Intel's Low End

I think pretty much everyone can agree that the Celeron is getting the snot kicked out of it by the Duron. Which leaves Intel's low-end offering pretty much where they were back in the day of the Celeron 300A (anyone else remember back that far?). And boosting the FSB speed to 100 is not going to do all that much for the Celeron: it is still going to be stuck with PC100 memory and a small on-die cache. So what's up with Intel's Low end?
First, Intel doesn't need the Celeron to be any better right now. Why? Because most of AMD's resources are going toward building Athlons, not Durons. So AMD cannot supply retailers or OEMs with anywhere near enough Durons, so a majority of low-end boxes are shipping with Celerys. As well, there are no low cost integrated chipsets for the Duron, something that is discouraging companies to build low-end systems with the AMD chip. So all Intel needs to do is keep the MHZ near the Duron level and cripple the Celeron enough so that anyone who cares about performance and wants to buy and Intel chip, has to buy the much more expensive P!!!
That is all going to change pretty soon. When the Mustang core comes around, it is going to enter the AMD low-end as the Morgan. Guess what . . . when that happens, I would bet that the lowest MHZ will be 750 and Morgan will be released to nearly 1ghz (Duron could hit 1ghz now, but then it would eat into Athlon sales). The P6 core in the Celery can just barely touch 1ghz. Once Morgan breaks that barrier, the Intel low-end is in serious trouble.
Or is it? The P6 core is due to get at least one more revision before it dies (man that core has been around forever), and that is a shrink to .13. By the time that happens, the P4 (for better or worse) will be established as the High-end Intel offering, meaning the only place for the .13P!!! will be in the low-end. Right now, the P!!! is as fast or faster than the Athlon at clockspeeds over about 800mhz (except for Intensive FPU stuff). On the .13 process, the P!!! will scale to at least 1.5GHZ, and I suspect it might go a bit higher, especially with voltage tweaks. So if Intel is smart, and positions this chip against, the Morgan, it will have pretty much an equal performer in the low-end and it will be able to produce many more such chips than AMD will.
So despite the fact that it looks like Intel's low-end is falling apart, Intel is actually doing nothing wrong. It is keeping its costs down, saturating the market, keeping its ASP up, and not introducing new technology until it needs to. We all like to hate Intel, but we should remember that it can execute and has the resources to make some amazing chips (The P6 core is a prime example).
35 answers Last reply
More about intel
  1. Do not be so sure about the P!!! hitting 1.5ghz so soon because around the same time the .13 P!!! is set to arrive the P4 is supposed to enter mainstream market and this means lots and lots of them invading our homes... Yeah, it will be much faster ( near 2ghz maybe???) but if the p4's IPC is lower than the P!!!'s, then it would be suicide to get P!!! clocked too high on the market.

    Also, I do not think that 1.5ghz cpu would be easy to produce; they had one heck of a time getting good yields over 933 mhz and couldn't produce stable cpu's over that... Yes a die shrink may help, but will it be enough? I don't think that it'll work miracles.
  2. Granted Intel has to keep the P!!! clocked way under the P4, so I do not expect a 1.5 P!!! anytime soon. But remember, P4 is supposed to clock all the way up to 2ghz (and my bet on that is by late summer 2001). All Intel has to do with the .13micron P!!! is match the clockspeed of the Morgan. Now because the Athlon is not going to be able to match the P4 clock for clock, let's just assume that when the P4-2000 comes out, AMD (and this is optimistic) counters with a 1.666mhz Palomino. According to current release schemes, the highest low-end chip would be 400mhz under the high-end, so Morgan would start at 1.2 Which means that the P!!! could be positioned at 1.2 or even 1.3 to squeeze Palomino a bit. So a 1.5 P!!! would not have to come out for quite a while, mostlikely not until Clawhammer hits the market and AMD can start ramping and matchin Intel clock for clock again. 1.5 was my estimate of a theoretical max for the P!!! core.
    Now, as to whether or not the P6 core will hit 1.5 Let's see how the core has ramped so far
    at .35 micron, the P2 hit 300
    then came a process shrink, which carried over into the P!!!
    at .25 micron, the P!!! hit 600
    then can CuMine.
    at .18 micron, the P!!! hit 1000 (1050, really with OC) and did hit 1.133, but with bugs
    So a gain of 300 and 400 respectively for each process shrink. I would assume that going to .13 will get at least 350-400mhz more out of the core, and because Intel has a lot of good engineers working for them (though you wouldn't recognize that recently, 1.133), they can probably milk it to a 500mhz gain.

