Re: [lwip-users] LWIP/WIN32 UDP performance (throughtput)

[Bill Auerbach]

I totally agree with Simon, at least with RAW API – I can’t speak for NETCONN performance.  My “embedded” system is like yours (667Mhz CPU and 256MB  667Mhz DDR2) and is zero-copy with optimized checksum (aligned accesses and in assembly).  I can reach 500-600+MbS with TCP.  In my testing of UDP the PC in some instances was not able to keep pace receiving.  I do not run lwIP on the PC – we use Win32 sockets so far without issues.

 

It would be nice to see a standard Win32/Linux program to communicate with a simple lwIP test program (RAW, NETCONN, sockets) to create a list of lwIP platforms and the results of the tests.

 

From: address@hidden [mailto:address@hidden On Behalf Of Simon Goldschmidt
Sent: Tuesday, February 07, 2012 12:57 AM
To: Mailing list for lwIP users
Subject: Re: [lwip-users] LWIP/WIN32 UDP performance (throughtput)

 

Zayaz Volk <address@hidden> wrote:

Obviously, it raise the questions

1. Did anyone try to work with lwip on linux/Win32 and what throughput you've been able to achieve ?

 

In my opinion, the win32 port is mainly limited by the netif driver: winpcap doesn't seem to give the best performance, since we have to copy packets twice on RX (and I don't remember if it's once or twice on TX). Also, lwIP doesn't really benefit from multiple CPU cores. Instead, you're right that it is targeted on low resources rather that throughput. 

 

Still, i think lwIP should be able to achieve the performance you want if:

A) your netif driver correctly supports zero copy and

B) you handle the stack "correctly" so that zero copy is possible (I.e. data passed to a TX function must be left unchanged in memory until the netif driver has actually sent it out).

 

Simon