Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 29 Aug 2013 14:51:46 -0400
From: Yaniv Sapir <>
To: john-dev <>
Subject: Re: Parallella: Litecoin mining

Why not define buffers (pointers) in the number of cores in the workgroup,
each pointing to some base_address+corenum*sizeof(buf), where
corenum=(row*rows+col)? On the device side you can resolve the core's buf
address in the same way, using the e_group_config object members.

(or maybe I missed your point altogether?)

On Wed, Aug 28, 2013 at 6:54 PM, Rafael Waldo Delgado Doblas <> wrote:

> Hello,
> 2013/8/27 Yaniv Sapir <>
>> Rafael, if this question is still relevant, then yes, you can dynamically
>> allocate memory using standard malloc(). However, malloc() (and other
>> functions using its mechanism, like printf()) cannot be placed in internal
>> memory and have to be executed from external mem, which is a bad thing
>> w.r.t performance of critical code.
>> That said, I think that "to allocate dynamic memory on an specific
>> address" is kinda contradiction, since "specific" and "dynamic" do not work
>> together... If what you're looking for is a memory region at a specific
>> address, but who's life period is limited (i.e., the region can be reused
>> for other purposes), then use pointers set to those "specific addresses"
>> instead of static allocation of buffers.
> Thank you Yaniv, I just wonder if there was a way to do it, because would
> be nice in order to work with any core number. Each core need to write its
> hash in a know address based in its row and col. With the current approach,
> only a predefined number of cores can work, but if there was a way to write
> to a specific address, we would have a code that would work in any
> epiphany processor, independently of the core number.
> Regards,
> Rafael.

Yaniv Sapir
Adapteva Inc.
1666 Massachusetts Ave, Suite 14
Lexington, MA 02420
Phone: (781)-328-0513 (x104)
CONFIDENTIALITY NOTICE: This e-mail may contain information
that is confidential and proprietary to Adapteva, and Adapteva hereby
designates the information in this e-mail as confidential. The information
 intended only for the use of the individual or entity named above. If you
not the intended recipient, you are hereby notified that any disclosure,
distribution or use of any of the information contained in this
transmission is
strictly prohibited and that you should immediately destroy this e-mail and
contents and notify Adapteva.

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.