    just my thoughts
  3. LONG LIVE COMPETITION!! Is all I can say. We all know the driving force that is getting us the processors we want, good 'ol fashioned competition. If AMD didn't start getting serious about making not only low-end but high-end cpu solutions we'd be stuck in the dark ages with Intel slowly releasing processors to milk the market as much as it possibly could. I'd be willing to bet that Gigahertz speeds would perhaps just now be on the horizon. One thing that TICKS me off about Intel is that their newest celeron's STILL are locked at 66Mhz fsb, what a joke! I can see why they limit the cache, that lessens the die size and cuts down costs, so theycan pass it to the consumer. But to limit the FSB is rediculous, that is why I stand beside the DURON. Only gripes I have about Duron is it runs hot, and doesn't shut off no matter how hot it gets till it has a thermal meltdown -- but hey, noone's perfect? Again I say LONG LIVE THE COMPETITION !!! And thank god AMD couldn't be bought off, cause you can bet Intel tried every dirty trick it could muster. In the end, we the consumer win !!
  4. if you look at any benchmarks on the P4 vs Tbird, the Tbird trashes it on a clock for clock basis.
    the P4 is out now but there are no mobos available yet.
    I expect that when a consumer chipset is available for them the performance will increase over preview models, but not that dramatically, AMD has pretty much said they expect their 1.2 Ghz tbird to outperform the 1.5 Ghz P4. Thats quite a brag but given the benchmarks... not outside the realm of possibility. The P4 is not a well designed processor, it was designed to hold over the market and Intel's position in it until they can release their next processors. The P4 has an extremely short term position in Intels roadmap. Do your homework and you will see for yourself. There is nothing spectacular about the P4, certainly nothing to brag about.

    life as we know it is absurd
  5. All the benchmarks out there, as far as I know, were done on pre-release samples. Pre-release sample benchmarks ARE NOT indicative of final performance. Remember the early Athlon Benchmarks? When Sharky's wrote the chip off as another AMD failure? Wait until Intel releases the final version to start commenting on benchmarks.
    However, I will agree with the fact that the P4 is not an amazingly designed chip: but I am holding off on condemning it until I see real benchmarks.
    Oh, and the P4 is going to be around for a long time. All that is going to change is the packaging. Intel is moving from a 432-pin grid to a 478-pin grid. The chip core is staying the same.
    Oh, and the Tbird beating the P4 on a clock-for-clock basis means nothing if the P4 is clocked higher. Look at the situation now, the P!!! 1000 beats the TBIRD 1000, but because AMD has the 1200 out, that doesn't matter.

  6. Just a little something I stole from Aces's (

    Intel's 0.13-micron Pentium III, codenamed Tualatin, will make its first appearance beginning in Q3 of 2001 at 1.26 GHz on the "B-Step" i815e platform. Furthermore, the desktop Tualatin variant will ship with a 256 KB L2 cache, just as the current Pentium III Coppermines do now. This is in contrast to previous reports indicating a 512 KB ondie cache for the 0.13-micron Pentium III. We can only speculate as to the motivations behind this change, but with the appearance of the 1.3 GHz Pentium 4 and therefore more aggressive pricing, it is safe to assume that Intel is more interested in moving the Pentium 4 into mainstream market segments than prolonging the life of the P6. Considering Tualatin's small die size and low production costs, as well as its established and inexpensive platform, it could represent a considerable threat to the Pentium 4's acceptance in the mainstream market. By maintaining the current 256 KB L2 cache size and restricting Tualatin to a 133 MHz FSB using standard SDRAM, Intel is able to effectively limit the competitive pressure Tualatin can exert against the Pentium 4 platform. </end theft>
  7. I am a bit quick to judge, perhaps its mostly because I side with the underdog in most situations.
    However, I guarantee you AMD will still cost less for the same level of performance.

    life as we know it is absurd
  8. Ain't that the true. That's why I love AMD so much, it's like Jerry Sanders said.
    "We offer High performance at a fair price. Our competitors offer fair performance at a high price."
    But I think having Intel in the market is also a good thing, as is having Cyrix (or Via now) and Transmeta and Motorola/IBM (PowerPC 4). The more innovative architectures there are, the better it is for the consumer and for the computer world in general. So while I will continue to buy AMD processors and to recomend them to all my friends, I am more than willing to buy from anyone else who can offer me better performance at a better price. And when you come right down to it, what more could anyone want?
  9. What more could anyone want?


    I have been playing around with PC's for many years, and until I bought a Duron 600 and FIC AZ11 mobo, I thought this was a stable world (AMD). I have had some very bad experiences with the VIA chipsets (I do not blame AMD). If you just jump over to the mobo or graphics boards, you will see a lot of problems. I don't see very many of them for the PIII. Most (if not all) are dealing with the VIA based products. What does this tell you?

    After my bad experiences with the AMD/VIA combo, I am purchasing an Intel machine. I am tired of spending hours downloading, flashing, and fidgeting to get things to work.

    It is time to stop playing with the system, and play some games.
  10. Pre-release samples? Is that p4 system I seen at best buy this weekend a pre-release system as well? What is next for Intell's marketing people, are they going to release a cpu rating like the Cyrix of old with there perfomance ratings? IE, a 366 is actually called a 450?
    Lets face it, it all comes down to marketing.

    A little bit of knowledge is a dangerous thing!
  11. Actually, that's another crazy thing about the Celeron. The cache size is half, but the die size isn't decreased because all intel does is DISABLE the cache. it's still there. Just not being used by the chip. Stranger and stranger.

    The Jolly Wizened Oaf!
  12. The only thing that makes the Celeron cheaper to produce is that they arn't clocked as fast thus higher yeilds. The die is exactly as large as the P3 and the amount of silicon is the same. The cache is mearly disabled, so there is unused silicon on that chip. That is why AMD can undercut Intel so baddly. The Duron's die is smaller that the TBird's and yeilds are higher because the clock speeds are lower. AMD is actually doing what Intel would have you believe it was doing.
  13. Eventhough not many has talked about it yet, I think one product will eventually emerge: Pentium 4 Celery! a Pentium 4 with half of its cache disabled and costing next to nothing (?????) for a 1.6Ghz part some time 2002. Combining this with some sort of Intel DDR solution, 333Mhz/400Mhz DDR SDRAM for all the overclocking pleasure. This system should serve me well in case I happened to come across one or two SSE2 optimised software at that time and wanted to see if it's just hype or actually does something before Hammer's price come down.

    Intel could still be lovely at times :)
  14. ...not to mention that you can purchase a <font color=blue>Celeron</font color=blue> bundled and strapped together, in either HP or Compaq boxes, with a monitor, printer, keyboard, etc. at many outlets like Wal-Mart, K-Mart, Staples, and the like.

    When was the last time you walked into a large retail chain and saw an AMD-compatible motherboard in a display case? (OK, maybe you ARE that lucky, but you can't in my neck of the woods). AMD products sell like a donut stand at a police station, but not to the common Joe Consumer. They are going to knowledgeable users usually as an upgrade, or systems being custom-built (personal or corporate) by such a user. Joe Consumer buys what he can see, especially when he can compare prices in front of him.

    When AMD can ramp up even more volume (isn't another Fab30 coming online?) and place <font color=red>SYSTEMS</font color=red> at Joe Consumer's eyeballs, and show him the price difference and cost savings, then you can watch Intel get nervous.

