Planning to build a new development workstation

Hi.

I'm planning to build a new PC in the near future. I've been considering using the machine build described here (core details: Athlon 64 X2 3800+, Asus A8V, 2x Kingston 512MB ECC DDR400, Seagate 250GB SATA), but the page is a little old (in PC-technology time) so I thought I'd look for some feedback first.

My target price range is $500--1000, and I'm more interested in stability and price-performance ratio than utmost performance. The system will primarily be used for development (editing, compiling, regression testing, etc.). Because I can hold off on this purchase for a bit, I figure can wait until after AMD's planned price cuts.

Does anyone suggest other parts? I've read the 939 is a dead-end upgrade-wise, but then I'm not much of one to bother upgrading machines so maybe that's not a big issue for me. Also, Core 2 looks pretty slick, but I get the impression they'll be more pricey than the X2 after AMD's price cuts (although I'm willing to be convinced otherwise).

Any feedback greatly appreciated.

Thanks.
4 answers Last reply
More about planning build development workstation
  1. Thanks for the quick reply.

    SupremeLaw: See the review in Computer Power User ("CPU") magazine for August, page 59, of the ASUS M2N32-SLI Deluxe. This might be just the ticket for your needs.

    Any relevant reasons I should spend $215 on that motherboard instead of e.g. $95 on the Asus M2V?

    You can start by overclocking an 805, [...]

    I'd rather not overclock (unless Intel and AMD have started covering that under their warranties).

    FYI: the future is DDR2 memory, not DDR.

    Maybe, but I thought benchmarks indicate little benefit for DDR2 memory today, the higher latency outweighing the additional bandwidth except at the highest frequencies.

    Thanks again.
  2. Quote:
    Thanks for the quick reply.

    SupremeLaw: See the review in Computer Power User ("CPU") magazine for August, page 59, of the ASUS M2N32-SLI Deluxe. This might be just the ticket for your needs.

    Any relevant reasons I should spend $215 on that motherboard instead of e.g. $95 on the Asus M2V?

    You can start by overclocking an 805, [...]

    I'd rather not overclock (unless Intel and AMD have started covering that under their warranties).

    FYI: the future is DDR2 memory, not DDR.

    Maybe, but I thought benchmarks indicate little benefit for DDR2 memory today, the higher latency outweighing the additional bandwidth except at the highest frequencies.

    Thanks again.


    1) Never ever buy a motherboard based on a VIA chipset. They suck. Instead get the Asus M2N-E (around the same price, but with vastly superior nVidia chipset).
    Asus M2N-E

    2) With AM2 you dont have a choice, you have to use DDR2. Go AM2 for drop in 65nm CPUs in December, and quad core in 2007. Get Corsair XMS CAS4, DDR2-800 RAM.
    http://www.newegg.com/Product/Product.asp?Item=N82E16820145034
  3. shadowduck: Instead get the Asus M2N-E (around the same price, but with vastly superior nVidia chipset). Asus M2N-E.

    Thanks, that seems like a more reasonable suggestion.
  4. Quote:
    My academic training biased me towards
    long-term planning, for better or for worse.

    I save $100 buying a MB with only one Gb LAN port,
    or I spend that $100 buying a MB with 2 Gb LAN ports
    (all other things being equal).

    I would rather have an unused LAN port
    out into the indefinite future,
    than to need a second LAN port but
    not have it when I need it.

    So, when I can afford it, I prefer
    "just in case" over "just in time".

    It's just my bias, but it works in important
    ways, e.g. when we upgraded to an HSDPA
    broadband modem, we used a free PCI slot
    for a PCI-to-PCMCIA adapter. Without that
    empty PCI slot, I would have needed to
    change to a better motherboard.

    As it was, we popped the adapter into the PCI
    slot, re-booted, and Windows XP detected
    it automatically. Then, the modem worked
    the first time in this adapter, because our
    workstation now functioned like a laptop
    with integrated PCMCIA slot.

    Also, the value of labor usually increases
    with time, and an extra $100 investment
    now will very likely save more than that
    amount in the foreseeable future.

    At $150 per hour today, that difference is
    only 40 minutes of time, whereas upgrading
    a motherboard takes a lot more time
    than that, particularly when shopping,
    shipping and travel are all factored in.

    Note that I didn't even mention the time
    required to re-load Windows and all other
    software which we routinely need and use
    on a daily basis.

    Last but not least, the machines we build
    need to work in a Gigabit LAN group,
    and that requirement by itself calls
    for a number of specific requirements
    which don't ususally come together
    on low-end boards.

    I hope that answers one of your questions;
    if not, chalk it up to different management
    styles.


    Sincerely yours,
    /s/ Paul Andrew Mitchell
    Webmaster, Supreme Law Library
    http://www.supremelaw.org/


    You know, considering the need for a 2nd NIC is very small anyway, it is still not worth $100 more for a motherboard just to have one. NIC cards cost like $10 to add in later, if there is ever a need for more than one NIC.

    There is no reason to spend extra money on a "workstation" board unless there is a feature needed that cannot be found on any other motherboard or added otherwise.
Ask a new question

Read More

Homebuilt Development Workstations Build Systems Product