VERY slow networking performance using Gigabit Switch

ryanv12

Senior member
May 4, 2005
920
0
0
I'm having this nagging issue that I can't seem to resolve - I get very slow network performance when I plug my 3 computers into a gigabit switch, but only when I use distributed rendering in 3D Studio Max.

It will be fine at first, and then get progressively slower and then nearly not work at all within about 5 minutes. This doesn't occur during regular network use though.

I think I isolated it to the switch because when I plug all the computers directly into the router it works like a charm. I tried using a different switch but I have the same issue.

For what it's worth, I'm using a Linksys WRT54G router with one cable going from it to a Linksys SD2005 or Netgear GS605 switch.

If anyone has any insight please let me know! Thanks!
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
Most likely a cabling problem or a duplex mismatch. Make sure you are using store bought CAT5e or higher cables (not homemade) and that the network cards are set to autonegotiate speed and duplex.

Gigabit ethernet is different from a cable perspective from 100 Base-T, it uses all 4 pairs and a much higher baud rate. So if something works fine for 100 Base-T, doesn't mean it's OK for gig.
 
Dec 26, 2007
11,782
2
76
Yeah, gig switches are a bitch. I still need to do some work on mine to figure out wtf is going on with it and get back to the tech with DLink....
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
Originally posted by: DisgruntledVirus
Yeah, gig switches are a bitch. I still need to do some work on mine to figure out wtf is going on with it and get back to the tech with DLink....

To be fair, there's nothing special about a gigabit switch. Ethernet performance problems still stem from the physical layer almost every time - cabling and duplex mismatch. These two are far an away the root cause, almost every single time. The technology is over 8 years old, there really isn't anything special about it.
 

Crusty

Lifer
Sep 30, 2001
12,684
2
81
He's right... there is nothing special about gigabit. For instance...

matt@svn:~$ wget 10.0.0.202/green.tar
--21:55:17-- http://10.0.0.202/green.tar
=> `green.tar'
Connecting to 10.0.0.202:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 1,513,881,600 (1.4G) [application/x-tar]

100%[============================================================================================================>] 1,513,881,600 110.10M/s ETA 00:00

21:55:30 (108.55 MB/s) - `green.tar' saved [1513881600/1513881600]
 

Crusty

Lifer
Sep 30, 2001
12,684
2
81
ifconfig -a
eth0 Link encap:Ethernet HWaddr 00:c0:92:90:e5:c0
inet addr:10.0.0.2 Bcast:10.0.1.255 Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe90:f8c0/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1060527 errors:0 dropped:0 overruns:0 frame:0
TX packets:116258 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1588671114 (1.4 GB) TX bytes:8607498 (8.2 MB)
Base address:0x2000 Memory:d8920000-d8940000

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:12 errors:0 dropped:0 overruns:0 frame:0
TX packets:12 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:760 (760.0 B) TX bytes:760 (760.0 B)


ifconfig -a
eth0 Link encap:Ethernet HWaddr 00:30:48:b9:2b:00
inet addr:10.0.0.202 Bcast:10.0.0.255 Mask:255.255.255.0
inet6 addr: fe80::230:48ff:feb9:2b00/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:12182246 errors:0 dropped:0 overruns:0 frame:0
TX packets:7932450 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:16843239625 (16.8 GB) TX bytes:4426387415 (4.4 GB)
Memory:faee0000-faf00000

eth1 Link encap:Ethernet HWaddr 00:30:48:b9:2b:01
BROADCAST MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
Memory:fafe0000-fb000000

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:37256 errors:0 dropped:0 overruns:0 frame:0
TX packets:37256 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:3396605 (3.3 MB) TX bytes:3396605 (3.3 MB)
 

BoboKatt

Senior member
Nov 18, 2004
529
0
0
Wait so all items connected to my Gswitch or Grouter should be on autonegotiate instead of forcing say 1000T if that option is there?
 

Crusty

Lifer
Sep 30, 2001
12,684
2
81
Originally posted by: BoboKatt
Wait so all items connected to my Gswitch or Grouter should be on autonegotiate instead of forcing say 1000T if that option is there?

The only time I've ever needed to set something other then auto negotiate is when using crappy equipment or if your cabling needs work. 99.999% of the time I leave it at auto negotiate.
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
Originally posted by: BoboKatt
Wait so all items connected to my Gswitch or Grouter should be on autonegotiate instead of forcing say 1000T if that option is there?

Yes, absolutely. There is hardly EVER a good reason to force speed and duplex, especially with gigabit ethernet. It's extremely common for folks to force 1000/full or 100/full thinking they'll get better performance, but what really happens is the switch will operate at half duplex because your NIC is not sending any autonegotation information. This is called a duplex mismatch and will give really bad performance.