But I run the provided demo from ..\RT-STM32F100-DISCOVERY (3.0) and ..\ARMCM3-STM32F100-DISCOVERY (2.6.8)
Also tested ARMCM4-STM32F407-DISCOVERY too.
I modify nothing in it.
One thing is that I checked chconf.h file and select in 3.0 classic tick mode with the same frequency as in 2.6.8 (1000Hz).
Anybody may repeat my tests.
ChibiOS 3.0 benchmark
- Giovanni
- Site Admin
- Posts: 14444
- Joined: Wed May 27, 2009 8:48 am
- Location: Salerno, Italy
- Has thanked: 1074 times
- Been thanked: 921 times
- Contact:
-
- Posts: 276
- Joined: Mon Sep 24, 2012 3:52 pm
- Location: Donetsk
- Been thanked: 32 times
- Contact:
Re: ChibiOS 3.0 benchmark
Hi!
I tested, tested, tested, even downloaded and ran ChibiStudio.
That's what conclusion: optimization level is crucial.
ChibiOS 3.0 code without high level of optimization runs much slower then ChibiOS 2.6.x
Thats my mistake: I thought that compare results independent from LEVEL of optimization, i.e. if code runs faster in high level optimization it must runs faster in low level of optimization. Thats wrong.
In end of ends realy Chibi 3.0 bit faster than 2.6 but only with high level of optimization of the compiler. It is true for IAR and KEIL.
I tested, tested, tested, even downloaded and ran ChibiStudio.
That's what conclusion: optimization level is crucial.
ChibiOS 3.0 code without high level of optimization runs much slower then ChibiOS 2.6.x
Thats my mistake: I thought that compare results independent from LEVEL of optimization, i.e. if code runs faster in high level optimization it must runs faster in low level of optimization. Thats wrong.
In end of ends realy Chibi 3.0 bit faster than 2.6 but only with high level of optimization of the compiler. It is true for IAR and KEIL.
Who is online
Users browsing this forum: No registered users and 15 guests