Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Intel Moves To 533MHz FSB 147

homerj79 writes: "Intel has launched an upgrade 850 chipset and faster Pentium 4's today. The new chipset, dubbed the 850E, supports a 533MHz (133MHz x 4) front side bus, as do the processors. Supporting processors come in speeds of 2.53, 2.4 and 2.26GHz. The 2.4GHz part is denoted as supporting the new FSB by a 'B' tagged to the end of it. And it appears as if the new chipset gives the P4 a performance boost in most apps over the previous 400MHz FSB chips and the Athlon XP." Meanwhile, back at the other processor ranch, firemoth writes: "Today OCAU has something special - They've gotten their hands on 3 AthlonXP CPU's based on AMD's new "Thoroughbred" core. This is the .13 micron process, of course, with lower voltage. This article compares them to the older Palomino core in both speed and temperature.. and they throw one into a Vapochill supercooling case and see just how fast it can go."
This discussion has been archived. No new comments can be posted.

Intel Moves To 533MHz FSB

Comments Filter:
  • by questionlp ( 58365 ) on Monday May 06, 2002 @01:06PM (#3470700) Homepage
    According to Anand's article [anandtech.com] on the new 533Mhz FSB P4's and the i850E chipset (which provides official support for the 533Mhz, aka quad-pumped 133Mhz, FSB), that Intel isn't officially supporting the use of PC1066 RDRAM modules which would allow the memory bus and the processor FSB to run in tandem.

    Although quite a few Samsung PC800 modules will run at PC1066 speeds without any problems, but if any installed modules are not capable of running at the higher speed, the memory bus will get capped at the current max of 400Mhz (or 3.2GB/s).

    I guess for now, the new processors don't really, really need the higher memory bandwidth, but as the processor speeds start to hit 3+ Ghz, the extra amount of bandwidth will become more important.

    • I forgot to make a comment about the AMD T-bred processors. Why doesn't AMD start putting some type of protection on the naked cores? I would really like to see AMD to put on the integrated heat spreaders on the processors to help reduce the chance of cracked and broken cores, as well as help dissipate the heat from the now smaller cores.

      Sure, the T-bred processors won't be producing as much heat (due to the die shrink), but once speeds start hitting well over 2Ghz, I expect the cores to get very, very hot again.

      Just my $0.015 post-taxes.

      • Adding a heat spreader between the core and heatsink would reduce the heat transfer efficency and increase the production costs. As for protecting the delicate core from careless home builders, I'd much rather them push motherboard and heatsink vendors to provide a better retention system. They actually seem to be backing away from the 4 hole/screw mounting system (as used with the Swiftech 462 and some Alpha heatsinks) which I think is a mistake. Also, you need to remember that 99% of AMD's (and Intel's for that matter) customers are OEMs who presumably know how to attach a heatsink.
    • It gets worse... (Score:1, Insightful)

      by Anonymous Coward
      Why does Intel boost the FSB speed and then turn around and bork it with old crap on the southbridge?

      ICH2 South Bridge connected to the MCH via Hub Link 1.0 with 266MB/sec bandwidth;
      ATA/100, USB 1.1 and AC'97 sound support.

      ATA/133 and USB 2.0 are pretty much the standard now on Socket A motherboards.

      Intel killed this motherboard line before it was even made...
      • I agree... Intel should have updated the southbridge to include at least USB 2.0 and maybe improved AC97 sound support. The difference between ATA/100 and ATA/133 isn't really enough to make an upgrade, but the support for > 130GB hard drives is a must.

        I would like to see Serial ATA to be out in the wild as that would help reduce the wiring mess, mostly for those with more than 4 devices on multiple controllers :)

        • Although I am pleased to see that you can now buy nice, neat parallel cables that coiled up into a neat, round, (if somewhat thick) package. When I have some spare cash, I'll be buying some.
      • The ICH4 southbridge will be out over the summer or the fall this year. It'll have USB2.0 (_much_ faster than NEC's offering), better AC'97, etc.
      • It looks like Intel has an i850-E based motherboard that has USB 2.0 built-in (I'm guessing that they planted a separate USB 2.0 controller on the motherboard). The model is the D850EMV2L and one can be purchased here [gamepc.com] .
  • by dattaway ( 3088 ) on Monday May 06, 2002 @01:09PM (#3470737) Homepage Journal
    I remember the old ZX81 I built from a kit clocked its Z80 at 3.58MHz, so it could generate the synced television picture directly from the processor.

    With 2.4GHz, I'm sure there will be wireless experiments by attempting to use the CPU as a DSP.
    • The real fun won't hit until 2.4GHz memory busses come along. Only having the CPU core run at 2.4GHz is like having a radio with no antenna. But get a couple inch long PCB trace with a 2.4GHz signal on it, and things get more interesting.

      On the same note, the VGA output nowadays is fast enough to generate AM broadcasts. Or so I'm told. I've never tried it personally.
      • Or you can integrate the 802.11b or 802.11g controller into the processor and integrate a special type of heat spreader that allows someone to bolt-on a heatsink (a la Alpha) and use the heatsink as a mini-antenna ;-)

        Or the other option is to have the controller on the processor and have a couple of pins from the packaging go directly to some type of MII/PHY chip, which in turns does the RX/TX with an external antenna.

        Just some thoughts... dunno how feasible either of them are.

        • by BeBoxer ( 14448 )
          Yeah, I don't know either. I work with 802.11b stuff a fair amount, but I'm no R/F engineer. Still, I can't help but imagine this becoming an all to common occurence:

          D: Greetings, Dell tech support.

          A: Hi, this new desktop you sold me is junk.

          D: What model is it?

          A: The new 2.4GHz P4, with the integrated wireless ethernet and wireless bluetooth keyboard.

          D: And what seems to be the problem?

          A: Every time I try to make a call on my 2.4GHz cordless phone, the computer crashes! And when I surf the web, my phone rings! And everything I type is ending up in my Palm's ToDo list! Then while I was upstairs heating my coffee in the microwave, it caught on fire!

          I mean really, how much stuff can we possibly cram into the 2.4GHz band anyway? Interesting times anyway.
          • I mean really, how much stuff can we possibly cram into the 2.4GHz band anyway? Interesting times anyway.

            Hmm.. I've never ever heard of 900Mhz computer crashing when someone made a GSM (nor NMT900) call...
      • But get a couple inch long PCB trace with a 2.4GHz signal on it, and things get more interesting

        Get too many of those and you won't be able to meet gigahertz timings, much less multigigahertz timings. Also, at frequencies like that, you start running into transmission line issues with the longer PCB traces. If the PCB trace isn't an integral fraction of a wavelength (1/4 of a wavelength, 1/2 of a wavelength, it's really too complicated to explain here), the traces start to introduce complex impedances. This is more of a design issue for the engineers than anything, but it's another reason why you probably won't see much, if any PCB traces running at the full 2.4 GHz.
    • I can see it now, 4 years from now.

      Athlon 46XP 600 terahertz: $150

      Visible Light: 430 THz to 750 THz

      Removing a heatsink from an athlon to produce a useable "gas" flame: priceless.
  • by Anonymous Coward
    Can't believe that Intel isn't supporting the ICH4 with the i850 (ICH4 adds USB 2.0 support among other things.) This is reported at Tom's Hardware I believe. Intel's ass backwords manner of support for PC1066 RDRAM stinks too.

    So when are i845E boards coming out? Why didn't Intel announce that and the i845G today? Probably some stupid contractual thing with Rambus I suppose.

    Sure do hate it when marketing and politics overrule good engineering. Intel should be building ATA-133 and Firewire/IEEE1394 support into their chipsets as well. And the i845E should have support for DDR300.

    What a bunch of losers.
  • 533 Mhz?! (Score:2, Interesting)

    by adam613 ( 449819 )
    So Intel put the P4 on a quad-pumped bus to get the clock speed to look better. When AMD put the Athlon on a double-pumped 133 MHz bus and said it had a 266 MHz bus speed, nobody believed it. Now even Socket A motherboards admit that they run at 133 MHz now. What gives with Intel doing this? Am I missing something that's supposed to impress me?

    (if I am, please tell me, because otherwise I will be buying an AMD processor for my new computer)
    • I have an AMD 1.4 GHz and its too damn hot. I had to create a wind tunnel in my case to keep it cool.
      IS the noise really worth it?
      • I have an athlon XP 1.4ghz (1600+) overclocked to 1.57ghz (1900+ I think), and yes, heat is a definite problem. But with AMD coming out with a .13 micron part this summer (Thoroughbred), it will hopefully soon be solved.
      • I have a 1.3GHz Athlon, and it's fairly cool. With a Volcano 7, I haven't managed to get it above 50C (and that's in a room that was at about 30C). No wind tunnel needed. In fact, since it rarely got about 45C, I swapped out the fan with a quieter but less powerful one (40CFM, not sure how many dB), and it still rarely goes about 45C. My brother's Celeron (a 566, not overclocked) runs at about the same temp with the stock heatsink.

        My point is that, while Athlons do run pretty dang hot, all you really need is a decent heatsink and good normal case cooling (I have those same 40CFM fans as the front and rear case fans).
    • Intel was already using 4x100. Now they are using 4x133. I don't know that I'd call it impressive but using a 4xX bus instead of a 1xX bus DOES give better performance. As it stands right now AMD systems still have a price/performance advantage but the difference is small until you get up near the cutting edge. Up untill a few months ago I was building AMD systems almost exclusivly but now I'm building some P4 systems for those rare customers who actually run apps. that can take good advantage of Intels' SSE2 instructions.
    • Re:533 Mhz?! (Score:2, Informative)

      by jtshaw ( 398319 )
      Ya, it is a little miss leading, but effectively it is kinda true. Since it pushes 4 things at once at 133Mhz. it is the same as doing one thing at 533Mhz. Same as AMD pushing two things at once doing 133Mhz is the same as doing 1 thing at 266Mhz.

      In a way I understand too, it is just easier to explain to the non-technical you are running at 266 instead of 133 or 533 instead of 400 or whatever the case may be then it is to explain that you are using both edges of the clock or using such and such method to get 4 times as much through the bus at one time thus creating this thing, that while it runs no faster in Mhz. then before, preforms x times faster.

      I mean, Mac tried to tell people there PPC's were faster then Pentiums for years any many wouldn't buy it because it was slower in Mhz. So you can see why they do what they do.

      Afterall, it isn't about the Mhz. that should matter to us that really know what is going on, what matters is the increase in memory bandwidth. And that is a very real increase. As it was when AMD went to DDR.
    • Ok, the buss runs at 100mhz (133mhz now), but transfers 4 times per clock, giving it an effective rate equal to a 400 (now 533) mhz bus. This DOES make a difference on the P4. It's a high memory-bandwidth chip.
    • 533MHz is not an attempt by anyone to deceive you. It is a genuine number indicative of the throughput the bus provides. Nobody (educated, at least) is claiming the bus clock runs at 533MHz, only that 533 MHz is simply the rate of transfer. The clock rate is effectively multiplied by 4 at each end of the transfer and used to synchronize the transfers. This is analogous to your system clock running at 100MHz, yet your core claims to run at 1GHz.

      How can this be? The core has an internal multiplier (called a PLL) that allows the core to run at a higher speed. This allows the board to run at a lower speed which results in less board noise, less switching (so lower power), and also represents a common denominator reference so all devices don't need to run at the same high frequency. For example, the CPU might scale the reference clock by 10, but the chipset might not scale the clock at all.

      Hertz (Hz) is merely a rate, and represents an inverse second (1/s) and is commonly used to indicate the rate of periodicity of a clock. However, Hz is used to indicate bandwidth of a modem, for example. A 48K modem transfers 48 kilobits of data, and can be said to transfer bits at a rate of 48KHz. In actuality, there are not 48000 individual bits transferred every second. Instead, there are 8000 individual symbols transferred each second, and each symbol provides 6 bits.

      Best regards,
  • do they have any memory bandwidth issues? i mean, i'd be a shame if intel's new found speed was useless
  • who cares what the FSB is at this stage. untill we have SDRAM that supports this "bandwidth" it will not help out total system performence.
    • One or the other has to come out first, may as well be this. Sure it's chicken and egg, but necessary. Saying this won't help total system performance is like saying Watt's improvements on Newcomen's steam engine were no biggy until iron rails came into play. Improvements come in steps, lots of tiny steps, always has, always will.
    • PC3200 RAM, which can match the speed of the 400MHz FSB, is already available (although it's non-standard). If you want more bandwidth than that, use two channels.
    • Who cares what the FSB is at this stage. Until we have SDRAM that supports this "bandwidth" it will not help out total system performance.

      Um, I believe this was the whole idea behind RDRAM - that it does have the necessary bandwidth. (Latency, of course, is another issue.)

  • by Anonymous Coward
    Are there any reviews or benchmarks available for the new CPUs using 266MHz DDR memory? It looks like RDRAM is not the most popular memory, but Intel is only allowing reviews that use RDRAM and not DDR. Onquiring minds want to know how a normal P4 config compares to a normal Athlon config.
  • by KanSer ( 558891 )
    I liked it when the processor and RAM cycle didn't resemble the videocard race. I wish they could build around all the RAM I just bought instead of a processor upgrade requiring new MOBO's, RAM, and more patience then ever. Then again it's sooo fast. Me want... doom 3 ain't gonna run on this old p3-450. Damnit, I'm gonna cry now.
  • by colmore ( 56499 ) on Monday May 06, 2002 @01:25PM (#3470857) Journal

    I'm running a 900 mhz Duron right now, and have been for a year and a half.

    I do some pretty heavy photoshopping, media work, and compiling. I've never felt that I was being slowed down significantly by my processor.

    I upgraded to 1GB RAM, and that improved things, but I just don't feel the need to go out and get the latest processor any more, the need is simply not in the applications.

    Maybe if I were playing games or ripping DivX, but really it seems that for the vast majority of the home market, the technology has vastly outpaced the need, even the latest-greatest MS OS can't justify the need for this kind of iron.

    Though I guess this means that the next computer I buy will be that much cheaper, since "low end" systems (with slight boosts to hard drive and memory) are now all that I need.
    • Suit yourself. Me, I want an 8-way Opteron :-)

      I do software for film visual effects, and when you're dealing with dozens of layers of footage that's 250 MB PER FRAME, you need all the CPU power you can get, trust me. I have customers with 100-CPU render farms, and those farms do not sit idle. There is still an ever-growing need for high-powered machines, even if it's not in your house.

      Every time I hear a comment like this, I think of an article I read in 1980, from a columnist who grumbled that the new-fangled 16 bit CPUs were just too fast - his 8-bit CPU was quite enough to run WordStar and 1-2-3, thank you very much. I wonder if he still thinks the same.

      • umm well I am sure you are not representative of the whole population of computer users.

        we have reached a point in computing where there is very little that can leach performence from a PC in the software realm.

        the next big boost that software will give hardware is in the next big interface change, other than that, the standard Metaphore and all the tools you can use just are not pushing the limmits any more.
        • i agree with you on that. i had a p400 for 3 years before fiddling with the motherboard/cmos battery finally died on me. probably not a smart thing to do in the first place.... anyways, i ended up with an apple tibook, which does everything i need, and then some. wish it had an audio-in port :-/

          inherited my dad's p200 when he died and i use that mostly for kazaa and video capture. it's too slow to play divx movies, and i didn't want to invest in a video card to play counterstrike on it. still plays mpeg2, flash, and quicktime movies just fine, along with winamp and aim, ie and netscape - 99% of what i do anyways. needed a laptop for college next year, though.
          • that TiBook will erver you well.

            get a copy of Virual PC if you need to compile anything for X86, you can get a codewarior IDE for students for $99 that compiles both for Mac and Windows.
            • i have VPC 5.0 on there right now. running win 98 on there...slow as a dog. kazaa wouldn't install properly. codewarrior works like a charm, however. also, the 200 is in a case FAR away from where i sit, keeping my desk area much more quiet. VPC 5.0 seems to make the fan come on at high speed continiously. plus, it's easier to monitor my downloads on a 17" screen, though :)
    • even the latest-greatest MS OS can't justify the need for this kind of iron.

      I think you're vastly underestimating Microsoft's talent for bloat.
    • Speak for yourself, and it should be noted that this exact same argument has played out since the days of the 386. I virtually guarantee you that if you search the forums, you can find dozens of people proclaiming that the new 486 is well and good, but their 386 does more than they ever need, etc. It's a dated, predictable argument.

      I recently upgraded from a P3 850/512MB PC133 to an Athlon XP 1800+ with 512MB of DDR333, and the difference in simple day to day applications is stunning (one of those "You don't know better until you've tried better" type deals), and for applications like Java or Visual Studio.NET it's a requirement.
      • More likely something else was slowing you down on your old system. Maybe you had a configuration problem or a slower disk.
      • Hey, if you are actually into Java or Visual
        Studio.NET, great.

        Me, I'm into writing documents, um... that's
        words, with (maybe) a bit of formatting, um...

        Actually, CPM/80 with WordStar is about all
        I really need, and BOY is it fast on my
        current 'puter. Only thing is, everyone
        else seems to like *ffice style products.

        Me, I'm content with WordStart, vi on unix,
        I use troff (runoff under CP/M). Works good.
        Fast, _and_ Cheap.

        Upgrades? Maybe not in this lifetime (I could
        care less). Standards? Sure, every is ASCII
        or close to it. WSWG? Nope, but if I go so
        far as to ASK for (say) Centuury Schoolbook, 11pt
        on 12pt lead, I EXPECT that it will be delivered.

      • No... I've been using computers for near 15 years, and for the whole time I've felt like there was more software out there than I could run. (i've never really had an up-to-the-minute system)

        Only recently have I found budget systems to be completely adequate for everything I see myself doing .

        I'm not saying running all of my apps faster wouldn't be *nice,* but it's not the same as saying there are new apps that will only be functional on these faster processors.

        what I'm saying is, like always, I can't afford an upgrade, but unlike say, when my pentium was two years old, I don't feel like I'm missing out on anything.
        • I can totally appreciate the idea of buying bargain PCs, and the Athlon XP 1800+ system that I bought I bought only once it got to the high value point (the chip can be bought for $189 CDN here, which is quite a sweet point). I recently bought a GeForce 3 Ti200 because it was unbelievably cost effective.

          However, that doesn't man that one should yawn at new PC components coming out: I love every new graphics card, processor, and memory technology -> Maybe I'm not going to buy it today, but in 3 months it'll be bargain basement and I'll be very grateful.
    • I was using a Celeron at 450 up until last month.
      I u/g to an AthXP 1700, and I wouldn't say I feel it's even twice as fast generally. Only compiling is noticibly faster.

      Then came an ATA100 HD to replace my old HD as the backup disk. A little faster for things that need lots of HD access, but really, how many apps do that?

      Anyway, my primary SCSI disk from 98 still beats it.

      Then I tried a few games. Maybe I don't have very intensive games, but again, OK it loads a ~little~ faster, 6s instead of 9s. Gameplay unaffected.

      Really different from the old days when I remember the huge difference adding 16MB ram to my P133, or the jump from that to the 450.
    • Doesn't it make anybody wonder why clockspeeds rise with so small steps at a time? Why FSB and other buses always get only this kind of 100 -> 133 improvements? Why memory channels were only doubled and not quadrupled at once? Why consumer-level SMP mobos are mostly 2 fixed sockets and not 8 slots for processor cards and possibility to use riser cards after riser cards for extra processor slots infinitely :)
  • by Anonymous Coward on Monday May 06, 2002 @01:36PM (#3470930)
    when the clock speed of the latest chipsets is faster than the clock speed of your processor, motherboard, and memory. Combined.
    • As other have said in other posts, that depends entirely on what you're using it for. I have an old P2-266MHz with 128MB of EDO ram, and an really crappy video card (does 640x480, 8 bit color ONLY) - I'd imagine my cumulative clock speed, counting everything, is under 400 MHz. That's like 5 times less than the current P4 offerings. But it's running as an OpenBSD firewall and filesharing device (SMB and NFS), and I'd be a fool to waste my money on CPU upgrades. Except for the occasional make build to upgrade to the next obsd version (which takes under a day anyway), i'd be wasting my money investing in a faster anything, since its all IO bound anyway.
    • Yeah, no forkin' kidding. My many year olde dual 500MHz Pentium III is really starting to feel left out. Muther McRae and her sister too.

      My, oh, my... [extremefineart.com]

  • I didn't see any mention of SMP support in either article for the new chipset. Does the P4 even support SMP? What are the current MB offerings for SMP?
    • And the Lord sayeth onto Moses, 'Thou shalt not SMP with Intel Desktop processors.' You must buy their Xeon's to get supported MP. I'm pretty sure no one has released a desktop SMP P4 board, but I wouldn't put it past SiS, who is finaally starting to make decent parts, after years of being the OEM boardmaker everyone hates.
    • Currently, Intel only officially supports dual processor configurations with their P4/Xeon processors (as well as their P3-S processors, but those are not based on the P4 core). Via is planning on releasing a chipset to enable dual processor support for the Northwood (both original and B-series) and possibly the Willamette processors, and using PC2400/PC2700 DDR memory. I'm not sure what the status is on that processor, but it would finally bring back relatively-affordable Intel dual processor machines.

      Right now, I'm just eyeing out a dual T-bred setup for myself :)

    • No SMP on the P4. The Xeon line is the only line that Intel currently produces that has SMP capabilities.

      • And the Itanium line, of course...
      • And the P4/Xeon line is pretty decent. I'm currently testing:

        Dual P4/Xeon 2.2Ghz
        1GB Ram (not sure what type of mem, didn't look it up)
        72GB HD.

        I'm testing multiprocessor programming tools (OpenMP compilers and analyzers) using KCC, Intel's OpenMP Compiler, and the Assure/Guideview thread analyzer. So far, this machine rocks out!
  • AMD Thoroughbred reviewed here [overclockers.com.au]

    Also for now this should only concern the Overclocking community because of the ability of this chip to run at higher clock speeds due to the .13 process
  • I guess my 8088 laptop running at a blazing 3.77 MHz is out of date now :-)
  • This should increase the speed at which Windows XP crashes dramatically. I can probably fit 1.5 times as many crashes in during the same period of time as on my old motherboards.
  • What kind of RAM? (Score:2, Insightful)

    by Lord Ender ( 156273 )
    Buying RAM is a bit confusing now. Athlons have 200MHz FSB and 266MHz FSB. What kind of DDR should I use with each? There is DDR1600, DDR2100, DDR333, and some other stuff. Now there is this intel chip with 533MHz FSB. Is there an easy rule to remember what number after the 'DDR' coresponds with the number before the 'FSB'?

    It gets so confusing, building your own systems is becoming less appealing.
    • Athlons? first, ddr333 will work with any of the athlon offerings. but it costs a heap. ddr 2100 should be used if you have 266mhz bus, but if you only have 200mhz bus, you should just go with the old sdram, because ddr1600 doesn't help much (at all).
    • Re:What kind of RAM? (Score:3, Informative)

      by Keeper ( 56691 )
      There are several kinds of FSB's out there...most of them operate at 100 or 133mhz. What's all this business about 200 and 533mhz busses you say? It has to do with "when" the data is being sent.

      Let's look at a clock signal for a second in primative ascii form:
      ___ ___ ___
      .../ \.../ \.../ etc

      An SDR bus sends data once every mhz. The components send a bit of data around each time the signal rises.

      A DDR bus sends data when the clock signal changes. So a bit gets sent when the clock rises, and again when it falls. While the clock may only be 100mhz, you're sending 2x as much data around so the equivilant clock rate is actually 200mhz.

      A QDR bus sends data at the different edges of the clock signal. Notice that the clock signal doesn't instantly flip on and off -- there is a transitional period. So it sends a bit when the signal starts to rise, sends another when it reaches the top, sends a 3rd one when it starts to fall, and sends a 4th one when it reaches the bottom. While the clock may only be 100mhz, it's effective clock rate is really 400mhz.

      So, Intel really just moved from a 100mhz system bus to a 133mhz system bus.

      DDR1600 ~= CAS 2.5 ram running at 200mhz (100mhz * 2)
      DDR2100 ~= CAS 2.5 ram running at 266mhz (133mhz * 2)
      DDR2400 ~= CAS 2 ram running at 266mhz (133mhz * 2)
      DDR2700 ~= CAS 2.5 ram running at 333mhz (166mhz * 2)
      DDR3000 ~= CAS 2 ram running at 333mhz (166mhz * 2)

      CAS latency kind of represents the "ping" time of the ram. The lower the latency the better. The numbers after the "DDR" represent the theoretical bandwidth that can be obtained by the chip.

      I'm not up to date on markings on RDRAM memory modules or what they mean, so I can't really offer any insight into it.

      There isn't a rule for remembering which number goes with what FSB frequency. Most places that sell the stuff will list both bits of information though, so it isn't too big of a deal. I'm sure a mathematical formula could be made, but it's easier just to memorize which is which than a forumula which requires the use of a calculator.
      • Correction:

        DDR1600, DDR2100, DDR2400, DDR2700, and DDR3000 should have actually been written as PC1600, PC2100, PC2400, PC2700, and PC3000.

