[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <20180424.110807.1097944821929554881.davem@davemloft.net>
Date: Tue, 24 Apr 2018 11:08:07 -0400 (EDT)
From: David Miller <davem@...emloft.net>
To: andrew.gospodarek@...adcom.com
Cc: talgi@...lanox.com, netdev@...r.kernel.org, tariqt@...lanox.com,
saeedm@...lanox.com, f.fainelli@...il.com
Subject: Re: [PATCH net-next V3 0/3] Introduce adaptive TX interrupt
moderation to net DIM
From: Andy Gospodarek <andrew.gospodarek@...adcom.com>
Date: Tue, 24 Apr 2018 11:02:46 -0400
> There is also an advantage that since this is done a per queue basis one
> queue that may be handling a bulk transfer can have its coalescing
> parameters adjusted while others stay at a setting that keeps traffic
> flowing at low latency. This is helpful when a system is receiving a
> large amount of traffic on one queue but also sending data on another
> queue and quick processing of acks keeps data flowing at high rate with
> low CPU utilization in both directions.
Ok, that's the missing piece on my end. My original analysis of this
problem space was on uni-queue NICs, and the problem there is that the
sampling algorithm is exposed to the traffic anomalies of the entire
link rather than a specific sub-class of traffic as is the case with
multiqueue NICs.
Powered by blists - more mailing lists