RTB

Good speed. Enjoy it

I'm doing good.

If you want to test your maximum bandwidth using wireless is often a poor choice. Wireless g has this theoretical speed of 56mbit, which is in practice at the very least halved. Wireless n might have enough bandwidth to cap out a 75/35 connection, but only if the laptop is really close to the router.
The ethernet result shows that something is wrong. It can be a hyperactive anti-virus, a virus, a cheap router unable to hit 75mbit, congestion, routing at the ISP side, or a bad signal for the modem, etc. Definitely worth investigating, since you're not even close to what you're paying for.

That would not solve the issue, and would essentially give you a single average number that is better left as a set of numbers for more information regarding routing.
There are two speed graphs that I think are most important for an internet connection: a graph of speed tests to an internal server (within the ISPs network) to determine the average and variance during the day/week/year/millennium of your connection to your ISP; and a graph of speed tests to a server outside the ISPs network, to find out how good your ISP is at handling that. Latency tests can also be very useful.
You are right in that a single server can be hampered by a single bad network, but it does point to a problem that is not supposed to exist, and will affect many more servers which you want to interact with.

Much depends on where you need the system for, so I'll just go for the generic details.
 
They're both refurbished and probably returned for a reason.
 
Both have Nehalem CPUs, which offer good performance but at a cost. The W3520 has a TDP of 130 watt, and if the cooling system was skimped on, it will overheat. The other system has 2x60W CPUs, which should be easier to cool. The dual socket system has 8 slower cores, so it will only provide a notable speed increase over the single socket system if the software you run is heavily multithreaded.
 
Oddly enough, both systems are triple channel yet run 4GB, so that's dual channel at best.
 
The single socket system has a mere 500GB harddisk, and the dual socket one is anemic at 250GB, which would probably require you to replace it.
 
The FX1800 is based on the peculiar 8800GS, and I have no idea if it's useful in any way.

That would not solve the issue, and would essentially give you a single average number that is better left as a set of numbers for more information regarding routing.
There are two speed graphs that I think are most important for an internet connection: a graph of speed tests to an internal server (within the ISPs network) to determine the average and variance during the day/week/year/millennium of your connection to your ISP; and a graph of speed tests to a server outside the ISPs network, to find out how good your ISP is at handling that. Latency tests can also be very useful.
You are right in that a single server can be hampered by a single bad network, but it does point to a problem that is not supposed to exist, and will affect many more servers which you want to interact with.

That would not solve the issue, and would essentially give you a single average number that is better left as a set of numbers for more information regarding routing.
There are two speed graphs that I think are most important for an internet connection: a graph of speed tests to an internal server (within the ISPs network) to determine the average and variance during the day/week/year/millennium of your connection to your ISP; and a graph of speed tests to a server outside the ISPs network, to find out how good your ISP is at handling that. Latency tests can also be very useful.
You are right in that a single server can be hampered by a single bad network, but it does point to a problem that is not supposed to exist, and will affect many more servers which you want to interact with.

That would not solve the issue, and would essentially give you a single average number that is better left as a set of numbers for more information regarding routing.
There are two speed graphs that I think are most important for an internet connection: a graph of speed tests to an internal server (within the ISPs network) to determine the average and variance during the day/week/year/millennium of your connection to your ISP; and a graph of speed tests to a server outside the ISPs network, to find out how good your ISP is at handling that. Latency tests can also be very useful.
You are right in that a single server can be hampered by a single bad network, but it does point to a problem that is not supposed to exist, and will affect many more servers which you want to interact with.

Let's see how well it works:



As opposed to east coast server:


A little surprised that it won't go faster to Washington. Would it be an idea and even possible to include a small ping test inside the download/upload test, so one could see whether latency is the bottleneck or not?

Or wait for the spambots to take over, whichever comes first

I do wonder if this is a thorough way to test ping. Most of the time I'm wondering if the connection here is stable, I hop onto a local quake3 or quakelive server and check the lagometer. Since those packets are being sent anywhere between 40 and 125 times in a second, you can really see any little stutter out there. It would be cool if CA3LE developed a method to test ping stability at low bandwidth and at high bandwidth usage (aka during a speed test), but with similar granularity. Since TCP is so sensitive to ping, this would explain most of the drops and thus lower speed results if worked out properly,