        Additionally, I left out a bit of info I had intended to include (ah, the joys of posting while distracted by work ;)). "DDRxxx" generally refers to the speed the ram is running at. So DDR333 would be running at 333mhz (166mhz * 2), referring to PC2700 or PC3000.
  • AMD has processors that run at 1.8ghz, so 533mhz is nothing! Oh wait, it said fsb. :-)
  • When, do you think, will we see processors with the main memory built in? I mean you may as well stuff it when you get it, it'll make it cheaper because we'll all get the same thing. Or is that just too far fetched an idea? ie. The new 2.5GHz with 4GB RAM right on the CPU. Hunh? Maybe it'll get rid of all this crap surrounding buying or trying to match RAM to...whatever!!!!
    • Well, first of all, what happens if you burn out a stick (or chip for that matter) of RAM? You're stuck replacing the whole assembly. Secondly, I'm building a PC now and am only buying enough RAM to get me off the ground. I'll add the full GB when I have the cash to do so. Third, these things would probably run extremely hot. That means we'd need to see a whole new generation of cooling devices, probably a (consumer viable) liquid cooling system.
    • I mean if you make it so cheap that it's like if your CD-ROM drive bites it, do you repair it? No, you buy a new one because it isn't worth the time and expense to repair it. That would be modular. But I guess if it does go you're stuck without a system. Maybe you could buy a six pack of processors. :-P Or one big RAM chip with everything as software, Transmeta Crusoe 2015?
      Maybe someday we'll see it. Maybe one unit for the mobo/cpu/memory/OS...
      • It's not just the expandability. There are several reasons.

        RAM generates more heat from the same package
        People will always want more RAM, so you can't reduce the pin count.
        Increasing die size will decrease yield and drive up costs.
        Chipset functionality would need to be moved into CPU as well, further adding size.
        RAM is a high-volume, very low margin commodity. CPU developers don't want to dilute their margins across mere RAM. (i.e. $400 for a P4 + $100 for 256MB RAM = a mere $500 for twice as much silicon)

        But, there is hope. Intel and AMD are adding dramatically larger L2 and L3 caches to their CPUs, so you are getting more memory for your money, although it actually provides better memory performance, and not truly a larger physical memory space.

        best regards,
    • Done. Gamecubes use a unified base of SRAM(what modern proc's use as cache) for the proc AND the vid card. Dimms will not be on the proc chip in the forseeable future purely because of size.
  • Bus speeds are catching up with Apple's CPU speed. I guess that CPU cache will become less important :-/
    • CPU cache will only become more important as CPU speeds increase. A hierarchical memory organization is intended to address the disparity between lower levels of memory and the CPU speed. Higher levels of cache run at faster speeds, until you get to the highest which generally runs at or near core speed. The goal is to access memory from cache as much as possible, and there is a lot of memory controller hardware designed to support that goal. CPU speeds still are still increasing at a much faster rate than memory speeds, and as this disparity increases, cache only becomes more important.

      best regards,
      • HELLO!!! Apple CPU speeds are NOT increasing. Not much, anyway. Thus, it won't matter if the memory is in the core, at CPU clock speed, or off core across the bus - which is rapidly approaching Apple's CPU core speed.

        I keep praying that they will just ditch moto and go with IBM. I know that IBM could up the Mhz.
        • I sympathize with you that Moto is not increasing the speeds of the PPC. They did the same with the dragonball that prompted Palm to consider the switch to an ARM architecture.

          However, you are significantly underestimating the importance of cache. Cache lowers cost by allowing the significantly larger storage memory to reside off chip. It provides a low latency access to memory to keep up with processor speeds in addition to reducing external accesses which save a very significant amount of power and reduce board noise.

          The key here is low latency. A core relies on the highest level of memory to have a latency of typically one core clock cycle. Main memory doesn't even come close to that. A memory transfer rate of 533MHz does NOT mean the access latency of the main memory is 1.88ns (1/533x10^6Hz). In fact, the access latency is actually several bus clock cycles. Since the memory bus clock is 133MHz, we are talking about a latency on the order of perhaps 25ns. When the core clock is close to 1 ns, there is a huge disparity between the core demands and the main memory capabilities. Without cache, the core would need to wait many tens of cycles per memory access. For applications that require heavy memory usage, this would have a devastating effect on your performance.

          This doesn't even begin to comment on how expensive the 533MHz QDR DRAM really is. By providing more cache, Moto is reducing the cost of the system, which inevitably impacts the end cost you must pay.

          Don't be so quick to second-guess the designers who bring you one of the better-architected PC CPUs and systems out there. We know what we're doing.

  • it appears [hothardware.com] as if the [tweak3d.net] new chipset [tweaktown.com] gives the [simhq.com] P4 a performance boost [anandtech.com] in most apps [tomshardware.com] over the previous 400MHz FSB chips

    Been spending too much time by the memepool, have we?

"So why don't you make like a tree, and get outta here." -- Biff in "Back to the Future"