Tile is based (loosely) on MIPS, but it's not quite the same. The TilePro
only has a simple test-and-set instruction.
I have getting Open MPI compiling on the Tilera on my todo list for
another project, so I may be able to make progress on that front over the
next couple of weeks. We're not interested in performance just yet, so
I'm pretty sure it's not going to be the most performant implementation on
the planet when we first start. Charles, would you be interested in
working on some performance work once I get the atomics working? That
might be an "easier" place to start, as it's less build system gorp :).
On 8/2/11 7:48 AM, "George Bosilca" <bosilca_at_[hidden]> wrote:
>I'm curious if you did manage to get Open MPI compile on the TilePro64?
>Based on information gleaned from google, the Tile is supporting the MIPS
>atomic operations, so this problem is just a matter of convincing the
>Open MPI configure system to use the right asm file. You might want to
>start looking at opal/asm/asm-data.txt to figure out how this is done.
>On Jul 5, 2011, at 16:29 , Ralph Castain wrote:
>> I very much doubt we have Tile support as it hasn't come up before. If
>>you look in opal/asm/base, you'll see a MIPS.asm that contains the MIPS
>>code - perhaps you could use that as a starting point?
>> I didn't write any of that code, but I think if you poke around that
>>directory looking for "MIPS" you'll see what needs to be done.
>> On Jul 5, 2011, at 4:15 AM, Charles Bouillaguet wrote:
>>> Hi all,
>>> We have gotten our hands on a machine with a bunch of TilePro64
>>> processors (which seems to be a MIPS-like architecture). It runs linux,
>>> has gcc, and I was only dreaming to compile openmpi on it. However
>>> compilation fails with the errir "no atomic primitives available for
>>> I was wondering whether this problem has arisen already? If not, I
>>> be willing to try to address it and write the corresponding assembly
>>> code in opal/asm/..., but in this case I would really appreciate some
>>> Charles Bouillaguet
>>> users mailing list
>> users mailing list
>users mailing list
Brian W. Barrett
Dept. 1423: Scalable System Software
Sandia National Laboratories