DFI NFII Ultra: Mean Green Dream Machine
by Wesley Fink on July 30, 2003 6:25 PM EST- Posted in
- Motherboards
DFI NFII Ultra: Stress Testing
We performed stress tests on the DFI NFII Ultra in several different areas and configurations, including:
1. Chipset and motherboard stress testing, which was conducted by running the FSB at 223MHz; and,
2. Memory stress testing, which was conducted by running RAM at 400MHz in Dual Channel mode with two DIMM slots filled, and at 400MHz with all three DIMM slots filled at the lowest memory timings possible.
Front Side Bus Stress Test Results:
As standard practice, we ran a full range of stress tests and benchmarks to ensure the DFI NFII Ultra was absolutely stable at each overclocked FSB speed. These stress tests included Prime95 torture tests, which were run in the background for a total of 24 hours.In addition, we proceeded to run several other tasks, such as data compression, various DX8 and DX9 games, and apps like Word and Excel while Prime95 was running in the background. Finally, we ran our benchmark suite, which includes Sysmark 2002, Quake3 Arena, Unreal Tournament 2003, SPECviewperf 7.0, and Jedi Knight2. 223MHz FSB was the highest overclock we were able to achieve with the DFI NFII Ultra without encountering any reliability issues.
Memory Stress Test Results:
This memory stress test is very basic, as it simply tests the ability of the NFII Ultra to operate at its officially supported memory frequency (400MHz DDR) at the lowest supported memory timings that our Corsair TwinX LL modules support:| Stable Dual DDR400 Timings (2/3 banks populated) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 4T |
| Precharge Delay: | 2T |
| Command Rate: | N/A |
It’s not surprising to see the DFI NFII Ultra achieve such low memory timings – we often see 2-4-2-2 timings are possible with the better nForce2 boards. The nForce2 Ultra 400 chipset, like Corsair LL memory, seems to be designed to give the lowest possible memory timings. As we have seen in other reports of memory performance, this does not always translate into the fastest memory performance, but the lowest memory timings is a means of comparing motherboards. It is most useful when comparing boards based on the same chipset.
Filling all three available memory banks is more strenuous on the memory subsystem than testing two banks in dual channel mode, as it tests the rare occasion that a desktop user will install three DIMMs running 400MHz DDR at the most aggressive memory timings available in the BIOS:
| Stable Dual DDR400 Timings (3/3 banks populated) |
|
| Clock Speed: | 200MHz |
| Timing Mode: | N/A |
| CAS Latency: | 2.0 |
| Bank Interleave: | N/A |
| RAS to CAS Delay: | 2T |
| RAS Precharge: | 5T |
| Precharge Delay: | 2T |
| Command Rate: | N/A |
We were very pleased to see the fast timings that we achieved with all three DIMM banks filled on the NFII Ultra. Those who wish to use all three memory banks will not have to relax timings very much on this motherboard. We found 2-5-2-2 worked well with three banks filled, which is only slightly slower than the 2-4-2-2 timings that worked well with two DIMMs in dual-channel mode.
We tested all these memory timings using several stress tests and general applications to guarantee stability. We started the tests by running 24 hours of Prime95 torture tests. Prime95 ran successfully at the timings listed in the above charts. We also ran Sciencemark (memory tests only) and Super Pi. All of these stress tests ran on the DFI NFII Ultra without problems.
Facebook
Twitter
RSS
46 Comments
View All Comments
Anonymous User - Thursday, October 16, 2003 - link
I have been considering this mobo for some time and I read the reviews here and at Tomshardware. Both seem to be based on boards tweeked by the mfgr. I also looked at the forum at Amdmb.com recommended by Angry Games. What a mess. This board, like so many others using Nforce2, seems to be extremely fussy about memory timings. Several brands of memory strips lock up during posts. I say the reviews should be redone with retail versions after initial bios fiddling has settled down. And the question of stability at various timings should have as much importance as performance speed.Anonymous User - Friday, September 19, 2003 - link
I'm having a hard time figuring out the difference between the NFII Lanparty Ultra and the NFII Ultra-AL.The Ultra-AL seems to go for about half the price.
Do they perform/overclock the same?
Anonymous User - Sunday, August 31, 2003 - link
Its a cool mobo...whats all the fussLonyo - Tuesday, August 19, 2003 - link
640x480 for game tests would make most sense, then it's mostly CPU limted not GPU limited.1024x768 is silly if you're comparing motherboard using different graphics cards.
And doing a graph showing one board is pointless.
You should have delayed this review until you'd done another one with the same testbed setup.
Anonymous User - Sunday, August 17, 2003 - link
i myself was skeptical about the whole raid 15 deal, that just seems like marketing scheme to me as well. this does seem to be the case doesnt it? well i do believe these benchmarx are quite abbreviated, being that they lack a a7n8x! and a few others, and while on this topic, very disappointed with the whole video card fiasco...please review and update this article guys....sryque
ATConundrum - Saturday, August 16, 2003 - link
I could be wrong, but this comment board is for the review on the mainboard, not the performance/build of the AnandTech.com website itself. Send those comments to the webmaster instead, and do the respectable thing and post constructive comments about the review here.Anonymous User - Wednesday, August 13, 2003 - link
Anonymous User - Monday, August 11, 2003 - link
Just another anonymous post...I think he did a fairly decent job of reviewing the board, and the choice of an Radeon for the benchmarking due to directx 9 compatability is understandable. If you wanted to wait longer for them to dig out all of the other mobos, rebuild them in to systems, then retest each one again so that you could have your nice little benchmarks... sorry.
Just wait, either they will re-do the other benchmarks again, or do the new benchmarks using this card. I've already seen tomshardwares guide showing 10 mobos with this chipset, and the bechmarks between them were so similar that they didn't make any difference in my buying descision.
next to last: I agree with the previous plugin complaint, as I also use Mozilla. Is there ANY advantage in using flash over JPG of GIF for the benchmark results? Also the flash banners are annoying as... but I guess they pay the bills.
Finnally, I know I'm not spelling well either due to no spell check, stiff keys on this keyboard, and I ran out of give a care juice about 45 minutes ago.
Anandtech is still all good.
Anonymous User - Sunday, August 10, 2003 - link
just to add to my comment in 37 if you had a post for pics i could have placed a pic on just to show you i does look smart and others have started making enqiries to my tower AND LANPARTY motherboard tower from www.thermaltake.com take a look if u dont know the case im talking aboutthermaltake tower + lanparty motherboard = 1 smart looking system
Anonymous User - Sunday, August 10, 2003 - link
hey wesley 1 point of interest for u m8 when you made the board why didnt u make the IDE cables longer as i have a thermaltake tower so when i saw your board and read the speck i was impressed and imediately boutght it only to be dissapointed in the cables as i cannot connect from the raid connectoins to my HD=2off + dvd rom and dvd-re writer as the cables are too short this post is for ALL u tech wizzards that make motherboardsGIVE US LONGER CABLES FOR GOD SAKE now im stuck with using normal IDE connections AND NO im not reverting back to the old cable strips it takes the look away from a promissing motherboard when you add in a uv light to light up the inside