------------------+-------+------+
| /home/runner/work/ethrex/ethrex/crates/vm/levm/bench/revm_comparison/src/compile.rs   | 63    | +63  |
+---------------------------------------------------------------------------------------+-------+------+
| /home/runner/work/ethrex/ethrex/crates/vm/levm/bench/revm_comparison/src/lib.rs       | 85    | +27  |
+---------------------------------------------------------------------------------------+-------+------+

Total lines added: +124
Total lines removed: 0
Total lines changed: 124

[github-actions]

Summary: 6475/6475 (100.00%)

Prague: 2373/2373 (100.00%)
Cancun: 3579/3579 (100.00%)
Shanghai: 221/221 (100.00%)
Paris: 62/62 (100.00%)
London: 39/39 (100.00%)
Berlin: 35/35 (100.00%)
Istanbul: 34/34 (100.00%)
Constantinople: 66/66 (100.00%)
Byzantium: 33/33 (100.00%)
Homestead: 17/17 (100.00%)
Frontier: 16/16 (100.00%)

[github-actions]

Benchmark Results Comparison

PR Results

Benchmark Results: Factorial

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
revm_Factorial	237.3 ± 2.2	234.2	241.4	1.00
levm_Factorial	841.0 ± 3.0	835.4	844.2	3.54 ± 0.04

Benchmark Results: Factorial - Recursive

Command	Mean [s]	Min [s]	Max [s]	Relative
revm_FactorialRecursive	1.452 ± 0.103	1.318	1.593	1.00
levm_FactorialRecursive	15.372 ± 0.126	15.256	15.523	10.58 ± 0.75

Benchmark Results: Fibonacci

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
revm_Fibonacci	211.5 ± 1.3	209.0	213.1	1.00
levm_Fibonacci	837.1 ± 3.1	831.5	841.5	3.96 ± 0.03

Benchmark Results: ManyHashes

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
revm_ManyHashes	8.7 ± 0.1	8.5	8.8	1.00
levm_ManyHashes	17.5 ± 0.2	17.3	17.8	2.02 ± 0.03

Benchmark Results: ERC20 - Transfer

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
revm_ERC20Transfer	249.0 ± 2.5	246.9	255.7	1.00
levm_ERC20Transfer	3239.4 ± 173.1	3165.5	3728.1	13.01 ± 0.71

Benchmark Results: ERC20 - Mint

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
revm_ERC20Mint	141.0 ± 1.2	140.0	144.2	1.00
levm_ERC20Mint	1676.4 ± 13.1	1660.5	1694.7	11.89 ± 0.14

Benchmark Results: ERC20 - Approval

Command	Mean [s]	Min [s]	Max [s]	Relative
revm_ERC20Approval	1.037 ± 0.006	1.026	1.046	1.00
levm_ERC20Approval	11.332 ± 0.095	11.188	11.432	10.93 ± 0.11

Main Results

Benchmark Results: Factorial

Command	Mean [s]	Min [s]	Max [s]	Relative
revm_factorial	7.373 ± 0.013	7.342	7.382	1.00
levm_factorial	26.197 ± 0.500	25.741	27.564	3.55 ± 0.07

Benchmark Results: Fibonacci

Command	Mean [s]	Min [s]	Max [s]	Relative
revm_fibonacci	8.212 ± 1.085	7.484	10.280	1.00
levm_fibonacci	24.232 ± 1.955	23.570	29.795	2.95 ± 0.46

[fborello-lambda]

Amazing.

The PR's note says:

I think the best is to merge like this and then change the CI again in main in another small PR?

Yes, i would do a small PR after this one is merged.