nForce4 SLI Motherboards: Premium Performance at a Bargain Price
by Gary Key on January 3, 2006 12:05 AM EST- Posted in
- Motherboards
Disk Controller Performance
With the variety of disk drive benchmarks available, we needed a means of comparing the true performance of the wide selection of controllers. The logical choice was Anand's storage benchmark first described in Q2 2004 Desktop Hard Drive Comparison: WD Raptor vs. the World. To refresh your memory, the iPeak test was designed to measure "pure" hard disk performance, and in this case, we kept the hard drive as consistent as possible while varying the hard drive controller. The idea is to measure the performance of a hard drive controller with a consistent hard drive.
We played back Anand's raw files that recorded I/O operations when running a real world benchmark - the entire Winstone 2004 suite. Intel's iPEAK utility was then used to play back the trace file of all IO operations that took place during a single run of Business Winstone 2004 and MCC Winstone 2004. To try to isolate performance difference to the controllers that we were testing, we used the Seagate 120GB 7200 RPM SATA (8MB Buffer) in all tests. The drive was formatted before each test run and a composite average of 5 tests on each controller interface was tabulated in order to ensure consistency in the benchmark.
iPeak gives a mean service time in milliseconds; in other words, the average time that each drive took to fulfill each IO operation. In order to make the data more understandable, we report the scores as an average number of IO operations per second so that higher scores translate into better performance. This number is meaningless as far as hard disk performance is concerned as it is just the number of IO operations completed in a second. However, the scores are useful for comparing "pure" performance of the storage controllers in this case.
It is impressive that the ATI SB450 IDE performance tops the NVIDIA nForce4 SATA 2 performance in both benchmarks.
Firewire and USB Performance
After looking at many options for Firewire and USB testing, we finally determined that an external USB 2.0, Firewire 400, and Firewire 800 hard disk might be a sensible way to look at USB and Firewire throughput.
Our first efforts at testing with an IDE or SATA drive as the "server" yielded very inconsistent results, since Windows XP sets up cache schemes to improve performance. Finally, we decided to try a RAM disk as our "server", since memory removed almost all overhead from the serving end. We also managed to turn off disk caching on the USB and Firewire side by setting up the drives for "quick disconnect" and our results were then consistent over many test runs.
We used 1GB of fast 2-2-2-5 system memory set up as a 450MB RAM disk and 550MB of system memory. Our standard file is the SPECviewPerf install file, which is 432,533,504 bytes (412.4961MB). After copying this file to our RAM disk, we measured the time for writing from the RAM disk to our external USB 2.0, Firewire 400, or Firewire 800 drive using a Windows timing program written for AnandTech by our own Jason Clark. The copy times in seconds were then converted into Megabits per second (Mb) to provide a convenient means of comparing throughput. Higher Rates therefore mean better performance in this particular test.
With the variety of disk drive benchmarks available, we needed a means of comparing the true performance of the wide selection of controllers. The logical choice was Anand's storage benchmark first described in Q2 2004 Desktop Hard Drive Comparison: WD Raptor vs. the World. To refresh your memory, the iPeak test was designed to measure "pure" hard disk performance, and in this case, we kept the hard drive as consistent as possible while varying the hard drive controller. The idea is to measure the performance of a hard drive controller with a consistent hard drive.
We played back Anand's raw files that recorded I/O operations when running a real world benchmark - the entire Winstone 2004 suite. Intel's iPEAK utility was then used to play back the trace file of all IO operations that took place during a single run of Business Winstone 2004 and MCC Winstone 2004. To try to isolate performance difference to the controllers that we were testing, we used the Seagate 120GB 7200 RPM SATA (8MB Buffer) in all tests. The drive was formatted before each test run and a composite average of 5 tests on each controller interface was tabulated in order to ensure consistency in the benchmark.
iPeak gives a mean service time in milliseconds; in other words, the average time that each drive took to fulfill each IO operation. In order to make the data more understandable, we report the scores as an average number of IO operations per second so that higher scores translate into better performance. This number is meaningless as far as hard disk performance is concerned as it is just the number of IO operations completed in a second. However, the scores are useful for comparing "pure" performance of the storage controllers in this case.
It is impressive that the ATI SB450 IDE performance tops the NVIDIA nForce4 SATA 2 performance in both benchmarks.
