Dual Ch. DDR - How Much Faster in "Real World"?

How much faster will Dual Ch. DDR be for...

* Pro-level Music (Audio) Editing?
* Home Movie Editing?
* Graphics/Design Work?

[ assuming Same speed -- e.g., DDR 2700, at 333 Mhz, Non-overclocked -- for Dual and Single-Channel. ]

If I'm going to find it hard to detect <b>IN PRACTICE</b>, I'll just go for a Gigabyte or Asus PE mobo.

Barry Sindlinger (<A HREF="mailto:Barry@compleatpianist.com">Barry@compleatpianist.com</A>)
Int'l Award-Winning Pianist-Composer

"Giving Good Pianists the <i>Keys</i> to...
Unlock your <b>Unique</b> Brilliance,
<b>Transcend</b> ROTE Teaching and
<b>Create</b> audience-Dazzling ORIGINAL Performances"
23 answers Last reply
More about dual faster real world
  1. Right now the early (pre-release) Granit Bay boards are showing themselves equal in all performance arenas to the i850E chipset P4T533/RIMM 4200. So if you want a frame of reference, some shootouts compare the current DDR boards to that RDRAM board, which has the same performance as the new Granit Bay boards.

    I somehow expect the SiS 655 to offer better performance than the Granit Bay, just like the 645DX had better performance than the i845E, just like the 648 has better performance than the i845EP

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  2. The 655 should be a heluva lot cheaper too.

    To start press any key. Where's the "any" key? --Homer Simpson.
  3. close to 100 CAD lees

    Now what to do??
  4. close to 100 CAD less but for what i have look granite bay is still under RDRAM but much better for overclocking

    Now what to do??
  5. Quote:
    ... just like the 648 has better performance than the i845EP ...

    IIRC, that is not true ... I don't have the time (or I'm just to lazy) to search for the appropriate benchies, but I think the 845PE is about 1 to 2% faster than the 648, in general.

    Greetz,
    Bikeman
    PS: My excuses for the shortness of this post, but I'm into a very challenging msn-session right now ...

    <i>Then again, that's just my opinion</i>
  6. All the test I can find handy show the 648 beating the i845G/DDR333. Was there an improvement with the PE besides the "official" support for the DDR333 already supported on the i845G boards used for these test?

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  7. Officially DDR-333-memory was not supported on the i845G-chipset, I thought. I was capable of doing that, though. And now that I have the time, I looked up a comparison between the P/GE version and the SiS648: <A HREF="http://www.anandtech.com/showdoc.html?i=1723&p=20" target="_new">http://www.anandtech.com/showdoc.html?i=1723&p=20</A>.
    That page and the ones that come after this one, show that, when used with DDR-333-memory, the Sis648 motherboard that is used (Gigabyte one) ends up in the mid to lower field of the intel-based motherboards. When used in combination with DDR-400, the SiS is able to outperform the other chipsets, except for the i850E, most of the times. Another thing you can see, is that the 845G, running out of spec with DDR-333-memory, is slower than the 845GE. Anand's conclusion is this one: <i>"Now that we've seen all the benchmarks, we can safely conclude that the 845PE and 845GE combined with DDR333 memory are more or less a couple percent faster than the 845G with DDR333 memory. The SiS 648 with DDR333 memory is more or less equal to 845PE and 845GE. With DDR400 memory, the SiS 648 takes a tiny lead over the 845PE and 845GE."</i>
    Can we conclude out of that that Intel not only validated DDR-333 memory, but also improved their memory-controller a little between these two versions? But maybe this is due to the fact that mobo-makers found a PCB-layout that is less sensitive to noise, could also be.
    I hope this is enough to make my previous statement an acceptable one.

    Greetz,
    Bikeman

    <i>Then again, that's just my opinion</i>
  8. Hmmm.... Thanks for the help, Crashman!

    Now, can anyone point me directly to any particular "shootouts" (between the i850E chipset P4T533/RIMM 4200 & the current single-channel DDRs)? (I'll search, but if any of you would kindly recommend some, I'd appreciate it ; )

    Barry Sindlinger
    "Giving Pianists the <i>Keys</i> to...
    Unlock your <b>Unique</b> Brilliance,
    <b>Transcend</b> ROTE Teaching and
    <b>Create</b> audience-Dazzling ORIGINAL Performances"
  9. OK, well, it is possible that they optimized the timing for DDR333 on the PE/GE, but it's still only equal to the 648 when used with the same memory, while the 648 supports faster memory. This kind of leaves it up in the air as to the direction SiS chipsets are taking as compared to Intel chipsets in performance.

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  10. <A HREF="http://www.anandtech.com/showdoc.html?i=1708&p=6" target="_new">start this page and on</A>

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  11. Oh, those benches are for the P4T533-C and a few DDR boards. You'll have to search again for P4T533 to P4T533-C comparisons. The reason is that I don't know of any P4T533 benchmarks against the 8SG667.

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  12. Thanks, Crashman!

    However, I sure wish I had a clearer real-world comparison -- e.g., "It takes xx seconds to process y in PhotoShop" OR "xx seconds to encode an MP3 (or MPG)" (RDRAM or Double DDR vs. Single DDR)... :frown:

    Have you seen anything like that? (Again, my point is that if most of my work can be done with a nominal (nearly unnoticed) difference between Single & Double DDR, I'll go for Single DDR NOW; otherwise, I might wait for Double DDR!)

