Does anybody use PC as hard real time controller?
Do you have any experience with eliminating SMI (system management
interrupts)? I have found out that power management, ECC checking, ACPI
(APM), thermal monitoring (for INTEL processor) will cause SMI assertion.
Anybody now other causes of SMI?
Many thanks!

Hi Jing,

Jing Zhang wrote:
Does anybody use PC as hard real time controller?

Do you have any experience with eliminating SMI (system management
interrupts)? I have found out that power management, ECC checking, ACPI
(APM), thermal monitoring (for INTEL processor) will cause SMI assertion.
Anybody now other causes of SMI?
Any suggestion at this point will likely be futile, as this strongly
depends on motherboard, chipset etc. how hard is your "hard real-time".
If it's hard RT and safety critical I would recommend looking at other
HW. If it's mission critical you might want to go down this path, but
it's surely a bit of a bet or at least an aweful amount of work to make
sure.

Regards,
Stefan.

Tue, 15 Feb 2005 19:36:12 -0700, "Jing Zhang" <zhjing@unm.edu>
wrote:

>
>Does anybody use PC as hard real time controller?
>
>Do you have any experience with eliminating SMI (system management
>interrupts)? I have found out that power management, ECC checking, ACPI
>(APM), thermal monitoring (for INTEL processor) will cause SMI assertion.

Fortunately most of these are easily eliminated e.g. by turning power
management off etc.

>Anybody now other causes of SMI?

If you are looking for timing challenges, look at Transmeta (code
translation) or Geode, in which missing hardware features are emulated
in SMM, such as text mode VGA :-).

Paul

"Paul Keinanen" <keinanen@sci.fiwrote in message
@4ax.com
Tue, 15 Feb 2005 19:36:12 -0700, "Jing Zhang" <zhjing@unm.edu>
wrote:
>
>
>Does anybody use PC as hard real time controller?
>
>Do you have any experience with eliminating SMI (system management
>interrupts)? I have found out that power management, ECC checking, ACPI
>(APM), thermal monitoring (for INTEL processor) will cause SMI assertion.
>
Fortunately most of these are easily eliminated e.g. by turning power
management off etc.
>
>Anybody now other causes of SMI?
>
If you are looking for timing challenges, look at Transmeta (code
translation) or Geode, in which missing hardware features are emulated
in SMM, such as text mode VGA :-).

Paul

and watch out for legacy mode usb!!! btdt, ended up watching an osciliscope
to catch the problem.
Pat

Does anybody ever written an SMI handler for the BIS?

Any experience with SMI handling in hyper threading enabled CPU?

"Jing Zhang" <zhjing@unm.eduwrote in message
news:cuubd8$t0d$1@nunki.unm.edu

Does anybody use PC as hard real time controller?

Do you have any experience with eliminating SMI (system management
interrupts)? I have found out that power management, ECC checking, ACPI
(APM), thermal monitoring (for INTEL processor) will cause SMI assertion.
Anybody now other causes of SMI?

Many thanks!
--

"Paul Keinanen" <keinanen@sci.fiwrote in message
@4ax.com
Tue, 15 Feb 2005 19:36:12 -0700, "Jing Zhang" <zhjing@unm.edu>
wrote:
>
>
>Does anybody use PC as hard real time controller?
>
>Do you have any experience with eliminating SMI (system management
>interrupts)? I have found out that power management, ECC checking, ACPI
>(APM), thermal monitoring (for INTEL processor) will cause SMI assertion.
>
Fortunately most of these are easily eliminated e.g. by turning power
management off etc.
>
>Anybody now other causes of SMI?
>
If you are looking for timing challenges, look at Transmeta (code
translation) or Geode, in which missing hardware features are emulated
in SMM, such as text mode VGA :-).

Is that AMD GeodeT SC1100 Processor and AMD GeodeT DBSC1100 Development
Board you were talking about? I take a look at AMD website, I am not sure it
offers hard real time.

Is Transmeta a supplimentary circuitry in the motherboard that could
generate SMI?

Thanks.

Jing

Paul