Hi Paul,

engine management is dominated by two players, Freescale and Infineon. 
Neither is using a Cortex based to device to control an engine, both 
have been using higher performance devices than a Cortex-M for many 
years. I would not know about Cortex-R in engine management, however 
afaik they are used in ABS systems.
Currently the dominating architecture for engine management is the 
Infineon TriCore, which has been providing DSP and Floating Point 
options for many years.
While Cortex-M4 is probably targeted at Automotive applications I doubt 
ARM envisions the Cortex-M in engine management. There is so much more 
to automotive than this, although I have to admit, I still consider 
engine management the heart of all automotive applications.
There will be even more computing performance required for all the car 
infotainment and there high end Cortex devices seem to do really well.

Cheers, Robert

Hi Robert,

Thanks much for the reply.

I know that Cortex-M3 is used for industrial control and motor 
applications, but not engine management.  The Cortex M4 seems to be a 
bit more power and can do floating point.  But the Cortex-R4 seems to be 
targeted at, among other things, engine management.

Do you know of any automotive controllers that use the R4, or is this 
still an emerging application?  Also, do you know what kind of functions 
that Freescale (I assume MPC5xx and MPC5xxx) and Infineon Tricore have 
that distinguish it as an engine management device?

thanks,
Paul

Mr. TPU guest is right. One major difference is the timer capability. 
TriCore as multiple timer arrays that can be combined in any thinkable 
way and then some more. IIRC there are approx. 100 timers on a TC1796.

One brief quote from the TC1796 data sheet "286 interrupt sources, 
generating up to 92 service requests" that is a little bit more than 
what a Cortex-M3 could or should handle. 416-pin package, 2 MB Flash 
with 256 KB SRAM and some more big chip items show that these devices, 
the TriCore and the MPC5xxx belong in a totally different league.

The timers are used to trigger injection on a hardware dependence rather 
than software triggered with jitter.

Or in a nutshell, the Cortex-M3 core does not cut it for this 
applications and existing peripherals are even much further below the 
required complexity level.

Robert

How about the Cortex-R4 however?

Paul Ware wrote:
> How about the Cortex-R4 however?

Chances are I just haven't heard about the R4 being used in Engine 
management but I can't fight the feeling that the Cortex-R family some 
off to a rough start with questionable future. Cortex-Ax and Cortex-Mx 
are doing great and having 2 out of 3 families being a huge success is a 
great achievements.
Again, the R-4 is not very visible to me, doesn't really mean there is 
no business with it. The only automotive implementation I've heard is a 
dual R4 design from TI, afaik for ABS.
The performance of the CPU would probably be adequate for engine 
management, the secret is in the peripherals though. The timer 
description for TriCore is more than 200 pages thick.

If you are with a company that does engine management, invite ARM to 
present their success stories, they would be happy to come. If it is 
just for academic interest, the lesson is there are specialized 
suppliers for this high end segment.

Robert