On 05/01/2020 10:49 PM, John Deere wrote:
I've also tried the standalone patch by Ben and it seems to have reduced the latencies on top of AQL by another 5 ms.
Hello,
Did you notice any throughput changes or system load changes in the test that you did with my patch?
Hi Ben,
I've ran some basic throughput tests with iperf and it seems that there was also slight hit in throughput, but only by 5-20 mbps or so.
If you are using smaller buffers anyway, maybe the 1/4 restart limit is too low. I'm running closer to 2000 tx buffers,
so my restart level would be at around 500 buffers. I did not see any drop in tput in my test case, but rather saw
improvement.
Thanks,
Ben
--
Ben Greear <greearb@xxxxxxxxxxxxxxx>
Candela Technologies Inc http://www.candelatech.com