    <font color=green>Buy, build, abuse, and replace... </font color=green>
    What else is there? :wink:
  15. Trust me intel is nervous.
  16. Well, I own a computer services company, and in the low-end systems I build I use Durons.
  17. The Athlon 1.2GHz on a DDR system beats the 1.5GHz P4 RDRAM system on most benchmarks, especially business applications. Clock-for-clock comparisons have always pissed me off. I prefer buck-for-buck comparisons, and top-of-the-line comparisons, which AMD wins in both cases. With the Intel roadmap project 2GHz P4's in Q1 AMD may lose the performance crown, but it'll still win the price/performance ratio, which is what I'm looking for.
  18. I have haad alot of similar problems with Intel combo's.... I think you will find ,alot of the people here are AMD fans so will post AMD problems.... I could be wrong but the P3 systems I have at my office are no better than the Athlon's but the Athlons are cheaper....


    one of the first UK T-Bird users....
  19. One product that Intel will probably (they'd be stupid not to) produce will be the P4 based on a 133 (or better yet, a 150) MHz clock instead of the crappy 100 MHz clock. Imagine when Intel does and implements DDR SDRAM how much faster the P4 will be then, especially using a .13 micron process.

    Intel isn't the one who is scared, AMD is.

    And yes, at the moment, running unoptimized software, the Athlon chips score better in benchmarks than the P4 does. But considering just how easy it'll be to optimize new software and how everyone 'loves' Intel, in a year what new software releases WON'T be optimized for a P4?

    And when that happens, what will AMD have to fight against Intel with now that the P4s are running clock-for-clock more efficient than an Athlon using readily available SSE2 optimized software?

    - Anything can be fixed with duct tape, a swiss army knife, and WD-40. :)
  20. Palomino! Sure it doesn't have SSE2, but still....
  21. Don't get me wrong. I will salute the AMD flag as well as everyone else here. I truely believe it is a superior product. But then I also was on the OS/2 bandwagon. I am just sick and tired of the most marketed (not best) product winning. I guess I am just burned out for all the years of tweaking, upgrading, fidgeting, and what not. I truely wish that the AMD 760 platform will be rock solid. Until I see that day, I am just going to sit in the bleachers.

    I guess I have lost my faith in VIA more than AMD. I feel that they do NOT do enough testing before rushing a product to market.
  22. With rip-roarin' PIII's over 700MHz available for under $200, who cares ? Intel chips are CHEAP today. Anyone stuck with a dog is out of the loop. Like, not even observing gravity. I hear chip watchers talking about "DDR around the corner" and "P4 is not ready" or whatever. SooooWhat!!! $30 to $50 dollars separates Celery from coppermine (Whoaa, I'm getting dizzy, how will I put beans on the table with these whopping processor prices !!) For an additional $200, you might be faster than me with yer 1.2 GHz thang (whoever made it), but yer not happier. I paid WAY less for O/C BX, get WAY more out of O/C AGP, and I'll patiently wait for the next golden opportunity to pounce on the cheap speed when it is available, whether it comes from AMD, or Cyrix (Oh yeah, 'm sure), or Intel. If I'm really lucky, You won't be looking. Hee Hee Heee. Chill, I say. No need to be mean. Lots of MHz for everyone out there...
  23. Well this is an interesting subject atm isnt it. The AMD fans are just looking at the pretty pictures and making a decision. The Intel fans are reading between the lines and coming up with their own conclusions. Being someone who is sitting on the Intel side of middle (I have no plans to upgrade in the near future, and if I do it shall be for a P3 1Ghz to save changing RAM and mobo - PS does anyone know if a slot 1 BX based mobo with PC 100 RAM will fit a Ghz P3 in it? In theory it should, but there are always complications with theories about computers).

    It is clearly apparent that the P4 is not a processor for today's market. It is a processor to last Intel as long as the P6 core has. And in my honest opinion it will. It puts in a performance with today's software, which granted isnt quite as fast as Athlon, but still isnt [-peep-] by ANY means. And after Intel has a "quiet" word with developers and get them to update their code, it will tear through tomorrow's software. So I actually applaud Intel for what they've done, made a processor which moves nicely with todays software, and will fly with tomorrows. They're also finding that you can't please all of the people all of the time.....

    I personally wouldnt go near an AMD processor until they can sort out 2 things. Heat and 100% compatitbility, both software and hardware. And before the AMD zealots start saying "but everything runs fine on mine", trust me, a lot of software still doesnt like Athlons.
  24. I don't think it is the Athlons that software does not like. Again, I blame it on the chipsets. They are charged with "making it all work". There are a lot of excelent components, and based on thier hipe they should all work. But it is just isn't so.

    Testing testing testing

    That's what it is all about...
  25. So now your saying INTEL is better for price/performance? HA!

    "Rip roaring 700MHz for $200!"
    LOL! Try "Rip roaring 950MHz for $200!" (talking about AMD here folks)
    Or try this one: "Rip roaring 1.2GHz for as much as P3 1GHz!"
  26. It's cool Dr. Grizely1, I'm sure your PC is fast. My 700E cost me $170, but it does 980MHz@140MHz FSB on this O/C BX system. Never chokes, never freezes, doesn't complain about even the wierdest office, mp3, er whatever software, but that's cause it is Compatible. That 140MHz FSB gives my ancient GeForce a 93MHz AGP speed (2/3xAGP=93MHz... yah). OK, so I'm not even 1GHz (almost, but I'm so ashamed) I have to underclock to watch DVDs, though, cause the inlay card needs PCI33MHz/AGP66MHz to communicate with the GeForce.. Waa. Wouldn't do it if I didn't get AC3 (5.1) from the inlay card (berrrly nuf processing power to do it wid software decoding, dis wheazing sputter, sputter 980 MHz system). An' the Ram is doing 140 too...

    So, What's yer AGP channel cyclin' at Dr. Grizely ???
    93? 92?...88? did you say 66MHz!!!!!!!!!!

    Oh, and can you O/C that AMD chip ?? I am getting 40% (YES 40%) more MHz than I paid for (Uh, Price/Performance. I think SO !!!!!)
  27. LOL. AMD is much better for overclocking. I dont have the thunderbird 900, but they have been overclocked to 1200 with air cooling and no stability problems. Sorry.

    My 750 is overclocked to 1050MHz with air cooling.

    So, that's 1200MHz for $150. You do the math, Dr Clottan
  28. sorry- I can't just "trust you". Please name some programs- ones that people actually use please- that don't like athlons. As for the "rip roaring" pIII- that's just funny. Besides athlons are just as easily overclocked and although they may be hot, they can take it, and one can always get a bigger heatsink. I mean if heat is your reason for not getting an athlon, then you just need to work a bit harder at cooling it off. And the price of a better heatsink is nowhere near the price of a faster pIII.

    "Are you saying that I can dodge bullets?"
  29. Haha! AMD fans just looking at the pretty pictures, eh? Nobody goes with AMD unless they have done some research. If you're going with the 'pretty pictures', you're gonna get an Intel system. I love my Duron system and have had exactly zero (I counted them, twice even) problems with it. I had no problems with my K6-III before it, either. I will not say that the Coppermine isn't a good processor, either. It's the one processor by Intel I'd buy. I think it is great. But it was also very expensive when I got my Duron, so I didn't go with it. I don't need the extra power and I certainly see no need to use RDRAM. As I see it, the three things that have really been hurting Intel are it's prices, it's chipsets, and it's low-end (or lack of). If they would do some adjustments to the celeron to get it up to speed, put out a good SDRAM-based chipset, and cut thier prices some (which they have been doing lately), I would go with them.

    BTW - I don't overclock (much). My games run incredibly well at the current speed, so why would I?
  30. Yes but I dont know WHY I should have to fork out an extra 30 for a cooler because the chip I've bought was designed by paraplegic muppets. I'll just pay the extra and buy a chip that runs fine with the chip that comes in the box, has the most stable chipset known to man (BX), and isnt bought by the holier than thou hardcore tech head.
  32. Why is everyone dancing around and talking about Intel's better stability and compatibility? Has everyone forgotten the MTH, the 1.13 GHz, and that whole bios thing with P4? It's not like Intel has a clean sheet with Stability and product quality. Both Companies have their problems. And Compatibility? what the hell are you talking about? I have a first generation Athlon and and a Tbird and I haven't come across ANYTHING I can't run, and I play so many games it's not even funny. I play ones from 10 years ago, I buy on average a game a week, I use emulators, and I haven't had a single damn problem yet. Please list some software that will run on an Intel and not an Athlon, unless someone purposefully designed it that way. And please don't start with hardware incompatiblities/conflicts, cause Intel has those too. That's usually more of a chipset problem anyway. Again show me examples, cause i really don't buy that crap.

