Timing benchmark tweaks and fixes

[cameel]

This is a bunch of independent tweaks to the test/benchmarks/ runner script to make using it for evaluating optimizer sequences a little less tedious. Each could really be a separate small PR.

• The script would not continue when legacy compilation failed with "Stack too deep".
• Bytecode size is a useful additional data point that we can easily get here.
• Putting the values manually in a table gets old quickly. Now the script prints a table I can just copy paste into a comment on Github.
• The lighter tests now run first so you can start seeing results immediately. Also, now the script does not wait with printing legacy results until IR finishes.
• solc binary is now a parameter so that I can easily use system-wide solc. For testing sequences I find it easier to use the release binary and just give it different --yul-optimizations.

[cameel]

Before

=======================================================
            chains.sol
-------------------------------------------------------
legacy pipeline took 0.96 seconds to execute.
via-ir pipeline took 69.23 seconds to execute.
=======================================================
=======================================================
            OptimizorClub.sol
-------------------------------------------------------
legacy pipeline took Command exited with non-zero status 1
2.78 seconds to execute.
via-ir pipeline took 16.33 seconds to execute.
=======================================================
=======================================================
            verifier.sol
-------------------------------------------------------
legacy pipeline took 0.74 seconds to execute.
via-ir pipeline took 3.21 seconds to execute.
=======================================================

=======================================================
Warnings and errors generated during run:
=======================================================

After

| File                 | Pipeline | Bytecode size | Time     | Exit code |
|----------------------|----------|--------------:|---------:|----------:|
| `verifier.sol`       | legacy   |    4983 bytes |   0.79 s |         0 |
| `verifier.sol`       | via-ir   |    4408 bytes |   3.51 s |         0 |
| `OptimizorClub.sol`  | legacy   |       0 bytes |   2.97 s |         1 |
| `OptimizorClub.sol`  | via-ir   |   22427 bytes |  16.99 s |         0 |
| `chains.sol`         | legacy   |    5911 bytes |   0.88 s |         0 |
| `chains.sol`         | via-ir   |   21443 bytes |  66.46 s |         0 |

=======================================================
Warnings and errors generated during run:
=======================================================

Formatted

File	Pipeline	Bytecode size	Time	Exit code
verifier.sol	legacy	4983 bytes	0.79 s	0
verifier.sol	via-ir	4408 bytes	3.51 s	0
OptimizorClub.sol	legacy	0 bytes	2.97 s	1
OptimizorClub.sol	via-ir	22427 bytes	16.99 s	0
chains.sol	legacy	5911 bytes	0.88 s	0
chains.sol	via-ir	21443 bytes	66.46 s	0

[r0qs]

Looks great! And much better than the previous output of the script :)
I just have a small nitpick to mention.