[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAFYn=yDr8UOyHLNo1rWR6qysRp-4zHshrwPUwh09ENUOabCJaw@mail.gmail.com>
Date: Mon, 16 Sep 2013 14:20:00 -0400
From: Yaniv Sapir <yaniv@...pteva.com>
To: john-dev <john-dev@...ts.openwall.com>
Subject: Re: bcrypt-parallella on 64-core (was: Katja's weekly
report #13)
BTW, the restriction is not just with fetch, but with load in general. The
only safe method for a core on this row to read data from ERAM is using
eDMA. Writing to ERAM, or writing data from host (or external space in
general) onto these cores is no problem.
On Mon, Sep 16, 2013 at 1:38 PM, Yaniv Sapir <yaniv@...pteva.com> wrote:
> Katja,
>
> As for the empty buffers - the E64 chip has a design bug - cores at the
> 3rd row cannot fetch instructions from external memory. Due to this, the
> kernel programs loaded on these cores are not run, because portions of the
> code reside in ERAM (b/c we use prtintf() calls). I believe that in your
> program, you use the internal.ldf profile, so this is not a problem.
>
> As for the other problem, I will have a look at your code later to see
> why. can you post the path to the program on the Z64 board?
>
> Yaniv.
>
>
Content of type "text/html" skipped
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
