Budget CPU Shootout: Clash of the 'rons

by Derek Wilson on 12/4/2003 10:55 AM EST
POST A COMMENT

97 Comments

Back to Article

  • arejerjejjerjre - Wednesday, January 07, 2004 - link

    Someone said that P4 EE loses to A64 3200+ HAH LOL
    Reply
  • arejerjejjerjre - Wednesday, January 07, 2004 - link

    THIS PROVES THAT INTEL IS THE BEST!!!!!!!!
    A RESPECTED FINNISH SITE HAS DONE SOME BENCHMARKING ON A64 3000-3400 AND P4 3.2GHZ NORMAL AND EE!!!!!!

    LOOK WHAT THE RESULTS ARE!!!!!
    http://www3.soneraplaza.fi/pelit/muropaketti/artik...

    Even the P4 3.2GHZ rulez over A64 3200+!!!!
    The Extreme Edition rulez quite easily over 3200+ and 3400+!!!
    Reply
  • Skandaloes - Saturday, December 13, 2003 - link

    "so while Intel is quite competitive in the mid-range and high-end segments, their value processors are inexcusably slow compared to AMD."
    What are you talking about? Intels mid-range segment, 2.4-2.8GHz P4's, offer the same performance as AMD's value segment Barton processors. You should do a comparison of the 2500+ and 2800+ Athlon XP's and the 2.4C, 2.6C, and 2.8C Intel P4's. I think you would be even more surprised.
    The problem is you're comparing the AMD processors to Celeron ones based on their ratings which were meant for P4's. A 2500+ Athlon ($90) has about the same performance as a 2.53GHz P4 ($175) or a 2.4C P4 ($165). A 2800+ ($145) will do almost as well as a 2.8C P4 ($215).
    AMD really has Intel beat in all price segment. The Value segment is completely obvious, and that's because AMD offers mid-range performance processors at value prices. The Athlon 64's control the high performance sector already, and the 3000+ should extend that advantage into the upper mid-range segment. The 3200+ costs about $20 less than a 3.2GHz P4 when configured, and the FX-51 is as fast or at least nearly as fast as the extreme edition for $200-$300 less. It's simple really. AMD offers faster processors for less in every price segment, not just the value one.

    Reply
  • arejerjejjerjre - Saturday, December 13, 2003 - link

    I am not going to post anymore its pointless you wouldn't believe me anyway! Reply
  • arejerjejjerjre - Thursday, December 11, 2003 - link

    By the way Celeron 2GHZ performed better in raw calculations (Pifast 3.3) than Pentium 4 1.8A

    Basis for claims about celerons low performance was that it was not good in raw calculations! It's better than you think!
    Reply
  • arejerjejjerjre - Thursday, December 11, 2003 - link

    You would probably have a heart attack if I could prove to you that the review is wrong! Reply
  • arejerjejjerjre - Thursday, December 11, 2003 - link

    How will I prove it? You wouldn't believe anything! You'd have to see for your self that's the problem!!

    Celeron has a different potential with different setting I meant!! And my test's used low end parts if you meant that!!!

    Noone reads magazines anymore :)

    Hard proof is hard to get! Like I said! And even if I had you on my side what does it matter? Still noone else wouldn't believe what I say!

    Yes it's true I just spitted out what I have said crappy text,bad english and all that! :)
    Review got me quite angry because the truth is far from what the article says.

    I cant stop thinking thats my problem! What's yours? If you do not even consider something to be truth you are in denial but that's not me whos in denial then!! :)
    Reply
  • JAGedlion - Wednesday, December 10, 2003 - link

    you know someone's in denial when they start to spam the boards like arejerjejjerjre. Seriously, if that isn't spamming, what is? At least the spam in my e-mail comes with pr0n, with arejerjejjerjre its just filled with the same old crap over and over.
    arejerjejjerjre, just think for a sec, I'm only asking for one second, so I am hoping that you can comply.
    Yes, some things will do better when on different platforms, that is rather obviouse... So whats your point? This is anandtech, not amdzone or w/e, and as such, they are successful if and only if they are reliable. Who buys a magazine and reads it through (real reading, not liek a comic book) knowing its all crap? Just as always, you have to learn to pick your battles. arejerjejjerjre, you started off at a disadvantage because the article strongly says the opposite of what you claim, and I grant you, that makes your job even harder. But to say they lied? Come on. Admit that they are right first, then say what you want. I will belive you saying that the celerons potential was not revealed, but realize, it was not only compared to durons. It was compared to XP's. These are full fledged chips, it makes sense that they would perform better, just like the worst viper outperforms the best tercel. Second, just stop with your proclamations about the unreliability of AMD procs, try to prove your point, not bash another's. Just the same, your idea of calling people liars is also moot. Compare the palimno to the throughough bred, in paper, not in what you could do. Performance increase may not have been much for you, btu there is no doubt it is a core designed to perform better. I'm not saying your right or wrong, but you need to work on your debate skills. Lastly, these are budget chips. The key word there is budget. 50% off the price of the chip isn't a little. Add to that the fact that amd mobo's are generaly less expensive and you see that for all of Intels capability in higher end systems, little can be said for their low end element.
    Reply
  • arejerjejjerjre - Wednesday, December 10, 2003 - link

    I only seen a duron run at 40 Celsius :) nothing else!

    Palomino 1700+ 60 C
    Thoroughbred 1700+/1800+ 50 C
    Duron 1300mhz 70 C the case was quite small! :)

    All the other cpus had big tower cases and good cooling!

    And intels stock cooling is silent too!!
    Reply
  • arejerjejjerjre - Wednesday, December 10, 2003 - link

    Justly if you got the money you should try and test it yourself that much better choice then to trust these reviews! Of Course if your afraid of what the results might be...

    Moronbasher what I said about the scores going high in 3dmark01 and not in quake 3 is completely true!! Hard to believe though I didn't believe it my self first!!!

    And again Intels stock cooling has allways been quite efficient enough my 2.4C is running at 42 degrees celsius!!!

    Hmm my friend had maxtor 80gb in serial ata! So that is the problem then but I've seen in some people say in nforce forums that they had had corruption with other harddrives not just maxtor!

    Yes VIA had had driver problems big time! But their chipset has workt greatly! And the performance difference between nforce 2 and via kt400a isnt much!
    Reply
  • MoronBasher - Tuesday, December 09, 2003 - link

    *i meant 9800 non pro Reply
  • MoronBasher - Tuesday, December 09, 2003 - link

    "3dmark scores went from about 9000 to 10000... Quake 3 the performance only improved from 210fps to 215fps"

    OMG!!! that's the funniest statement i have ever read. Think, in ut2k3 my 2.8C can go 220fps on flyby. but 52fps on botmatch on a 1280x1024 res. my 2500+, 190 fps on flyby, but guess what, i get 55 fps on botmatch both systems are using 9500Pro.

    yes, rely on 3dmark to show you how well your system performs on real world benches. Look at reviews of the 9900Non pro and the 9800XT, the difference between the card are at least 5-10 frames

    on nforce 2 issue, i used to have a hard drive that used to currupt alot on my a7n8x-x. The problem was not the board but the hard drive itself. I replaced my maxtor 160Gb to a seagate 160GB, guess what my hard drive never corrupted. and that was 6 months ago. it is a driver problem most likely, as 1 driver release for the nforce cottupts hard drives.

    and arguing about via as the better solution, bull crap. they have one of the worst driver record ever.

    on intel stock cooling being the best... yea right, if you want your proc to be nice and toasty at 60 degrees celcius!!! the exact temerature that my barton was at stock cooling. Now that i got myself a vantec aeroflow for both my barton and my 2.8C, my barton is at 29 degrees celcius, while my 2.8C is at 37 degrees celcius.
    Reply
  • Stlr22 - Tuesday, December 09, 2003 - link

    jesus, bout time this thing works! Reply
  • Stlr22 - Tuesday, December 09, 2003 - link

    Hi Reply
  • justly - Tuesday, December 09, 2003 - link

    Again, "Holes can be punched in all of your statements, and quite easily I might add".

    If your can't recognize the obvious, is pointless for anyone to waste their time trying to educate you, so believe what you want even if it can proved otherwise.
    Reply
  • arejerjejjerjre - Tuesday, December 09, 2003 - link

    And when bying Intel cpu you should keep in mind that it comes with a large heatsink and good fan those things cost a almost 60e etc. alpha some cost less some more!

    Amds cpus come with a tiny heatsink and small cooler they are not efficient enough for cpus they were made for!
    Reply
  • arejerjejjerjre - Tuesday, December 09, 2003 - link

    And I was saying that Celeron would win Duron I did't say anything about athlon!

    In your example they used a much better graphic card! If you remember I did the test's with a TI4200 Core:250 MEM:550 Ti4600 Core:300 MEM:600

    There is also one thing that if you use the orginal opengl32.dll wich comes with quake3 you'll most likely get bad results! With Amd that does not occur atleast not if you have the newer processor driver or a 3dnow patch dont quite remember and dont need to!

    Reply
  • arejerjejjerjre - Tuesday, December 09, 2003 - link

    Haven't heard that anyone ever had any faulty Northwoods thanks for telling :) I seen a 3.06ghz blown to pieces when overclocked too much! :)
    And for the Thoroughbred death sundrome I read about it on nvplanet or some other major site don't remember! (Of course they can deceive if that's not too hard for you to believe)

    My friends don't do much of adjusting settings!!! :)
    Nforce 2 is quite annoying when hardrive is being used in serial ata! It corrupts the harddrive safe to say once a week and there is nothing that could fix it friend used the latest bios and instructions from manufacturer no use!

    Because many of my friends have amd based system I have been able to run some test on how different a palomino core and thoroughbred is!
    I found out that when changed from palomino to thoroughbred 3dmark scores went from about 9000 to 10000 quite an increase on the performance I'd say!!! But when tested game performance (AGAIN) with Quake 3 the performance only improved from 210fps to 215fps. So the conclusion is that amd didn't made the new core so that it would be faster but they made it for marketing reasons because 3DMark01 was very popular testing program and the new core would give quite a boost to performance in reviews!
    Reply
  • justly - Tuesday, December 09, 2003 - link

    vinicastro, it is very possible that the 2.6GHz celeron is bandwith limited (quake is a very bandwith hungry benchmark) and the difference is less than 1.5 FPS or 1%, well within a range of error.

    Also the Athlon/Duron core uses exclusive cache while Intel uses inclusive cache design so in all actuality the duron has a total of 192KB of cache compared to the celerons 128KB.

    When comparing an overclocked celeron you have to aslo realize that the increased FSB provides more bandwidth and faster responce to memory calls, and that has a great impact on the P4 core design.
    Reply
  • justly - Tuesday, December 09, 2003 - link

    The problem is you are not using an analytical approach in your statements, take this one for example.

    "How the hell is that even possible to get 150 fps in quake 3 and with 640x480 ??????Anandtech sure knows how to cheat in benchmarks!!
    I got sometime ago a Celeron 2ghz and TI4200 and I scored 170 fps with 1024 resolution!!!!!!!(Every other option to the best grahics mode!)
    so how is it possible that they could get such bad results??!!!(AND they even had a RADEON 9800!!!) CHEATERS!!! Trying to mock Intel!!"

    You say you did this some time ago, well Tomshardware did a celeron review some time ago also and he got a score of 178.6 FPS at a resolution of 1024 x 32bit using a 2.0GHz celeron with a TI4600. This does not prove that you are correct since Tomshardware tested using "Demo001" and Anandtech has been using Demo004 for their reviews. In fact Tomshardware lowest speed Athlon in the test was the 1600+ and it scored 211.6 FPS.

    As for this comment.

    "Amds cpu really suck they dont work correctly and the life span is quite sort if you have a amd cpu dont be surprised if someday your computer wont start! ITS JUST AMD QUALITY!!!!! "

    I find this funny since I have four Amd systems in my house right now each with AT LEAST 2 years running time and all of them run fine and they range from a AMD 486 to a socket "A".

    Again you come across with inadequate detail with this statement.

    "There was a significant amount of processors wich suddenly just died!"

