And finally (I'm kinda done messing around with this): I took out all other cards save for the AGP video card and re-ran iperf, and got 162 Mb/s with the XP PC sending to the Linux server and 612 Mb/s sending from the server to the PC. Unpleasantly asymmetric.<br>
<br><div>An abbreviated "lspci" (from the server) for those interested:</div><div><div>03:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 01)</div>
<div> Subsystem: ASUSTeK Computer Inc. Device 81aa</div><div> Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx+</div><div> Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-</div>
<div> Latency: 0, Cache Line Size: 64 bytes</div><div> Interrupt: pin A routed to IRQ 26</div><div> Region 0: I/O ports at de00 [size=256]</div><div> Region 2: Memory at fdfff000 (64-bit, non-prefetchable) [size=4K]</div>
<div>...</div><div> Kernel driver in use: r8169</div><div> Kernel modules: r8169</div><div><br></div><div>Numerically, the PCI codes are 10ec:8168 (rev 01); the PC's card reports as 10ec:8169.</div><div>
<br></div><div>I'll wait a bit to see if any new AMD CPUs are coming out, and then buy a quad- or hex-core CPU/motherboard/RAM and hope for a decent network interface.</div><div><br></div><div>Kevin</div></div><div><br>
<div class="gmail_quote">On Sat, Apr 10, 2010 at 5:52 PM, Kevin McGregor <span dir="ltr"><<a href="mailto:kevin.a.mcgregor@gmail.com">kevin.a.mcgregor@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Well. I plugged in my iMac (Intel, Core 2 Duo T7200, 2.00 GHz) to my gigabit switch and ran iperf on it in server and client mode. The only copy I could find for download (executable) was 1.70, and was compiled for the PowerPC Macs. Here are the results between the iMac and my server:<div>
<br></div><div><div class="im"><div>$ iperf -s</div><div>------------------------------------------------------------</div><div>Server listening on TCP port 5001</div><div>TCP window size: 85.3 KByte (default)</div><div>
------------------------------------------------------------</div>
</div><div>[ 4] local 192.168.27.10 port 5001 connected with 192.168.27.29 port 49371</div><div>[ ID] Interval Transfer Bandwidth</div><div>[ 4] 0.0-20.0 sec 2.17 GBytes 930 Mbits/sec</div><div><br></div>
<div>$ iperf -t 20 -c 192.168.27.29</div>
<div>------------------------------------------------------------</div><div>Client connecting to 192.168.27.29, TCP port 5001</div><div class="im"><div>TCP window size: 16.0 KByte (default)</div><div>------------------------------------------------------------</div>
</div><div>[ 3] local 192.168.27.10 port 44201 connected with 192.168.27.29 port 5001</div><div>[ ID] Interval Transfer Bandwidth</div><div>[ 3] 0.0-20.0 sec 2.18 GBytes 936 Mbits/sec</div><div><br></div>
<div>The server and the iMac seem pretty happy to talk to each other -- that's twice the performance of any other TCP result I've had! Just as a check, I plugged the PC into the same cable as the iMac had been plugged into:</div>
<div><br></div><div><div class="im"><div>$ iperf -s</div><div>------------------------------------------------------------</div><div>Server listening on TCP port 5001</div><div>TCP window size: 85.3 KByte (default)</div>
<div>------------------------------------------------------------</div>
</div><div>[ 4] local 192.168.27.10 port 5001 connected with 192.168.27.23 port 4701</div><div>[ ID] Interval Transfer Bandwidth</div><div>[ 4] 0.0-20.0 sec 386 MBytes 162 Mbits/sec</div><div><br></div>
<div>Much lower results. It seems to me that the problem is with the network hardware or TCP/IP stack on the PC side. Does anyone else want to venture an opinion? Or another test to run?</div>
<div><br></div><font color="#888888"><div>Kevin</div></font></div><div><div></div><div class="h5"><br><div class="gmail_quote">On Wed, Apr 7, 2010 at 4:11 PM, Trevor Cordes <span dir="ltr"><<a href="mailto:trevor@tecnopolis.ca" target="_blank">trevor@tecnopolis.ca</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div>On 2010-04-07 Adam Thompson wrote:<br>
> Actually, I was mainly hoping to verify that it was, indeed, a<br>
> hardware problem. One person (Trevor) reporting similar results has<br>
> fairly decent-quality GigE NICs on both sides – or at least what I<br>
> *assumed* to be fairly decent-quality NICs!<br>
<br>
</div>I should hope so, my NICs are Intel server grade gigabit on the server<br>
and Intel high-end workstation grade gigabit on the client ($100-$300<br>
NICs, retail).<br>
<br>
Kevin, I didn't have time to scan your exact results, is it mostly<br>
pc->server that's slow or server<-pc? And your pc is Windows, I gather<br>
(XP?).<br>
<br>
My big problem has always been windows->linux performance (but never<br>
linux->windows). I've given up on it for now, but one thing that made<br>
a HUGE difference was turning OFF jumbo packets. I instantly got 5X<br>
better performance with jumbo OFF. Yes, my switch is jumbo capable,<br>
and it was enabled, and set properly on the pc and linux. Go figure.<br>
I blame the Linksys WebSmart switch, but who knows.<br>
<div><div></div><div><br>
_______________________________________________<br>
Roundtable mailing list<br>
<a href="mailto:Roundtable@muug.mb.ca" target="_blank">Roundtable@muug.mb.ca</a><br>
<a href="http://www.muug.mb.ca/mailman/listinfo/roundtable" target="_blank">http://www.muug.mb.ca/mailman/listinfo/roundtable</a><br>
</div></div></blockquote></div><br></div></div></div>
</blockquote></div><br></div>