|
Message
From: christian@m...
Date: Tue, 21 Oct 2003 19:14:13 +0200
Subject: Re: [oc] ARM 922T, Cyclone FPGA
So, in brief, the ARM922T would be a good choice for multi-threading. I
prefer to use Linux OS, like eCos RedHat. The memory management will
be perform by the MMU and the OS services. I think that using many
Nios processor for each thread would be hard to manage the memory
access by each thread.
With Excalibur, I could do different prototypes (using with or not 922T).
For a small project, i could select only APEX 20KE and use Nios os with
the Avalon Bus. I could install the uCLinux.
For a huge project with multi-threading, I could enable 922T and install
eCos for example.
But, I read that openrisc 1200 would be a good choice, but I don`t
know if it`s easy to install.
Regards
----- Original Message -----
From: "Jim Dempsey" <tapedisk@a... >
To: <cores@o... >
Date: Tue, 21 Oct 2003 11:24:09 -0500
Subject: Re: [oc] ARM 922T, Cyclone FPGA
>
>
> Christian,
>
> Multi-threading is generaly a function of the operating system. You
> can also
> perform multi-threading within an application by way of a custom
> written
> interrupt service routine or by use of co-routintnes. This is not
> hard to
> do. With the selection of ARM922T in Excalibur you generaly will
> end up
> using one processor. With the selection of NIOS in Excalibur or in
> Cyclone
> you have the choice of either multi-threading a single processor
> (via OS or
> application) or by using multiple processors (e.g. one per thread).
> Depending on the size of the core you can get a bunch of NIOS
> processors
> working on the problem. You can do a similar thing with and
> OpenCores
> processor. However, if you elect to go the multiple processor route
> my
> inexperienced guess would be that the Avalon bus would be a better
> choice
> than the Wishbone bus.
>
> Although ARM922T is faster than NIOS running a benchmark you have
> the
> flexibility of using the FPGA to add hardware accellerators. You
> may find
> the following article of interest.
>
> http://www.cl.cam.ac.uk/~rnc1/descrack/cracker.shtml
>
> You can thank John Dalton for pointing this out about two years ago
|
 |
Browse
