Are they approximately the same distance from the router? Both cables look good, not pulled, kinked or abused in any way?
Barring actual transmission problems, there can also be differences in anti-virus software (or its configuration) as well as general load of the computer with other tasks. One has 64-bit OS, another 32-bit? Do they use different Ethernet frame sizes? Maybe even different HD write speed plays a role although the difference is a bit too big for that. But in any case, the processor clock frequency is just one of dozens of things that can affect download speeds. Unless it's affecting some extremely important application, I'd say let it be. You'll spend a lot of time tracking the problem down yet and advances you're able to achieve may not even be all that significant. For great many applications 3.82 Mb/s is still a decent download speed.
------------------------------------- Best Regards, Dmitri Abaimov, RCDD
Cabling Guide, Cabling Forum, color codes, pinouts and other useful resources for premises cabling users and pros
"Dmitri(cabling-Design.com)" wrote in message news:e346b$4d4ad7fd$42bb6765$ email@example.com...
Yes, same distance, I'll go over cables this evening.
Same antivirus (Avast) and firewall (Zonealarm), I don't know about 32 bit/64 bit or what an Ethernet frame size is. I'd like to make it faster, I was hoping my son would use this computer, but he noticed the download speed was much slower and continues to use what I call, MY computer. Mikek
On a good day, Internet speed tests are unreliable because they use such a small chunk of data with which to test. You're really getting the tiniest snapshot there.
Instead, if you have a third data source on the LAN, I would download a large file from there, first by one computer and then by the second computer. Since the end to end path is within the LAN, it removes all of the unknowns and inconsistencies of the Internet connection, which will be identical for both computers.
During the download, watch the throughput of the download in real time. Does it repeatedly burst, pause, burst, or does it go to a certain speed and generally stay there? If you use a Windows OS that's less than 10 years old, you can see your real time network activity on the Networking tab in Taskmanager, or you can install a small utility to do the same, such as DUMeter or Bandwidth Monitor or a free program that is a clone of those, but I forgot the name of.
Alternatively, check out Iperf or Jperf, a couple of benchmarking utilities. Again, keep the testing within the LAN, if possible, even if you have to use the two computers as a source for the other.
Assuming your LAN testing continues to indicate one being faster than the other, check out your NIC settings (negotiated connection speed, duplex settings, frame size, etc.). I don't think the CPU differences are to blame here, but you can see what the CPUs are doing by watching them in Taskmanager during a large download.
Connected via wireless, wired ethernet, or two tin cans and a string?
Linksys does not make a 54G router. What is the real Linksys model number including the hardware version? Extra credit if you know the firmware verison. The operating systems would be nice to know.
Some routers are really slow:
My guess(tm) is that you have a buffer window issue with your unspecified operating system TCP/IP stack.
However, using the internet download speed to "optimize" your TCP/IP settings is kinda marginal. You might want to setup IPERF or JPERF and run a local test without the internet. You might be surprised with what you find. Bug me for instuctions.
My take would be to verify the Ethernet ports themselves are set up the same. One might be running at half duplex, 10mb while the other might be at full duplex 100mb. Most modern router ports will try and match what ever the computer is set at.
Another test might be to boot the same version of Puppy or Lucid Puppy Linux from CD's on both machines and see what the speed are like. That would eliminate those pesky Windows Ethernet problems and other configuration differences that may have crept into the two machines.
Yep. You supply the data and I'll supply the abuse. The more
*NUMBERS* you supply, the better answers you'll get (unless you enjoy decoding generalization, pontifications, and guesswork).
Barf. It's the hardware version. Please send it to the recyclers. I have a fairly large pile of WRT54G v5 and v6 routers that have been retired from various customers due to flaky operation, chronic hangs, and unscheduled reboots. Installing DD-WRT seems to help on some, but the hangs and reboots never seem to go away. Last time I tested one using iPerf, the best I could do through the router was a very inconsistent 10Mbits/sec. How you managed to get 12Mbits/sec through it will remain a mystery. Well, I was using TCP and Speedtest.net is probably using UDP.
They're basically the same operating system. However, there were a mess of network related improvements in SP3, so you might consider installing it on your Media Center machine, which should slow it down to the same speed as the unspecified model Dell.
Not really. The machines are so much faster than the ethernet that once you get above about a Pentium 133, the CPU speed doesn't make much difference. I'm sure you would hear the screaming if the network traffic sucked 100% of the CPU cycles on a slow machine. Try the task manager to see what it's doing, or bettery yet, System Explorer:
Interesting, I disabled Avast and the download speed more than doubled into the 8 to 10 Mb/s range verses high 3s before. When I disabled Avast on the faster HP computer it did not increase the speed. 12Mb/s may be near the max the cable system provides. Thanks, Mikek