Firewire and USB Performance
After looking at many options for Firewire and USB testing, we finally determined that an external USB 2.0, Firewire 400, and Firewire 800 hard disk might be a sensible way to look at USB and Firewire throughput.
Our first efforts at testing with an IDE or SATA drive as the "server" yielded very inconsistent results, since Windows XP sets up cache schemes to improve performance. Finally, we decided to try a RAM disk as our "server", since memory removed almost all overhead from the serving end. We also managed to turn off disk caching on the USB and Firewire side by setting up the drives for "quick disconnect" and our results were then consistent over many test runs.
We used 1GB of fast 2-2-2-5 system memory set up as a 450MB RAM disk and 550MB of system memory. Our standard file is the SPECviewPerf install file, which is 432,533,504 bytes (412.4961MB). After copying this file to our RAM disk, we measured the time for writing from the RAM disk to our external USB 2.0, Firewire 400, or Firewire 800 drive using a Windows timing program written for AnandTech by our own Jason Clark. The copy times in seconds were then converted into Megabits per second (Mb) to provide a convenient means of comparing throughput. Higher Rates therefore mean better performance in this particular test.
Facebook
Twitter
RSS
37 Comments
View All Comments
Gary Key - Tuesday, January 3, 2006 - link
Corrected. I apologize for that error and the previous mention of DDR2 in the Asus memory section. My right side brain was thinking DDR but the Intel side of the brain was typing DDR2. ;->
JarredWalton - Tuesday, January 3, 2006 - link
Nah, these are M2! Seriously, though, thanks - busted Gary on the copy/paste, eh? Although he did get the memory speeds right, so you can't go too hard on him. :)Martrox - Tuesday, January 3, 2006 - link
I've got a K8SLi that I picked up for $80. I'm currently running a 3000+ Venice at 2610(9x290) memory set to 166fsb(TCCD 2x512)2.5,3,3,7. Running 7800GT's SLi'ed with absolutely no problems.......Wish the voltage options were better, but it does work pretty well.lopri - Tuesday, January 3, 2006 - link
The title says, and the article begins with "Premium performance at bargain price" then it ended up praising (and recommending) ASUS board. The ASUS board doesn't belong to same ground as that of the other two boards, and the price of the board certainly proves it. In the first page how affordable NF4-SLI boards have become is mentioned ($70), then compare 2 value boards with a premium board($170)? What's the point of this review? Did ASUS ask to boost the "A8N-SLI Premium" board since they can't meet the demand of "A8N32-SLI"?I'm sorry but as an avid AT reader/forumer, I should say this review is totally off. It's nothing but an ASUS PR presentation. Also, it's been pointed out many times lately, but according to AT reviews, every single board on the market is a "great overclocker," "dark-horse," "keeper," "enthusiast-targeted". Everytime AT reviews a motherboard they miraculously find something to compliment and that becomes the main point of the article.
It's almost up to the point that I no longer completely trust AT motherboard reviews. I'm sorry to say this, especially the author of this specific article, but my comment is more general towards the whole AT motherboard team. (Wesley, Jared, et. al)
Is this going to continue? (Sigh..)
JarredWalton - Tuesday, January 3, 2006 - link
Just curious, Lopri, but you did notice that the Foxconn and Albatron received a recommendation based on price, right? There are people that want all the features of something like the ASUS Premium, and they're willing to pay for it. All three motherboards here are pretty decent, though, and while I didn't write the article I'm a little confused as to how this is seen as a PR piece.To a certain extent, all reviews are a form of PR. Companies send us a product and we review it. We review other products sometimes, but generally speaking we review the products that we get sent. The only things that really separates these boards from each other are features and overclocking capability. In most of the benchmarks, scores are close enough that you needn't lose sleep over the difference.
Wesley Fink - Tuesday, January 3, 2006 - link
We've complained plenty about the ATI SB450 southbrudge (poor USB 2.0 performance) even though we like the ATI chipset in general. We've also complained loudly about nVidia's mediocre AC'97 audio on their top boards when ATI and ULi are delivering hooks for HD Azalia in their AMD chipsets.Frankly, until very recently Asus had not had a really great AMD board since the Socket A days - they hadn't won a single AMD roundup in quite a while. Asus rested on really great Intel boards. Recently Asus got very serious about the AMD market and the A8N32-SLI Deluxe is a great board. So is the A8R-MVP at the price point of around $100. However, as another recent review clearly showed, the Asus 6150 is pretty pedestrian and not a great choice for a multimedia PC with all the important multimedia stuff an extra cost option.
A lot of the real crap boards we see get pulled from the market before release and our review never sees the light of day. An example is the original Asus Premium reviewed here, which was a terrible overclocker due to the switches that replaced the paddle for SLI switching. When we complained to Asus they decided to hold off release of the board and rework the switching. The result is the board reviewed here which is a huge improvement. Oddly, it overclocks FSB very well but really does not do a great job in overclocking memory. The A8N32-SLI Deluxe is still the better board for the Enthusiast.
The motherboard team really cares about the products, and they aren't looking for a sensational story like some other sites. We could write Inquirer-like exposes on plenty of the early BIOS' we receive, but we prefer to work with the manufacturer to try to resolve the issues on production boards. If we can't get things to work we ask lots of questions before we assume we are Gods and the board is junk. But if it's junk or just all hype in the end, we'll tell you that.
We began testing features more in-depth because the memory controller on the AMD CPU greatly reduced the performance variations we used to see in motherboards. Also most manufacturers have gotten more serious about going after the Enthusiast who is our basic reader. That's the reason you're seeing so many really good overclockers these days. Just when we think it can't get any better another board sets some new record. However there are lots of boards out there below these standards, and we do tell you that when we review them.
mindless1 - Tuesday, January 3, 2006 - link
I see nothing wrong with this concept, makes plenty of sense.I would not however consider this to be "Premium Performance at a Bargain Price". Back in the socket A days an AMD platform at a bargin price was $50-70 for your basic nForce2, of course a little higher when first deputed. It's very hard to make an argument that any board over $100 is bargain in this age with so many of the features chipset-integral.
If anything, the typical trends of motherboard features and performance are still holding true but at a far worse bargain than ever before. SLI should not cost much more except for the smaller market segment.
On a side note- Cooling the Asus board by moving the heat from the southbridge to the hottest zone on the board just to save a fan is a really bad idea. Almost seems like an afterthought for marketing purposes. Given the space that thing takes up they should've just use a giant low-profile passive 'sink if they really needed to stay fanless.
mindless1 - Tuesday, January 3, 2006 - link
"... by moving the heat from the southbridge"I too would like an edit function. ;-)
Beenthere - Tuesday, January 3, 2006 - link
...Mobo for the vast majority of PC users as the article states. Most people have no need or use for SLI and have no interest in paying a premium price for SLI when they will never use it. SLI is a gimmick for gamers and nothing more. If you want it, buy it, but I highly doubt mainstream PC users will cough up $170 bucks or more for SLI when you can have Asus's A8R-MVP Crossfire Mobo for $101. from Mwave.The $70. you save could go toward a nice Vid card and the A8R-MVP overclocks every bit as good as the SLI boards and performs better than all of the SLI Mobos per Wesley Finks recent review.
http://www.anandtech.com/mb/showdoc.aspx?i=2617">http://www.anandtech.com/mb/showdoc.aspx?i=2617
At $101. mainstream PC users can have the best performance, overclocking and not even worry about paying a premium for Crossfire or dual graphics cards. If in the future they really desire to blow more dough on a second video card, they could, but there would not be a price penalty for purchasing the superior A8R-MVP Mobo like there is with any of the SLI rigs.
Some1ne - Wednesday, January 4, 2006 - link
I agree, I feel that the article makes some unfounded and moreover highly irresponsible statements, such as "in fact, the current price structure almost ensures that your nForce4 purchase should be an SLI-capable motherboard." SLI is not worth it in any way, shape, or form from any cost/performance standpoint, unless you happen to be the enthusiast user who wants the highest possible performance available today no matter the cost. For everyone else SLI is worthess...and yet how many new users are going to go out and waste their cash on an SLI board because of statements made in the article like the one above?Irresponsible, and not backed up by any solid evidence. It's like the author started off with the assumption that SLI = good, so therefore any affordable SLI board is also good. It doesn't work like that though.