1. Register NOW and become part of this fantastic knowledge base forum! This message will go away once you have registered.

Is it worth upgrading to Socket 939?

Discussion in 'Recording' started by crunk, Jul 20, 2005.

  1. crunk

    crunk Guest

    Hi, just registred and this is my first post in this awesome forum :wink:

    I own a Pentium 3.4Ghz Northwood CPU + Intel D875PBZ Mainboard + 4x 512 MB Ram.

    As you know the socket 478 is dead, no further CPUs will be released for the socket.

    Will an upgrade to an A64 system with (maybe) Athlon 4000+ CPU increase my DAW power/speed...?

    I can't use a dual system cause i am working mainly with Fruity Loops & lots of Vst's, and unfortunately it doesn't support dual CPUs. :evil:

    Would it be worth upgrading to an A64 system? Or is the speed similar to my 3.4 Pentium system? I just want to be able to use just more Vsti's & Fx...
  2. anonymous

    anonymous Guests

    the 939 subsystem is better for VSTi and lower latency than intel.


    SONICA-X Guest


    not quite true;

    1. The AMD 3500+ priced at $270.00 does not outperfom the Intel P4 3.4GHz priced at $270.00 even on a PCIe platform.

    2. The AMD X2 Dual Core 4200+ priced at $535.00 does not outperform the Intel D840 Dual Core priced at $535.00.

    Every test we have done using the Thonex test shows that when equaly priced processors from Intel and AMD are compared, AMD does not outperform Intel running the Thonex test.

    Your AMD biased graphs show the AMD processor in better light as you are not comparing apples to apples.

    As an example;

    a. You are comparing the Intel D820 vs. the AMD 4400+. The 4400+ wins but it is a $600.00 processor and the Intel D820 is a $245.00 processor. For $355.00 more the AMD 4400+ better perform better!

    Just my 2 cents,

    Guy Cefalu
    Sonica Audio Labs
  4. anonymous

    anonymous Guests

    HI Guy,

    LOL do you know how funny that sounds.
    "should compare based on price"

    if i had said that 2 yrs ago (when AMD was the whore and Intel was overpriced) you would have cried foul, foul.

    now that Intel is the Whore (due to knowing they are behind)
    its a differant story.
    you know darn well the Intel DUal core should be priced higher.

    only $100 more than the same speed single core? come on!

    our tests were very fair.

    i choose the 2 lowest from either line with 1meg cache for testing.

    2) even using a D830 it still dont beat the 4400.
    and dual vs dual middle procs on both.

    and if nothing else in our benchmarks only the AMD could do the lower buffer settings, how are you going to argue that?

    again your more than welcome to do your own shoot out between all the systems and post on your website.
    like i did on mine



    SONICA-X Guest


    it doesn't matter the fact is that a 4200+ is $535.00 and a Pentium D840 is $535.00 when I run the same test on both performance is the same. There is no AMD advantage. On the contrary, if we go Intel we can use system boards with all the latest technologies a luxury not available for AMD audio systems.

    The fact that Intel priced the dual cores cheaper is a good thing for consumers and bad for AMD. That is not my problem and lets face it these n3 boards are getting a bit old.

    Don't get me wrong I have nothing against AMD processors but I can not recommend to any of my customers the 4200+ over a D840 as performance is the same, price is the same and the Intel board has much more to offer.


    Guy Cefalu
    Sonica Audio Labs
  6. crunk

    crunk Guest

    Hi guys, interesting to read - which Amd64 CPU would be similar concerning to the speed of a 3.4Ghz Pentium Northwood 512 kb L2 Cache?

    Or in other words - to make the question more precise - which SINGLE amd64 can outperform my 3.4 Pentium CPU?

    Or shall i wait some time before upgrading & stick to (hotter) Intel CPUs for the future...? I don't know if DUAL is really an option, because my software doesn't support dual core (yet)
  7. Big_D

    Big_D Well-Known Member

    AMD gives you this information right in the chips name. An AMD 3400 is similar in performance to an Intel 3.4, a 3000 is similar to a 3.0 and so on.

    SONICA-X Guest

    Hi Big_D,

    we used to think that too until we got into benchmarks. When running the Thonex test a P4 3.4 with HT enabled outperforms the AMD 3500+.

    So I don't know waht to make of those numbers.

    My best,

    Guy Cefalu
    Sonica Audio Labs
  9. There is a problem with the Thonex test that can produce errors in real-world performance. It only measures usage and does not factor in efficiency. Benchmarking wouldn't be the proper term for it. Benchmarks measure how efficient a system is.
  10. anonymous

    anonymous Guests

    HI Brian,
    interesting you think that why?
    i agree its not the best of a real world. but it is a dang good test of a subsystem.

    how about our "Fudd" test. with looping 16 back in for record i dont think you can get more real world than that.
    and still have it be duplicable.


    FYI i would like to see a 3.4 P4 open that LOL

  11. SONICA-X

    SONICA-X Guest


    I agree 100%, yet when everyone uses it we have to use it too in order to have a common point of reference.

    Guy Cefalu
    Sonica Audio Labs
  12. SONICA-X

    SONICA-X Guest


    if the P4 3.4 can't open it then no sense on trying to run it on the AMD 3500+. :)

    Guy Cefalu
    Sonica Audio Labs
  13. For one thing, the Thonex test measures only CPU usage (according to the link). That is really only a minor factor in how efficient the CPU may be with the rest of your hardware. The architectures of different processors are different by design. Its not because of patents, but rather to achieve better performance in their target markets. Intel processors work much better on business applications, AMD works better on games. I don't use DAW's very often (I like my tape too much), but I do understand computers enough to know what should be factored in for benchmarking. I've worked with them for 12 years, and I am finishing my bachelor's this semester in IT (btw, I turn 20 at the end of the semester). Not trying to start a fight, just making sure people understand that that is not an accurate way to benchmark a system. I've done benchmark tests where the motherboard was the only changed component in either the AMD or Intel system and affected the results dramatically.

    Just because something can be a common point of reference doesn't mean it makes it accurate to base things on.
  14. anonymous

    anonymous Guests

    HI Brian,

    the test (thonex) includes a number of memory dependant VSTi
    and therefore is using both cpu and ram.

    of course there is no harddisk usage to speak of.

    the Fudd test indeed uses everything.

    however in testing a sub-system disk usage well has more to do with disk drives then testing a subsystem.
    and therefore not important.

    track count (untill large 96K) is all about harddrives we are trying to test harddrives.

    more to the point. there needs to be some "benchmark" used.

    as there is already so much bullcrap and yperbole out there, we have to be able to point to something as a standard.

    a good example is the nforce 4 subsystem. people have bought those and thought they had a good computer (mind you generally noobs)
    once they ran those tests NADA!

    same for a number of laptops, particularly the Clevo D900 that several so called "DAW builders" are selling as an audio solution.
    it chokes @ 12ms. and barely runs it @ 24ms.

    so while we may not have a "perfect solution" at least one thats not a 800 meg download.
    its far better than nothing.

  15. crunk

    crunk Guest

    i think i should stick to my pentium 3.4.

    got an asus p4c800-e deluxe & pc4000 ram from ebay and will do some overclocking up to 3.8 or 4.0 ghz. :wink:

Share This Page