VIA has been incredibly quiet the last couple of months.  There was the near silent release of PT880 and KT880 a few months ago, but on the surface the company has been totally unspoken.  However, a quick look at our VIA roadmap anticipates the exact opposite.  Pay particular attention to Anand's comments about VIA's PCI-Express implementation.

As far as future AMD chipsets are concerned, VIA has been working on K8T800 Pro for Athlon FX Socket 940.  While this looks slightly promising, SiS's 755FX chipset is said to provide better performance and cost.  We really won't see 755FX nor K8T800 Pro until Socket 939 debuts anyway.  VIA's real effort will appear in a few months in the form of K8T890.   With 1GHz HyperTransport, PCI Express and AGP, we could see nForce3 250 take the backseat in the Socket 939 race.  Of course, now we just have to wait for Socket 939.

We mentioned before that nForce2 still takes the crown for Socket A.  However, K8T800 demonstrated better performance time and time again over nForce3 150.  Better performance and cost is the reason why so many vendors ended up adopting the chipset in the first place.  Our Athlon64 choice this week has to be the ASUS K8V Deluxe Socket 754 board.  This board really packs a punch including onboard WiFi, Raid, SATA, and an excellent audio package.  The MSI K8T Neo wins close second, but only because of price.  You really can't beat the features on the K8V.  Kudos to ASUS.

SiS is slowly clawing their way into the Athlon 64 world.  ECS finally was able to get their 755-A board under control with a second revision, the 755-A2.  Wesley Fink drew some surprising conclusions from the ultra-cheap board that actually surprised everyone a little bit (particularly during gaming performance).  Unfortunately, it took SiS and ECS two tries to get it right, but the ECS 755-A2 might be the fastest Socket 754 board for some time.

NVIDIA Motherboards Intel Motherboards
Comments Locked


View All Comments

  • AGM130 - Sunday, February 15, 2004 - link

    Anymore word on weather or not the A64 will drop in price? I have my debit card in a holster, just waiting!
  • KristopherKubicki - Wednesday, February 4, 2004 - link

    Trogdor, have you any other memory to test it with? I have not heard great things about Geil recently. I am not sure...

  • TrogdorJW - Wednesday, February 4, 2004 - link

    For Athlon 64 (socket 754), you list the MSI Neo FIS2R as a recommendation. Having just built one last weekend, I have to say that everything was great... except for the memory support. 1GB Geil Golden Dragon PC3200 2-3-3-6 timings would not run stable. I tried quite a few settings, but finally dropped to DDR333 to get it stable. So that's probably why it has this huge price drop: people are learning that it isn't very stable. Just my two cents. Make sure you get RAM that you know will run in the MSI board!
  • KristopherKubicki - Tuesday, February 3, 2004 - link

    "Your indication that the 3000+ A64 model would drop is incorrect (I already have the new pricelist), A64 pricing will not be cut"

    I have heard about vendors say both will be cut. Please email me -

    I do agree AMD has been much better with paper launches since the A64 introduction. This does only account for 4 processors but its a good start.

  • KeithDust2000 - Tuesday, February 3, 2004 - link

    Also this: "You may recall that Opteron 248 was first announced at COMDEX 2003. Here we are two months later, and retail wise the 248 and 848 is virtually non-existent. We have seen a few chips here and there for review samples, but it appears that both AMD and Intel embrace the paper launch with open arms.

    The 248 model has become available from Monarch at launch day, and they have it in stock today. The 848, at a cost of more than $3000, is hardly a retail product, it´s not like 4way is the enthusiast´s choice. The P4EE wasn´t available anywhere for many months after introduction, so please don´t lump these together. AMD had terrible execution in earlier years, but since the Opteron launch, this has changed significantly, and it would be great if your statements would acknowledge that.

  • KeithDust2000 - Tuesday, February 3, 2004 - link

    Christopher, maybe you should check this review of the Athlon 64 3000+ from Anandtech

    to see that the 3000+ is much more comparable to the P4 3.2 Ghz instead of the 2.8 Ghz model, as you try to make it seem. The only relevant mainstream area where the 2.8Ghz model would be faster would be on certain encoding tasks, and those people who mainly encode video all day (???) and think a few percent faster really matter for them should choose that model. Others are surely better served by the 3000+. Your indication that the 3000+ A64 model would drop is incorrect (I already have the new pricelist), A64 pricing will not be cut, only AXP pricing will, and your guess that the 3000+ would be cut to 2.8Ghz P4 level is wrong, and wouldn´t make sense based on its performance and feature advantages (Cool´n´quiet etc.). It stays at the P4 3 Ghz level, which is more than adequate, considering what Anandtech (and everyone else) found out about it.

    As for your question:

    "When was the last time we saw a solid launch and release date from either AMD or Intel in the past 2 years?"

    A64 launched exactly on the day that AMD set many months before. It was available at launch day. The same is true for the FX, 3400+, and the 3000+ actually launched ahead of schedule. That is unlike the P4EE, or the Prescott.

    Buying the 3000+ doesn´t lead to a dead-end either, as it will be at least viable for the 3700+, and AMD has on Aceshardware given the indication that 90nm processors would also be made available for S754 boards.

    As for a dryup of Athlon XP parts, that is not what channel checks indicate.

    Just FYI.
  • KeithDust2000 - Tuesday, February 3, 2004 - link

Log in

Don't have an account? Sign up now