5 ms·

The GIL removal was implemented, apart from the GIL removal part?

• BerislavLopac 4y ago
  No. Sam Gross’s PoC included a number of optimisations besides the gilectomy. It were those optimisations that made it faster, while the GIL removal slowed it down again; so only the optimisations were implemented.
  • DonHopkins 4y ago
    When performing gilectomization, it's better to undergilectomize than overgilectomize.
    • _joel 4y ago
      It's ok, you can get a reverse gilectomy
• remedan 4y ago
  Some of "Sam Gross’s proof of concept" was implemented of which the GIL removal was only a part. The rest was performance improvements.
• coldtea 4y ago
  Yes. The "GIL removal" umbrella proposal was two-fold. It included (a) removing the GIL, (b) several optimizations to handle some issues with GIL being removed and offset the GIL removal overhead (due to more frequent lock checks, etc). The GIL-removal assisting changes and optimizations were merged, but the GIL removal was not.