Re: Ecs N2u400-a

Benchmarking, Continued

Science Mark 2.0

Science Mark 2.0 is vastly different than the original Science Mark and has a much more robust benchmarking suite.

Memory Bandwith

Here we have the bandwith in MB/s for the memory and L1 and L2 caches. It's not surprising the N2U400-A has come out ahead in the memory bandwith department, considering the past two memory tests. What is very intresting is how the L2 cache on the N2U400-A has the largest lead over the Soyo KT600. This could explain the large leads under Quake 3 and 3DMark 2001SE that we saw earlier.

Cipher

The graph seriously distorts the N2U400-A's lead over the KT600 here. The execution time is in seconds and lower is better. ECS nForce 2 proves to be around 0.4 seconds quicker than Soyo's KT600.

BLAS

The BLAS scores on the single and double precision benchmarks shows the N2U400-A doing more work than the KT600.

Overclocking

As seen in BIOS, this board does not have any voltage adjustments. The good news is that the AGP/PCI lock does work, at least if Clockgen is to be believed. There has been some question as to wether or not ECS did implement this on the N2U400-A, and from what I can see, it's implemented, but not user adjustable. In my Soyo KT600 review, I was able to hit 160MHz on the FSB, but was cut from any further overclocking due to the PCI bus being overclocked and causing data corruption. It was also seen that 166MHz was not attainable due to lack of voltage adjustments. 166MHz did not last long for the Duron 1.4GHz. I found the system had either locked up or rebooted and hung. Starting off at 160MHz, I worked my way up and found that 165MHz was the highest I could get. I hoped that hitting 165MHz would mean 166MHz was game, but no go. Still, an extra 300MHz+ without any extra voltage isn't too bad for the AMD side of things. While I highly doubt we'll see it from ECS, I feel that if a BIOS with voltage adjustments comes out, the N2U400-A could be a real gem.

Final Words

The N2U400-A's performance really impressed me. In nearly every benchmark it beat out the KT600 competition, and this wasn't even with both memory controllers in use. Except for the floppy connector placement and maybe a capacitor or two nearby the CPU socket, ECS has a very solid layout. I would really like to see a socketed BIOS chip instead of one soldered onto the board, however. Additionally, the 12V 4-pin auxillary power connector is on this board, which a number of socket A boards don't seem to have. The BIOS could use some work in a couple of areas. First, the hardware monitor is able to measure the 3.3, 5, and 12V power supply rails, yet ECS fails to make them viewable in BIOS. The second thing that would really compliment the inclusion of the mounting holes around the CPU socket is some voltage adjustments for the CPU, chipset and memory. With some voltage options I believe the N2U400-A could be a real winner for the overclocking crew. It's price is right around $50 on-line, putting it head to head with Shuttle's AN35 Ultra, which does have the voltage adjustments.

Make no mistake, the N2U400-A is a solid board and it was a real pleasure to work with. I've heard of some stores giving it away with an Athlon XP 2500-2600+ CPUs, making it all that much sweeter. If you're an overclocker, this certainly isn't the board for you. However, if you need a solid performer on the cheap without alot of extra's, you'd certainly do well to look at the N2U400-A.

4.3 out of 5.