[Performance] startup time #11106

Open
abose opened this Issue May 13, 2015 · 3 comments

Projects

None yet

2 participants

@abose
Contributor
abose commented May 13, 2015

Comparing the startup time of brackets to other editors.

  • Determine the time it takes to launch brackets in the standard configuration(with extract extension).
  • Table out the comparative data
  • Analyze the code for figuring out the areas where the performance needs to be improved.

Method used to measure the performance:-

The The measurements was done using a stopwatch where launching the app was the start point and complete loading of all UI elements was the end point.

Parameters to consider for testing this.

  1. Test this for small(<500 files), medium(5K files approx) and large(>25K) projects.
  2. Run it on Windows and Mac systems with most likely configurations.
  3. compare warm launch and cold launch times.
@abose abose added this to the Release 1.4 milestone May 13, 2015
@abose abose added the Development label May 13, 2015
@abose abose self-assigned this May 14, 2015
@abose
Contributor
abose commented May 17, 2015

System Configuration

Windows iMac
Windows 8.1 Pro 64-bit. Processor: Intel core i7, 3.40 GHz. RAM: 16 GB OS X 10.10.3. Processor: Intel core i7, 3.50 GHz.RAM: 16 GB

Cold launch test
For this test, we switched on the system and when the processor/memory/disk utilization went zero after the boot-up, each of the apps were started. The measurements was done using a stopwatch where launching the app was the start point and complete loading of all UI elements was the end point. The testing configuration for all the apps where the default installer feature set for each of the latest versions of the apps as on 5th may 2015.

Windows
Brackets atom sublime vs code
18.35s 21.82s 6.25s 15.43s
mac
Brackets atom sublime vs code
25.98s 21.93s 4.11s 17.32s

Observations

  1. Brackets was particularly slow on mac.
  2. Sublime being a native app with a small memory footprint loaded quickly.
  3. VSCode was quicker to start up and being a CEF based app, there is scope for improvement in brackets with a similar architecture.
  4. The launch times doesn’t depend much on the project size which is open in brackets.

Warm launch test
Once each of the app was loaded at least once; a stop watch was used to measure the launch times averaged over 3 launches.

Windows
Brackets atom sublime vs code
2.3 3.4 0.5 2.01
mac
Brackets atom sublime vs code
2.13 2.01 0.12 1.25

Observations

  1. Brackets quits on close in mac while other editors don’t (you have to do app>quit). This results in other apps opening a bit faster in mac.
  2. The launch times doesn’t depend much on the project size which is open in brackets.
@nethip nethip removed this from the Release 1.4 milestone Jun 3, 2015
@nethip
Contributor
nethip commented Jun 3, 2015

Since we have not gotten started yet on this, let us move this to 1.5.
@ryanstewart: Should this be ok?

@nethip nethip added this to the Release 1.5 milestone Jun 3, 2015
@abose abose modified the milestone: Release 1.6, Release 1.5 Sep 11, 2015
@abose
Contributor
abose commented Sep 11, 2015

Moving to 1.6 as 1.5 didn't cover any startup acceleration work.

@abose abose referenced this issue Oct 10, 2015
Closed

RESEARCH: Measure performance against other editors #9270

0 of 12 tasks complete
@abose abose removed this from the Release 1.6 milestone Mar 2, 2016
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment