Please report all spam threads, posts and suspicious members. We receive spam notifications and will take immediate action!
Page 6 of 7 FirstFirst ... 4567 LastLast
Results 51 to 60 of 64

Thread: Intel vs. AMD - The Road Ahead




  1. #51
    Join Date
    Mar 2004
    Posts
    2,910

    Default

    Mt Tweak I am not am AMD. I'll ask a question or two.

    Don't you think your readers have been misled given the title of the article yet the story body has little to do with the actual processors at all?

    If so, you should etiher you should heavily modify the title or the story or pull it altogether.

    There was an article recently at HardOCP about this very practice by the less "mature" sites. A lot of people read and trust what appears on TT yet such an article, to the newer users, is actually misleading. (I don't know what recourse there could be if someone actually made a commercial decision of that article).

    I look forward to your "highly critical" "review" of the processor; complete with benchmark comparisons (not supplied by Intel of course), with full system specs etc, architectural benefits of the different processors, actual availability of the bits (chip, mobo and the memory) and the comparitive prices of course; wouldn't be objective without those would they?

    Given the "forward looking" previous article, it would also be nice if you quiz your "source" as to whether it will be the same socket/power form factor as the impending Yamhill with its cut down 13 stage pipeline or will it be another dead end?

  2. #52
    Join Date
    Mar 2002
    Posts
    8

    Default

    Quote Originally Posted by Unregistered
    quote:
    Actually a RDRAM platform coupled with a P4 is pretty much to best platform (for desktop) you can get to date. Obviously the constraint has always been the high cost due to RDRAM. If you look at reviews such as Toms P4 2.4Ghz vs AXP 2100+ you'll see he uses a RAMBUS platform as this shows the P4 in the best light (See: http://www6.tomshardware.com/cpu/02...p4_2400-06.html)

    sorry but i think that you didn't understand me. I said clock for clock as in a 1.7Ghz P4 and 1.7Ghz Athlon not a rating like 2100+ and compare it to 2.4Ghz P4. if the Athlon was running at the same clock as a P4 it would make it burn. so I think that what I said is still valid.
    hmmm, its kind of a silly comparision as Athlons are behind in clock speed compared to P4's. So in the realworld (what you can get today, not tomorrow) you would be comparing one of the top speed Athlons to a low end P4. Even AMD don't feel Mhz for Mhz comparisions should be used any more, which is part of their reasoning behind the numbering scheme 2000+, 2100+ (and obviously marketing :) ), and I must say I agree with them, Mhz mean nothing between platforms.

  3. #53
    Join Date
    Mar 2004
    Posts
    2,910

    Default

    "Not mentioning that much about the Hammer does not mean this article is biased, it just means there is little information about it to report on."
    That's funny.... AnandTech has plenty of information about Hammer, and its been there some time now... You can find all the Hammer information Here , but then again Anand is a real tech expert!!

    PLUNK

  4. #54
    Join Date
    Mar 2004
    Posts
    2,910

    Default

    Given that the K7 core is near the end of the road for scalability, and it's incorporated into Hammer, how do you suppose that will affect Hammer's scalability?

  5. #55
    Join Date
    Apr 2002
    Posts
    19

    Default

    "Given that the K7 core is near the end of the road for scalability, and it's incorporated into Hammer, how do you suppose that will affect Hammer's scalability?"

    Well, it is and it isn't...first it will be more (and longer) pipelines to allow for much higher clock speeds. This will be offset by having the Northbridge built into the chip die itself (vastly reducing the latency). Also, because it uses HT (with 3 HT connections built into the die), the bandwidth to and from the processor will be MUCH higher with (again) much lower latency. Third, it will be .13u SOI moving to .09u SOI, allowing for much higher clockspeeds. Fourth, it's not really one K7 core but 2...(at least for sledgehammer/Opteron) which the OS can actually see as 2 seperate CPUs (both on the same die), at least it will be by 2H '03.

    Cheers,
    Charles

  6. #56
    Join Date
    Apr 2002
    Posts
    19

    Default

    Hmmm...it appears as if my hope for a Sept launch may be too optimistic (by a month...). I have a friend at AMD who (in response to my constant whinging) told me today to "read the Inquirer" if I needed good data...from the Inquirer:

    "SOURCES CLOSE TO AMD's plans continue to tell us that the 28th of October date we're currently projecting appears so far to be spot on.
    And we also have projected prices of both processors and motherboards, with CPUs costing $400 and mobos hovering around the high $200 mark, as an initial thrust into the marketplace"

    http://www.theinquirer.net/25040202.htm

    I hope that it's appropriate that I post this stuff here (in this thread)...if it's not, someone please let me know...

    Cheers,
    Charles

  7. #57
    Join Date
    Mar 2004
    Posts
    2,910

    Default

    Hi Asher Moses. I was wondering what you mean by real world benefit. If you were to use 333Mhz memory on a 133bus athlon you could still recieve a minor performance gain over 266Mhz memory. Of course you could get a larger gain running both the cpu and memory at 166Mhz. This seems to be available to everyone. Products that allow this are easy to find. If you are still wondering what the point is of a KT400 chipset, I'll let you know... higher performance. Apart from handling 200Mhz fsb (which apparently some KT333 mb's are doing today) it should come with 8xAgp and an improved NB/SB connecter (2x faster than the previous). Of course, to take advantage of these one would have to "overclock" their babies, sometimes things like that happen in the "real world"!

    Do you really believe that AMD will stick w 133fsb for its athlon line till 2003? What reason could there be for such a move? The things to consider (from AMD's perspective) would be quality of product, availabilty and pricing. Although 333Mhz chipsets and memory are new in the market their similarity to 266Mhz products will make their adoption much quicker than the Sdram to DDRam changeover for example. The quality of product and pricing are also helped by the similarity to 266 product. To expect AMD's acceptance of this standard to be delayed till 2003 for the athlon in light of intc's 533bus in addition to all the other improvements to the pIV line that will be introed in this year does not make sense. To those who think that "validating" the athlon for 166Mhz isn't worth it (cough, anand) maybe they should consider where AMD makes most of its money.

    As for the Hammer line, it seems as if it could be possible to intro products before 2003. Rumors are all over. How long will it be before it contributes a lot to the bottom line, better make sure the current product is competitive till at least then. AMD says that athlon will be around for all of 2003. I don't think they would last that long with you in charge (joke!) !

  8. #58
    Join Date
    Mar 2004
    Posts
    2,910

    Default

    quote:



    Originally posted by Unregistered
    What would AMD have to do to make the Athlon run at a higher FSB frequency? What is involved engineering- wise? Is it possible to do something like the quad pumped bus of the P4?





    An entire new architecture I'd say.

    __________________
    Cameron "Mr.Tweak" Wilmot
    Founder and Owner of TweakTown


    If one were to have a 166fsb athlon homejob would they be an engineer?

  9. #59
    Join Date
    Mar 2004
    Posts
    2,910

    Default

    Mr. LSD, do you think that there will be any improvements in the thoroughbred vs palomino? Not considering L2 cache. I think that would be a mistake. I guess you must have already seen pics of the cpu.


    Butthead

  10. #60
    Join Date
    Mar 2004
    Posts
    2,910

    Default

    For those that say the article is 'lacking' because it does not address hammer are also missing the point that the article does not address future intel chips.

    The article is taking an analytical look at the here/now and the next 6-9 months.

    Though we cannot say what hammer will bring, we equally cannot claim on what Intel's Prescott (.09 micron P4) will bring. This is due out sometime early next year, and rumors are abound on this chip...

    Rumors now...
    - 1 meg L2
    - 800MHz FSB
    - And more die changes in terms of functions...

    A lot of you seem to assume that while AMD comes out with Hammer Intel is going to go along with their current Northwood line like nothing's changed...thats far from...


    I think this is wrong for many reasons:
    1)When do you think Hammer is coming out? In just over 6 months it will be November. Clawhammer will almost certainly be out, Barton will have been out for a while, and Prescott will still be way off. Prescott is over a year away, Hammer is not. By the time Prescott comes out an improved Clawhammer will probably be on the way. There will at least have been many more speed grades released (3 speed grades above the initial one will be introduced within a quarter of its release).
    2)Imagine that Prescott comes out with 1 MB L2, 800 MHz fsb, and all of the other rumoured improvements it will have. Compare this to a Sledgehammer with 1 MB L2 cache and dual channel DDR (at least dual 333 MHz). You could speculate that AMD's high end part would be at the very most at an equivalent price point to Intel's top end CPUs. AMD states that 20% of the improvement in speed of a Clawhammer over an equivalent clockspeed Athlon is because of the on-die memory controller. I greatly doubt then that Prescott will be able to get any more real-world bandwidth than a dual-channel Sledgehammer if the latency advantage of the on-die controller is so large. Hammer is also designed to scale to much higher clockspeeds than Athlon and doesn't need to scale as high to get the same performance increase. It is hard to believe that Sledgehammer wouldn't be competitive with the next P4, but unless Clawhammer is a bit of a failure so far as clockspeed scaling goes then it should be competitive too.

    My take on the article is that it is basically saying AMD will be left behind so far as performance is concerned because not only will Intel be changing to a 533 MHz FSB, they will be introducing their dual-channel DDR chipset. AMD on the other hand will introduce a 0.13 micron Athlon and by the end of the period getting talked about should have an Athlon with 640 KB cache. This gives a warped opinion though, because this Athlon will only be AMD's replacement for the Duron, and it will be getting introduced well within 6 months. In reality, AMD will have a processor that out that scales much better than Athlon, is 25% faster at the same clockspeed, has the potential for another 15% improvement (figure from Ace's Hardware) when compiled with a complier that takes advantage of the extra registers etc afforded by x86-64, supports SSE2 (thus removing what is a major advantage for P4 in programs that support it), and supports 64 bit addressing for > 4GB memory. It is sort of like speculating that ATI will be leading the Geforce 4 easily by the end of the year, and ignoring to mention that nVidia will be introducing new technology too.

Page 6 of 7 FirstFirst ... 4567 LastLast

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •