Beyond Malloc Efficiency to Fleet Efficiency - A Hugepage-Aware Memory Allocator

Beyond Malloc Efficiency to Fleet Efficiency - A Hugepage-Aware Memory Allocator

USENIX via YouTube Direct link

Intro

1 of 24

1 of 24

Intro

Class Central Classrooms beta

YouTube playlists curated by Class Central.

Classroom Contents

Beyond Malloc Efficiency to Fleet Efficiency - A Hugepage-Aware Memory Allocator

Automatically move to the next video in the Classroom when playback concludes

  1. 1 Intro
  2. 2 normally reducing malloc cycles is good
  3. 3 business logic changes, infrastructure does not
  4. 4 malloc cycles do not matter. $$$ spent on hardware matter.
  5. 5 most cpu cycles do nothing
  6. 6 hugepages cheapen the page table walk
  7. 7 hugepages make the TLB bigger!
  8. 8 space efficient hugepage aware allocators are hard
  9. 9 demand oscillates wildly
  10. 10 emptying density - binpacking
  11. 11 mistakes can live forever
  12. 12 tcmalloc structure
  13. 13 spans back everything
  14. 14 change nothing but the page heap
  15. 15 Temeraire: the design
  16. 16 slack and donation
  17. 17 how does the HugeFiller make decisions?
  18. 18 HugeFiller tracks metadata per hugepage
  19. 19 we favor fragmentation over fullness
  20. 20 results
  21. 21 staged rollout
  22. 22 saved -1.3% of cycles
  23. 23 saving memory in the process
  24. 24 virtuous cycles: hugepage coverage

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.