startup reduction times with syzygy

People that I’ve told about the work with syzygy that I’ve been doing have, almost universally, two reactions:

  1. That’s cool!  (Thanks; I did very little work for it!)
  2. How does that translate into startup time?

Assuming that a 40% reduction in page faults leads to a 40% reduction in startup time is not reasonable, but surely there should be some reduction in startup time, right?  I finally benchmarked this today using the about:startup extension; these numbers are from cold start, freshly rebooted both times:

Version main sessionRestored firstPaint
Trunk build 125 1733 1671
Optimized build 125 1639 1577

So a 40% reduction in page faults translates into ~6% reduction in startup time; not stellar, but not too bad either.

The next step is making sure this all works with PGO builds on Windows. Then we get to have a discussion about whether to incorporate this into the regular builds and getting all the infrastructure on build machines.

7 comments

  1. I guess it depends on how slow of a computer you have. I should imagine it will make a bigger impact on a computer that always has its hard drive thrashing for reason what so ever, such as my work laptop.

    Anyhow great work.

  2. The win varies with disk speed, is this on SSD?

    • Yes, this is on an SSD. Might be more significant on a mechanical hard drive.

  3. First of all- great job :)

    Second- were this measurements made on the Win7 SSD laptop you mentioned in the previous post? if so, have you thought about measuring it on a non-ssd machine and seeing if that makes a bigger difference to start-up time?

    • Nathan Froyd

      Yeah, this was on that laptop. I will experiment with my desktop machine (old, creaky 80GB drive) and see if that makes a difference.

  4. [...] my previous post, I presented some startup timings with syzygy-optimized Firefox binaries.  People asked whether [...]

  5. [...] numbers are really encouraging; they’re actually even better than the initial numbers I posted earlier.  (Though I note that they are universally slower than the earlier [...]