Final Words

The AOpen AK89 Max is aimed at a different market segment than the AK86-L. The AK86-L excels by bringing very high performance and reasonable features to a value-priced board at less than $100. The AK89 Max that we are testing here is aimed at providing everything but the kitchen sink. It is feature-rich with Firewire, 4-drive SATA, Dual BIOS, and optical and coaxial audio IO. It brings with it just about every option that AOpen could cram on a Socket 754 nForce3-150 motherboard. Compared to other top nF3-150 boards, the AK89 Max can certainly hold its own.

If the AOpen AK89 Max is evaluated just on the basis of our benchmark tests, the performance would be considered about average compared to other Athlon 64 motherboards we have tested. However, there is a lot more performance waiting beneath the surface of the AK89 Max - performance that is unique among the Athlon 64 boards that we have tested.

The AK89 Max is the first nForce3-150 board that we have tested with a true working PCI/AGP lock. Other enthusiasts report the Gigabyte K8NNXP and the Shuttle AN50R also to have working locks, but we have not personally confirmed this. We can tell you that with the latest 1.06M BIOS, AGP/PCI lock, and working multipliers, we achieved the highest overclocks yet on the Athlon 64. We were able to achieve a stable 233 FSB setting at default voltage with a standard multiplier. We also quickly outstripped the on-board range to 250 FSB. However, with the free Clock Gen for nForce3 from www.cpuid.com, we were able to fine-tune multipliers and FSB - reaching a truly astounding 347FSB.

When we first looked at the AK89 Max, we were not particularly impressed. In fact, we considered passing on reviewing the board with the first nForce3-250 boards starting to appear. We are glad AOpen persisted, however, because 2 BIOS revisions later, we see a remarkable nF3-150 board with tremendous capabilities. If we have a complaint about the AK89 Max, it is not with AOpen's implementation. AOpen got the nF3-150 right; it is just that it comes very late to the nF3-150 life cycle. The AK89 Max is still hampered somewhat by the maximum 600 HyperTransport of the 150 chipset. While the AK89 certainly makes the most of the 600HT and carries it to heights we didn't think the nF3-150 could reach, it really makes us just that more anxious to see a nForce3-250Gb motherboard from AOpen.

The AOpen AK89 Max is the best nForce3-150 motherboard that we have tested, but we already have nF3-250 motherboards that we are testing, which promise even more with working locks and a 1000 HyperTransport option. We do wish AOpen could improve the CPU voltage range and widen the available FSB settings. The AK89 Max is also the first nF3-150 or VIA K8T800 motherboard we have tested that deserves and can definitely use a FSB range to 350. But even without these options, the AK89 Max set new overclocking records in every category.

If you're currently in the market for a Socket 754 motherboard based on the nForce3-150 chipset, you won't do any better than the AOpen AK89 Max. We can't wait to see AOpen's nForce3-250 motherboard with all the lessons they learned in developing this design.

High End Workstation Performance
Comments Locked

23 Comments

View All Comments

  • cowdog - Wednesday, April 21, 2004 - link

    I have had a very different tech support experience with AOpen and my AK89 Max. I have filled every online tech support option through the AOpen website over the past 4 weeks or so and have yet to receive any kind of reponse. I have additionally sent email and posted on the AOpen support forum.

    My negative AOpen customer service goes back to buying the board with an AOpen rebate. After numerous calls trying to obtain the rebate form, I twice talked to someone at AOpen who took my email and promised to reply. They never replied.

    I have to wonder about your glowing comments about AOpen's tech support. Either that or I have simply had the worst possible luck.
  • Pumpkinierre - Wednesday, April 21, 2004 - link

    Good review Wesley. You should include 3mark 2001 results for the FSB275 (1:1) and 347 async.. I suspect the asynchronous would be disappointing as the RAM is the bottleneck not the ondie mem. controller. Also the FSB200 and 250 results should be run at the same mem. latency (the looser one- CAS3) so that we can see how FSB increase alone affects performance on the a64. Sorry if I sound like your mother but you were missing her anyway.

  • Pumpkinierre - Wednesday, April 21, 2004 - link

    test again

Log in

Don't have an account? Sign up now