core: speed up header import

[holiman]

This is a follow up to #21471.

Charts and niceties can be found here: https://github.com/holiman/headerimport/blob/main/README.md

This PR implements the following modifications

• 1M blocks without header-shortcut (mod1)
  • Do not shortcut check if block is present, thus avoid disk lookup
• 1M blocks with less hashing (mod2)
  • Don't check hash ancestry in early-check (it's still done in parallel checker)
• 1M blocks with less syncronization (mod3)
  • Don't check for abort between every header EDIT: this optimization was reverted
• 1M blocks without Now
  • Don't check time.Now for every single header

TLDR: With these changes, writing 1M headers goes down to 62s for 1M headers, from 106s on (old) master, and 80s on PR 21471.

[ucwong]

Why remove this check ?

[holiman]

See https://github.com/holiman/headerimport/blob/main/README.md#1m-blocks-on-21471

[holiman]

Rebased on master. Master:

This PR:

TLDR: 52+22 = 74s vs 38+23=61s

[karalabe]

unix :)

[karalabe]

Pls fix utx->unix, othewise SGTM