Blazingly fast FUD busters

Over the last 5 years of my handheld development experience, I have heard quite a lot of FUD about Qtopia and it's offspring, Opie. Today I heard a doozy, "We didn't use Qtopia because GPE is blazingly fast".

I have had enough of hearing that Gpe and X are so much more faster than Opie and Qtopia. I know from 5 years of experience in the linux handheld development world, that is just not true. Here's some numbers to prove it.

I will do a memory usage analysis in the coming weeks when I have the time.

Gpe vs. Opie vs. Qtopia

Distribution: Openzaurus 3.5.4
Machine: Sharp Zaurus 5000d

Boot up time (2nd boot after initial boot from flashing procedure):
Gpe- 42 s to logon screen; 53 to home screen (not counting time for log in)
Opie- 45.8 s with 5 second "wait..."
Qtopia (hardfloat)- 40.1 s
Qtopia (softfloat)- 38.2 s

Calendar (empty):
Gpe- 4.8 s
Opie- 1.4 s
Qtopia (hardfloat)- 3.3 s
Qtopia (softfloat)- 1.3 s

Contacts/Addressbook (empty):
Gpe- 17.3 s
Opie- 1.4 s
Qtopia (hardfloat)- 3.1 s
Qtopia (softfloat)- 1.1 s

Todo (empty):
Gpe- 13.6 s
Opie- 1.2 s
Qtopia (hardfloat)- 2.8 s
Qtopia (softfloat)- 0.9 s

Gpe- 4.3 s
Qtopia (hardfloat)- 4.0 s
Qtopia (softfloat)- 1.7 s

Gpe- 5.1 s
Opie- 0.9 s
Qtopia (hardfloat)- 3.3 s (includes scan for documents)
Qtopia (softfloat)- 0.8 s (includes scan for documents)

Gpe- 4.0 s
Opie- 1.0 s
Qtopia (hardfloat)- 3.2 s
Qtopia (softfloat)- 0.8 s

qwerty Input method popup:
Gpe- 3.0 s
Opie- 0.5 s
Qtopia (hardfloat)- 0.4 s
Qtopia (softfloat)- 0.3 s

Gpe- 1.5 s
Opie- 1.9 s
Qtopia (hardfloat)- 2.7 s
Qtopia (softfloat)- 1.6 s

System Info:
Gpe- 8.2 s
Opie- 2.1 s
Qtopia (hardfloat)- 4.6 s
Qtopia (softfloat)- 1.5 s

Image Viewer:
Gpe- 5.1 s
Qtopia (hardfloat)- 2.7 s
Qtopia (softfloat)- 2.6 s

Additional Notes:

Gpe and opie from official oz 3.5.4 release.

Qtopia (hardfloat): Version 2.1.2 trolltech rom image
Qtopia (softfloat): Version 2.2.0 oe-3.5.4 sdk was used to compile. -msoft-float was added

Gpe: No way to 'multitask' or select running applications (?)

Blog Topics: