I've been hearing a lot about the Asus P3V4X on the big tech sites the past few months. I've seen it rated as the second fastest motherboard, barely missing first place. As I've been using Abit motherboards in most of the computers I build, most recently the BE6-II, I thought I'd give the newer Asus a try and see if I've been missing something. After all, I'm not generally loyal to one brand for it's own sake. To me that's something that a manufacturer has to earn with each model. I am, however, loyal to myself and my customers and will use whichever product I feel is best. That can mean the fastest, the most reliable or just the most compatible. I've been getting a bit of email asking me to recommend one of these boards over the other. I hope this article will help you make a decision.
There are several sites on the internet where you can find reviews of these 2 boards that go into technical detail and give scores with what have now become fairly standard tests for benchmarking. If any performance comparisons are done in games, they concentrate on First Person Shooters (FPS) pretty much leaving out the simulator community. I recommend that you read some of these reviews and weigh their results with what you read here. As always, Tweaks & Reviews is not as concerned with technical specs as it is with the practical differences between products. It's how they perform in ways you can see and feel that I consider important. Knowing that one board may score 1 or 2 more WinMarks may have it's place, but it leaves you unfulfilled if you want to know, in practical terms, how that translates into frames per second on your favorite sim. Most of you don't have access to the same tests and benchmarking utilities used in those tests, so you can't compare them to your own system's performance. I use Wintune98 for the raw computer scores. It's not quite as precise as other tests used elsewhere, but anyone can go to www.wintune98.com and easily test their computer and compare it's scores with others of similar specs. You can also get an idea of what kind of gains to expect if you upgrade your current system.
The test system consists of:
The "hot" Asus P3V4X has 4 memory slots, 1 AGP, 6 PCI and 1 ISA slot. The last PCI slot is shared with the ISA. In comparison, the "old standby" Abit BE6-II has 3 memory, 1 AGP, 5 PCI and 1 ISA slots. The last PCI slot is not shared with the ISA. I doubt that the extra memory slot in the P3V4X is much of an advantage to most users. I don't know anyone running more than 1500MB of memory anyway. I suppose if you have 4 sticks of 64MB SDRAMS it could come in handy. Both boards are very high quality. Their layouts are similar, but the Abit has 2 extra drive connectors for primary and secondary slaves. I do like the added connectors for convenience and expansion. If you add another hard drive and have a full house of CD-Rom, DVD and CD-RW drives, you'll need the extra connectors. I like them because it's sometimes easier to run 2 cables than it is to try to reach 2 drives with one cable. These connectors are mounted lower on the Asus board and are a little easier to reach, but their higher position on the Abit keeps the ribbon cables tucked away more for slightly better air flow and cooling.
Asus P3V4X Abit BE6-II
Both boards installed and setup easily, and both had drivers that had to be added once I got into Windows. The Abit needed the HighPoint ATA66 drivers and the Asus needed a suite of drivers (4) for it's VIA chipset. The Abit was one of the first motherboards to support ATA66 for hard drives. Although setting it up for ATA66 was not difficult, there was almost nothing extra to do for setting it up in the newer Asus board. The VIA chipset in the Asus did have some problems, however, as we'll discover later.
Abit was the first to come out with Soft Menu, the ability to set Front Side Bus speeds and a host of other things in the BIOS as opposed to using jumpers on the board itself. As anyone who overclocks their cpu knows, this is quite convenient. Both of these boards had Soft Menu as well as the ability to use jumpers on the board if you prefer to do it that way. The Abit board is capable of 2xAGP and the Asus is capable of 4xAGP. I'm not sure how well the 4xAGP worked, though. Even when set to this mode, the Asus monitoring utility showed it as 2x. Also, when I later installed a GeForce2 MX, there was no difference in performance when set to 2x or 4x. Of course, that doesn't mean you won't notice a difference in the future. I suspect that most people buying one of these boards is going to overclock, so I should point out some differences here. The BE6-2 has more options for setting various voltages, PCI speeds and FSB speeds. In fact, after 83mhz FSB, the adjustments go up one mhz at a time up to 200. This combination allows you to get the most out of your cpu. A definite plus for Abit. They also include an extra sensor (thermistor) on a wire. One end plugs into the board and the other end can go anywhere you'd like to get a temperature reading. The thermistor is paper thin and will fit anywhere. The P3V4X has a nicer utility for monitoring the motherboard. It runs in Windows and is very comprehensive. But even though it was fun to play around with, I soon turned if off by default. In my opinion, all you really need is "rain" anyway.
Each board had similar scores in Wintune98. I ran the tests several times and am posting 2 sheets for each that represents their overall findings.
Asus P3V4X (1)
Asus P3V4X (2)
Abit BE6-II (1)
Abit BE6-II (2)
It's interesting to note that although the cpu in both boards was clocked at 935Mhz, the Abit consistently showed a higher clock speed. This didn't only happen in Wintune98, it also showed this in the benchmark screen of Nascar Heat and every other cpu reading utility. Otherwise, I'd say the results are very similar on both boards. I made BIOS settings in each to maximize their performance. As each board was very stable, this was easy to do.
So, how did the games play? Well, this was a little surprising given the fact that the Wintune98 scores were so close. In Grand Prix Legends both pegged the fps counter at 36 most of the time. With the resolution set to 1600 X 1200 and sitting at idle in the pits at Monza, I did notice the frame counter showing in the 35's sometimes with the Asus board. This never happened with the Abit. Starting from the rear of a 20 car field, it also lost 1 fps more than the Abit. While running a race I didn't notice any difference. Similarly, in other games like Midtown Madness2 it was about 1or2 fps less than the Abit. These scores are close and you don't actually see the difference, but the big surprise was in Nascar Heat. The sim's internal benchmark showed 42 fps with the Abit board, which is what I've come to expect with the settings I use (see "GeForce2 MX vs. V5500" for details). With the Asus board installed, the frame rate dropped to 38. I would expect this score if the cpu was clocked at 850Mhz, but not at 935. I really have no idea why there was such a difference, but don't read more into it than exactly what's there. All it means is that the BE6-II is slightly faster in some sims than the P3V4X when running with a Voodoo5 5500 video card. This does not necessarily mean that it will be faster if you're using one of the nVidia video cards. I didn't test it that way.
When I was done with the testing, I installed the P3V4X into a computer I was building for a customer and decided to test it with a Celeron 300a that I'd used before and successfully overclocked to 464Mhz @ 1.95 volts. 2 volts is default on this chip, but it always ran perfectly stable with the core voltage reduced on the BE6-II. I was able to run Windows with the chip clocked at 473Mhz and some core voltage adjustments, but it would crash after a few minutes. With the BE6-II, I was never able to boot into Windows at this frequency. At 464Mhz, the computer ran much longer, but it crashed after a few minutes of running the Q2 timedemo. I tried several core voltage adjustments to no avail. Finally I set the chip to 450Mhz and the core to it's default 2 volts. After 24 hours of Q2 timedemo the system performed flawlessly. I was originally going to use some of my off the shelf parts for this computer, including a Diamond Sonic Impact90 sound card. No matter what I tried, I couldn't get that card to work properly. All the updated drivers loaded fine and the card was working properly according to Device Manager, but there was simply no sound. After a few hours of investigation, reading the VIA web site, the Asus web site and the Asus newsgroup, I found that this and several other sound cards have had problems with the VIA chipset used in the P3V4X. There were supposed to be additional drivers on the Asus and VIA web sites, but I couldn't find them (neither could anyone else). While I was on the Asus newsgroup I noticed that nobody from Asus tech support answered any questions there. It's been a while since I've been on the Abit motherboard newsgroup, but they did a pretty good job of fielding tech questions there. In time I may have been able to solve the issue with that sound card, but instead I gave up and installed an Ensoniq. Installation went smoothly and the computer is now absolutely stable.
Both motherboards performed very well. The Asus P3V4X has a lot of nice features and someday the 4xAGP may be an important feature. At this point, however, I don't see any reason to switch from the Abit BE6-II. The Asus offered no performance gains for my own computer, and the minor incompatibilities that are introduced with the VIA chipset make it less practical to use as a base for my customers' computers. After reading the P3V4X reviews on other sites, I must say that I expected it to be faster than the BE6-II. But for now, as a platform for running racing (or flight) simulations, I'd have to say that I prefer the small speed gains, compatibility and ease of use of the Abit BE6-II - the winner by a narrow margin.
GTX_SlotCar (Gary DeRoy)
January 19, 2001