git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: cloning the kernel - why long time in "Resolving 313037 deltas"

From
MRMarco Roeland <marco.roeland@xs4all.nl>
Date
Dec 22, 2006, 08:04 UTC
Message-ID
<20061222080450.GB9595@fiberbit.xs4all.nl>
In-Reply-To
<991F50E2-2C98-4710-84C5-3D7348E00B64@mac.com>
On Thursday December 21st 2006 at 20:44 Kyle Moffett wrote:
Show 6 quoted lines
> I've actually also seen filesystem operation latency double or triple  
> if you start trying to do operations from multiple threads at once.   
> Suddenly the already dog-slow single-CPU operations start bouncing  
> caches and the Mac OS X mostly-whole-of-BSD-BKL across CPUs and it  
> just crawls.  I can definitely see the local disk IO taking 100x  
> longer than the network I/O, especially with an 8-megabit internet link.

The mmap() implementation on Mac OS X ate 85% percent of system time on the old version for git-index-pack. Of that time nearly all was spent on some Mach locking function. So yes the BKL like locking inside the Mach message passing seems to be the big culprit, perhaps Andy Tanenbaum can explain this on LCA 2007. <ducks>

As most developers now run multicore CPU's we notice these differences even better now! ;-)

-- 
Marco Roeland
Previous: Shawn PearceNext: Andreas Ericsson
Message 12 of 51 in “Re: [PATCH] fetch-pack: avoid fixing thin packs when unnecessary”
  1. Johannes SchindelinDec 18, 2006
  2. Nicolas PitreDec 18, 2006
  3. Randal L. SchwartzDec 18, 2006
  4. Nicolas PitreDec 18, 2006
  5. Randal L. SchwartzDec 18, 2006
  6. Nicolas PitreDec 18, 2006
  7. Linus TorvaldsDec 18, 2006
  8. Randal L. SchwartzDec 18, 2006
  9. Martin LanghoffDec 18, 2006
  10. Kyle MoffettDec 22, 2006
  11. Shawn PearceDec 22, 2006
  12. Marco RoelandDec 22, 2006
  13. Andreas EricssonJan 3, 2007
  14. Linus TorvaldsDec 18, 2006
  15. Nicolas PitreDec 19, 2006
  16. Theodore TsoDec 19, 2006
  17. Shawn PearceDec 19, 2006
  18. Linus TorvaldsDec 19, 2006
  19. Shawn PearceDec 19, 2006
  20. Marco RoelandDec 19, 2006
  21. Shawn PearceDec 19, 2006
  22. Shawn PearceDec 19, 2006
  23. Marco RoelandDec 19, 2006
  24. Shawn PearceDec 19, 2006
  25. Marco RoelandDec 19, 2006
  26. Alex RiesenDec 19, 2006
  27. Juergen RuehleDec 21, 2006
  28. Theodore TsoDec 19, 2006
  29. Linus TorvaldsDec 19, 2006
  30. Shawn PearceDec 20, 2006
  31. Shawn PearceDec 20, 2006
  32. Linus TorvaldsDec 19, 2006
  33. Johannes SchindelinDec 19, 2006
  34. Junio C HamanoDec 19, 2006
  35. Jeff KingDec 19, 2006
  36. Andy WhitcroftDec 19, 2006
  37. index-pack usage of mmap() is unacceptably slower on many OSes other than LinuxNicolas Pitre, Dec 19, 2006
  38. Junio C HamanoDec 19, 2006
  39. Nicolas PitreDec 19, 2006
  40. Linus TorvaldsDec 19, 2006
  41. Randal L. SchwartzDec 19, 2006
  42. Randal L. SchwartzDec 19, 2006
  43. Jeff GarzikDec 19, 2006
  44. Junio C HamanoDec 20, 2006
  45. Linus TorvaldsDec 20, 2006
  46. Jeff GarzikDec 20, 2006
  47. Junio C HamanoDec 20, 2006
  48. Junio C HamanoDec 20, 2006
  49. Linus TorvaldsDec 20, 2006
  50. Junio C HamanoDec 20, 2006
  51. Nikolai WeibullDec 20, 2006

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.