Ethernet performance
We evaluated Ethernet performance using the NTttcp tool from Microsoft's Windows DDK. The docs say this program "provides the customer with a multi-threaded, asynchronous performance benchmark for measuring achievable data transfer rate."

We used the following command line options on the server machine:

ntttcps -m 4,0,192.168.1.25 -a

..and the same basic thing on each of our test systems acting as clients:

ntttcpr -m 4,0,192.168.1.25 -a

Our server was a Windows XP Pro system based on Chaintech's Zenith 9CJS motherboard with a Pentium 4 2.4GHz (800MHz front-side bus, Hyper-Threading enabled) and CSA-attached Gigabit Ethernet. A crossover CAT6 cable was used to connect the server to each system.

The nForce4 boards were tested with the NVIDIA Firewall and Jumbo Frames disabled.

Over the nForce4's history, driver updates have fixed and subsequently broken the chipset's ActiveArmor Gigabit Ethernet acceleration. Everything seems to be working on the A8N32-SLI with the latest ForceWare 6.82 drivers, though. Throughput is great, and CPU utilization for the nForce4 GigE controller is much lower than that of the board's integrated Marvell Gigabit Ethernet chip.

Unfortunately, the ForceWare 6.82 drivers are only available for the nForce4 SLI x16 chipset. The DFI board's nForce4 Ultra chipset is running its latest ForceWare 6.70 drivers, and GigE CPU utilization is higher than we've seen with previous drivers.

To test reports that ActiveArmor and the NVIDIA Firewall corrupt large files, we also downloaded several gigabytes worth of large game demos and video files on the A8N32-SLI. We used the nForce4 GigE port with both ActiveArmor and the Firewall enabled, and we didn't encounter any problems with data corruption. However, given the nForce4's history of intermittent problems, your mileage may vary.