    And this didn't happen with the northwood or have you forgot about SNDS. Failing is one thing having the user kill it is another, so show me some proof that AMD processors die without the user causing it.

    Holes can be punched in all of your statements, and quite easily I might add.

    Finally to finish this off lets take a look at this.

    "And justly I seen too many Amd system's they hardly work at all! All of my friends possess some sort of amd rig and theres allways something not working or not working correctly"

    Well if your friends possess the same computer prowess that you display I think it is safe to use one of your own statments as an answer.

    "You'd have to fuck it up yourself for it not work!"
    Reply
  • vinicastro - Tuesday, December 09, 2003 - link

    I dont' belive in this results. Simply because in others tests the results are very diferent. Note in Quake3 test that the Celeron 2,4 is better than the Celeron 2.6.

    The Celeron 2.6 have 2x cache and 1GHz clock plus the Duron 1.6.

    In others sites tests the Celeron 2.0 is a litte bit slower than an XP 1600. With overclock performs better than XP 1600.

    Also note the huge AMD advertises in the tests pages. Something smell very bad here.
    Reply
  • arejerjejjerjre - Tuesday, December 09, 2003 - link

    Lol justly!

    It quite obvious that celeron's can do much better even with more lower end parts than that wich was used in the review!

    Compatibility issues with Intel? Never heard of them! Didn't know even that such term existed! :)
    You'd have to fuck it up yourself for it not work!

    And justly I seen too many Amd system's they hardly work at all! All of my friends possess some sort of amd rig and theres allways something not working or not working correctly!
    I think that Nforce chipset's suck! Via works much better! And That's why my mini pc Shuttle has a via chipset inside and duron 700 ;) works quite well most of the time! Shuttles heatpipe cooler is GREAT!
    Thats right I got one Amd based computer! :)
    Reply
  • Doop - Monday, December 08, 2003 - link

    Intel isn't the perfect when it comes to compatibility. I have crippling problems with the Pentium 4 denormalizing bug that forces a system lock up when I use some audio software. I simply cannot use that software anymore because the work arounds are too time consuming for efficent work flow.

    I seriously regret buying a Pentium 4.

    As for the Celeron...people should be warned that they suck bad. It's funny that people critisize the article because it doesn't compare apples to oranges.

    What I saw was the Duron kicking the Celerons arse. No need to read more into it than that.
    Reply
  • Quixfire - Monday, December 08, 2003 - link

    Looks like the Athlon XP 2200+ will be my next processor. Reply
  • MoronBasher - Monday, December 08, 2003 - link

    arejerjejjerjre, go to the corner with AMDjihad, and maybe both of you could join the special olympics. Hell, you're both overly qualified. Reply
  • justly - Sunday, December 07, 2003 - link

    It should be blatantly obvious that arejerjejjerjre demonstrates such extreme Intel bias that his comments can only be classified as witless. His presence is only beneficial as a source of humor for the rest of us. His unrelenting and repetitive posts filled with inaccuracy only prove that he is unwilling to accept anything other than his own biased opinion that Intel is superior in all aspects of processor performance even when we have seen this proven wrong time and time again. I do believe Intel has some good products but not to the point that I am blind to the truth, because so does AMD. I think it is obvious to most people that AMD has the best price to performance ratio in all areas except where Intel can compete with a 800MHz FSB and hyperthreading or specific apps (such as encoding) and that is even becomming less apparent when compared to the Athlon 64.
    If I had any problem with this article it would be that Anandtech didnt test using chipsets that are more commonly used with these processors. Since this article is meant to show processor performance not system performance I see no problem not because it showed AMD in a good light but because Anandtech gave BOTH platforms the benifit of a better performing chipset than they probably would get in real life.
    Reply
  • arejerjejjerjre - Sunday, December 07, 2003 - link

    You are quite wrong about how much the celeron has to offer at least in games I have tested it with low end parts: celeron 2ghz,abit bd7-II,some 333mhz 256mt 32e memory so if that isnt a cheap computer nothing is! (Graphic cards much worse than yours ti4200 :( ) and my celeron performed quite well! Not as well as my 1.8A P4!
    With Celeron 2ghz I got 170 fps in quake 3 and with 1.8A I got 225 fps and with better graphics than that what was used in anandtechs review!