    Maybe the P4 will be some good stuff in the future, but right now it isn't any better than an Athlon, and since they're going to change the pins in 6 or so months, you'd have be an idiot to buy one now. Why not wait til there actually is some SSE2 optimized software before buying a mobo that won't support the newer chips. I'd rather go on current performance then promises about the future.

    Speaking of forking over 30 bucks for a cooler, why should should I put out 200 dollars more for a processor of the same mhz and worse performance?

    Please prove me wrong, show me an example of a software incompatiblity with an athlon, cause I totally don't buy that.
  33. What are you talking about? That's only if you have to over clock it- of course you can buy an athlon 900 for 10 bucks less than your chip and overclock it to 950 or faster easily with any half way decent heatsink. Personally I don't buy sucky heatsinks in the first place, so I really wouldn't have to shell out another 30 bucks. I mean, I understand why you're frustrated with these posts, after all no one likes to hear about the ubiquitous better deal after they spent too much on a lesser product, but your anger doesn't change the fact that a pIII 700, even O/C'd, is not a rip roarin anything. And if it was so stable you wouldn't have to underclock it to run dvds. athlon 900's are fine the way they are. The fact that you had to overclock your chip to make it worth while just goes to show how lame intel is. The majority of people don't over clock so they are left with a slow chip and less money. I mean you've basically admitted that a pIII 700 is only worth it if you over clock it. Otherwise you get gyped.

    "Are you saying that I can dodge bullets?"
  34. OK, grizely1, So your 750 really IS fast. As I said in my original post in this thread, I'd buy 'the next golden opportunity', ...'whether it comes from AMD, Cyrix,(Oh yeah, 'm sure), or Intel'

    And, 'there are plenty of MHz for everyone out there...'

    My purpose was to post a reality check to remind everyone that it's hard to get a slow computer these days, and I you actually take the advice of experts, you can really put together an incredibly powerful computer. The benchmarks are really getting up there, and it can be done for so little $, it is just plain goofy to puff up and say that mine is faster, or yours cost more than mine.

    Bottomline is that I'm really happy with the performance that I got out of this O/C BX, and, though it is currently running sub-GHz (by 20 cycles), I don't feel it on the hills. Sorry I got caught up in the AMD vs Intel rage. Ya know, maybe I could inch up the FSB another notch.. the CPU is running at 40to50C. That's just shower temp, Silicon must be able to deal with way more than that ! Sounds like another adventure brewing...
  35. Amen brother testify, TODAY the p4 with rdram platform and shakie code that give it crappy legacy software bench mark is just a bad deal. Maybe next year with the .13 structor ,non rdram platform with revised code it could kick ass what P4 ver2? Luckly for low end or what I like to call econobox buliders or upgraders there alot good choices. First of all if your building a new econobox from the ground up the duron is the best choice performance /price wise, the celron doesnt enter the picture unless your buying a god awful system from a retail outlet like a office store, if your building the system it your self or having it built for you go duron. If your upgrading a existing intel based system ,check your mobo web page and find out whats the fastest coppermine it will support and upgrade the bios. Any of the pIIIe's are a better choice, than any of the celrons (true sse1 support, mmx, and 256 full speed cache and 600,700,800e's overclock like a champ especially the new cCo manufactor PIII's).If you simple must get a celron because of cost make sure you get one of the newer 600 cCo manufactored because they over clock better. But In the LONG run the PIII e's are a better deal, even for "low end" upgrades, heck even if you have a slot 1 bx, and need a slotket, the performance boost you get will be worth it. That my $.02
Ask a new question

Read More

CPUs Celeron Intel