redo both Elements and Bitcoin benchmarks

[apoelstra]

I didn't actually want to do the Elements benchmarks, but I guess I failed to filter the benchmarks, so here they are..

[apoelstra]

On ab4e17e successfully ran local tests

[roconnor-blockstream]

Unfortunately these benchmarks are bad.

Firstly BuildTapbranch and TapdataInit indicating that there is still something wrong with how the Elements jet benchmarks are handled.

But the bigger issue is that there seems to be significant overhead in low cost jets.

Our calibration jet: CheckSigVerify is reasonably consistent: 55555 in 2024-07, 55137 in 2024-09, 54442 in 2025-10, 54916 and 54758 in this PR.

But the low cost jets have added significant overhead:

e.g. Eq1 was 41 in 2024-07, 43 in 2024-09, 36 in 2025-10, but now is 145 and 137 in this PR

Verify was 32 in 2024-07, 21 in 2024-09, 23 in 2025-10, but is now 96 and 136 in this PR.

[apoelstra]

Yep, you're right. I ran these with --debug on cargo criterion.

Will try again.