There is no implementation of xC for the ARM architecture. The idea of integrating an ARM core with the xCORE is to be able to reuse existing ARM code. xC code can currently only be run on other platforms using software simulation of the xCORE architecture.

<r>This looks like there's a problem with the compiler. If you could submit a bug report, with the details of which version of the tools you are using, at <URL url="https://www.xmos.com/secure/tickets">https://www.xmos.com/secure/tickets</URL> we will have a look into it.<br/> <br/> Cheers,<br/> <br...

Hi ers35, Back to your original post, the deadlock is due to the client's use of the 'pause' control token and I think xsim isn't freeing the channel end properly when the pause is received. Running the same program on AXE (https://github.com/rlsosborne/tool_axe) it looks to be okay. Incidentally, c...

Another XMOS-related story in The Economist:

How the West was won: Inmos, Picochip, Icera and their founders: the evolution of a tech cluster.

http://www.economist.com/node/21525434

On XS1-G the code produced by the mapper accesses a sswitch register to synchronise startup between the cores. Programs built for multinode XS1-L systems appear to work.

Oh okay, how does synchronisation work for multi-node XS1-L systems then?

This looks really interesting - I've been using xsim heavily for development and testing so it would be great to be able to simulate execution faster. In particular, performance is hit dramatically with multi-node simulations. From looking over AXE it doesn't seem that this is supported yet but I'm ...

Well there are some significant differences between XC and occam-pi but I'm surprised that you claim occam-pi is unsuited for a platform very similar to the transputer.

Why??? Occam-pi is a concurrent message passing language and the XMOS architecture is perfect match for this! Is it doomed to run only on single cores now?

Oops - looks like they've published a draft!

Pretty interesting article in the Economist yesterday about parallelism:

"Computing: Parallel programming, once an obscure niche, is the focus of increasing interest as “multicore” chips proliferate in ordinary PCs "

http://www.economist.com/node/18750706