    Barry Sindlinger
    "Giving Pianists the <i>Keys</i> to...
    Unlock your <b>Unique</b> Brilliance,
    <b>Transcend</b> ROTE Teaching and
    <b>Create</b> audience-Dazzling ORIGINAL Performances"
  13. Obviously RDRAM will do better in memory intensive applications like Photoshop than DDR. Applications like that are what has made RDRAM shine. But if the price difference is 10% of total system price, for a total system that performs 5% better in that application, it's really a matter of how much your TIME is worth. Most users can get by with less. And then there are people with lots of money to spend where nothing less than the best will do. I'm in the catagory of overclockers who want the best performance possible at a reasonable price. So neither solution is fully adequate, as I want RDRAM performance at DDR prices. This forces me to prefer the wait for Dual Channel.

    Now, I multitask continuously. Which means HT would be a great benefit for me. But the only HT CPU right now is the 3.06, and it cost too much. And I can't wait for the price to drop on that one. I'd be happy if Intel were to backpedal and allow HT on a 2.53 since that will be economically feasable around the time of the Dual Channel board's release.

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  14. Although these benchmarks are from SMP systems, you may find it an interesting read. <A HREF="http://www.2cpu.com/Hardware/previews/msi_e7505_master/" target="_new">http://www.2cpu.com/Hardware/previews/msi_e7505_master/</A>

    *PIII-800 @900 i440BX SMP and Tualeron 1.2 @1.74 i815*
  15. Crashman, thanks for your thoughtful response. BUT... :wink:

    I still don't know how much faster Dual is than Single (DDR or RDRAM). You say (rightly),
    Quote:
    Applications like [Photoshop] are what has made RDRAM *shine.* But *IF* the price difference is 10% [more]...for a total system that performs 5% better.... (emphasis mine)

    As helpful as that may be... I still don't know what *in fact* the real-world application difference *IS.* (Is it near the 5% you suggest for your hypothetical?)

    Can anyone point me to "real-world" comparisons?

    Barry Sindlinger
    "Giving Pianists the <i>Keys</i> to...
    Unlock your <b>Unique</b> Brilliance,
    <b>Transcend</b> ROTE Teaching and
    <b>Create</b> audience-Dazzling ORIGINAL Performances"
  16. Dual Channel DDR has twice the bandwidth of single channel DDR. Dual Channel PC2100 has the same bandwidth as RIMM4200 and PC1066 RDRAM (because PC1066 RDRAM has 2100MB/s and is also Dual Channel for the P4). So what are the advantages? Better performance than traditional DDR platforms at less cost/identicle performance to RDRAM. Everything is theoretical and based on real benchmarks, not real programs.

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  17. Hummmm...

    My logic is thus:

    A. Dual DDR is THEORETICALLY on par with RDRAM.
    B. Dual DDR is BENCHMARKED as on par with RDRAM.
    C. Dual DDR is likely to be essentially as fast as RDRAM on "real tasks" in "real programs."
    D. Ergo, all that remains -- to know how much faster Dual is in the real world -- is to find "real tasks" comparisons of Single DDR & RDRAM.

    Illustration: All other things being equal, if the RDRAM system cuts 5 sec from a Single DDR's 30-sec multimedia process, so will Dual DDR.

    Barry Sindlinger
    "Giving Pianists the <i>Keys</i> to...
    Unlock your <b>Unique</b> Brilliance,
    <b>Transcend</b> ROTE Teaching and
    <b>Create</b> audience-Dazzling ORIGINAL Performances"
  18. Polyphon,
    Have a look at this link : http://www6.tomshardware.com/cpu/02q4/021001/xp_2800-16.html
    the benchmark for Pinnacle. Even XP2400 beats Intel 2.66Ghz although the benchmark for MPEG4 (the top benchmark in the same page), shows intel is better by small margin. Can anyone explain this ? Is dual channel ddr better for Athlon XP in video encoding ?
  19. Would you happen to know when SiS655 is actually coming out? I've been hearing about it for quite some time and dual channel DDR333 is an appealing thought, but I can't find any mention of approximate release date, ANYWHERE.

    The reason I ask is a friend of mine is looking to build a new high performance P4 system around the third week of december and I told him SiS655 would be ideal if he could wait for it... but I still don't know when it IS coming out!

    It'd be awfully nice to see it with HT... but it's not actually that important, since he's not going to get a 3.06 as it's way too expensive to be feasible price/performance.

    -Col.Kiwi
  20. It's aready past it's release date, and will probably be delayed again because if the HT revision, for another month. Granit Bay boards are just hitting the market now, Asus has already anounced the launch of the P4G8X, while the 655 based P4SDX will probably wait for the new chipset revision.

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
  21. Sounds to me like the end conclusion is right around christmas day... damn =/

    At least there'll probably be some amazing boxing day sales related to such things...

    BTW for anyone that's american and doesn't know (apparantly it doesn't exist in most of the US?), boxing day is december 26th and in canada there are lots and lots of big sales on that day with major discounts and stores rake in lots of money.

    -Col.Kiwi
  22. Q1 the lastest i have read

    Now what to do??
  23. We have the sales, but we don't call it boxing day.

    <font color=blue>You're posting in a forum with class. It may be third class, but it's still class!</font color=blue>
Ask a new question

Read More

Motherboards DDR