Hi sdrgpu, yes. Each section has performance metrics specific to each block. They varied depending on the amount of work done in each kernel so it wasn't possible to give blanket throughput.
---------- Forwarded message ----------
From: sdrgpu <notifications@github.com>
Date: Thu, Apr 27, 2017 at 4:10 PM
Subject: Re: [ghostop14/gr-lfast] Is gr-lfast now faster than gr-clenabled? (#1)
To: ghostop14/gr-lfast <gr-lfast@noreply.github.com>
Cc: ghostop14 <ghostop14@gmail.com>, Comment <comment@noreply.github.com>
From: sdrgpu <notifications@github.com>
Date: Thu, Apr 27, 2017 at 4:10 PM
Subject: Re: [ghostop14/gr-lfast] Is gr-lfast now faster than gr-clenabled? (#1)
To: ghostop14/gr-lfast <gr-lfast@noreply.github.com>
Cc: ghostop14 <ghostop14@gmail.com>, Comment <comment@noreply.github.com>
I see. Page 40 of your article "Study on Implementing OpenCL in Common GNURadio Blocks" showed 378 MSPS performance for a block size of 24576. I assume that did not include the Costas Loop but only the Quadrature Demod block for PSK.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
No comments:
Post a Comment