Using LWIP 0.8 (the current Xilinx distribution) raw API on a 300 MHz PPC in Xilinx's ML403 development board, but otherwise a very similar setup to Bill Auerbach's, we can receive about 13 Mbytes/sec from an old WinXP machine (net payload). This is using programmed I/O to read the TEMAC FIFOs which is very very slow, so we are moving to using custom DMA. Lwip processing for a packet takes about 14 microsecs with the TCP checksumming turned off. (I had previously reported receiving 300-400 Mbps using older drivers, but I now think that figure was erroneous.)
Paul www.PaulTobias.com _______________________________________________ lwip-users mailing list [email protected] http://lists.nongnu.org/mailman/listinfo/lwip-users
