- Native Gigabit Networking
Something that we haven't seen before now is the inclusion of native Gigabit Networking. While we saw Intel move the networking chip off the PCI bus to their CSA bus, it still had a separate chip which naturally caused latency. nVidia has gone a step further and given the 250GB Gigabit networking on the actual MCP.
The easiest way to understand how native Gigabit networking is more beneficial is to think of the very first stages of Serial ATA. When Serial ATA was released last year, due to the bandwidth limitations of the PCI bus, you were unable to experience the maximum bandwidth of your Serial ATA hard disk drives. When moving the controller from a separate chip directly into the nVidia MCP, it reduces the traveling time (latency) of data to an absolute minimum providing the best possible performance. This method works the same way for the new native Gigabit networking from nVidia. By having the controller on the MCP chip itself, latency is reduced and as a result the maximum performance of Gigabit networking can be achieved since we have the required and free bandwidth to play with without worrying about flooding any other system bus.
One thing worth noting is that while the nForce3 250GB can offer this huge amount of networking bandwidth, unless you are connected to motherboards on your LAN utilizing this technology or switches being able to handle this amount of bandwidth, you will not see a difference as you will be hitting a bottleneck in the transfer speeds of not only networking but regular SATA and PATA desktop hard drives not in a RAID as Gigabit networking provides higher transfer rates.
In the past there were problems with the PCI bus being flooded which caused crackling to sound when onboard networking, sound card or other PCI devices were used at once. Fortunately now we have the full PCI bus bandwidth needed to run all the extra PCI cards including sound cards, TV tuners and so on due to native devices becoming more and more popular not by choice but by necessity.
- Increased Hyper Transport Bus
As we mentioned previously, the only real performance feature that we can see is the increased Hyper Transport bus. nVidia didn't produce the original HyperTransport bus fast enough. nVidia turned around and released the nForce3 250 with a true 16-Bit/800MHz each way as apposed to the previous 16-Bit/600MHz down and 8-Bit/600MHz up.
We will find if this little piece of information is something that affects the performance in some of our game tests today.
- A working PCI Lock
One feature that gave nVidia's older nForce3 150 the competitive edge over the VIA K8T800 was the locked PCI speed.
Although it didn't take long for people to quickly discover that the lock didn't actually work. While we didn't get a chance to test if the PCI lock was working, numerous tests from other media have shown that the new nForce3 250 does in fact lock the PCI bus (and subsequently the AGP bus).
It goes without saying that this is excellent news for overclockers finally wanting to make the move from their highly overclockable Athlon XP's knowing they can have an Athlon 64 solution capable of locking the PCI bus for maximum overclocking.
Find the lowest price on Motherboards!