    And the problem with comparing P4 EE with fx51 is that p4 ee is old tech and still quite the best! Amd had to use complete new structure :socket,chipset and other things to get comparable results! When Prescott arrives we will see again that Intel rulez! This same rutine has repeated it self for some time now and never has intel been defeated when they have launched their brand new processor or chipset!
    Reply
  • Wesley Fink - Sunday, December 07, 2003 - link

    #68&69 - Pentium 4 3.2GHz does indeed beat Barton 3200+, and we have stated that in our own tests. Barton 3200+ performs more on the level of a 2.8 to 3.0 Pentium 4 800FSB and the older performance rating is not very accurate compared to 800FSB Pentium 4 chips.

    However, Athlon64 3200+ is at least the equal of the P4 3.2 and in most cases it is actually faster. As we have stated in our reviews, AMD revised their Performance Rating with the A64 and it is actually a bit conservative. The top enthusiast Athlon64 FX51, which does not carry a Performance Rating, clearly outperforms everything we have tested - including the P4EE.

    This article is about the Bargain CPU's, where Duron/AthlonXP/Barton, the older technology AMD chips, clearly outperform the similarly priced Intel Celeron chips. This is simpler than it first appears and is a result of the differences in architecture. Pentium 4 requires huge bandwidth for best performance, and the Celeron can't deliver that bandwidth. AMD chips don't require the same bandwidth for top performance and do well with what the bargain chips can deliver.
    Reply
  • arejerjejjerjre - Sunday, December 07, 2003 - link

    And I didnt mean any small site articles!! Reply
  • arejerjejjerjre - Sunday, December 07, 2003 - link

    Theres this one thing I'am curious about!!
    I've seen many articles wich clearly say intel wins amd es p4c 3200 vs barton 3200+ and after a period similar articles appear and they seem to indicate a performance loss in intel based system! Results are clearly being altered!

    Reply
  • arejerjejjerjre - Sunday, December 07, 2003 - link

    Actually Its the P4 EE that wins by a mile :)

    and for this test seen here results are clearly tampered! Own experience proves it! I just don't know how someone could believe so blindly those test results! :) LOL
    If someone would send me some money or duron based computer and a celeron cpu I would perform the tests for you all and you would see that that the results are wrong!
    Reply
  • Gage8 - Saturday, December 06, 2003 - link

    grammar correction...
    (Ed seemed to have harsh words)
    Reply
  • Gage8 - Saturday, December 06, 2003 - link

    Ed over at OC didn't seem to have some harsh words over this article. However, I liked the duron/celery comparison. That was quite enlightening. Reply
  • Pumpkinierre - Saturday, December 06, 2003 - link

    Sorry #52 Derek, about that statement- must be the UV mod in my case blurring the eyes. However I have seen other articles where larger caches 'get in the way' with some apps. And I still stand by my base postulate that caches inherently increase system latency particularly for programs requiring fast spontaneous user control ie gaming. Unless the whole of the .exe program and settings can be contained in the cache (which basically voids system memory) then the cache has to be purged and refreshed with the required instructions and data at the whim of the user. The ideal for gaming is CPU and RAM running at the same speed- no cache. Large caches are perfect for predictable usage programs: Office,CAD/2D Graphics,Video encoding/streaming,server etc(hence large cache Xeons and opterons).The internal CPU registers and buffers are 32bit (4bytes)-so not much information is required but it must be the correct information. This is why the P4 L1 cache is only 8K cf. to the earlier P3 which had 16K-quicker to purge if wrong info, and the fact that it is inclusive to the L2 cache which again decreases the latency should the user decide to return to that part of the game in the next instant of play. This (in conjunction with large memory bandwidth) is why P4s feel smoother in play. The other alternative is no L2 cache and a slightly larger L1 cache (128 or 256K split data/commands is enough) explaining the Duron's longevity despite slow CPU speed and hence my ideal K8 CPU(see #51).
    CPU testing for gaming should be carried out by an operator playing the game. Demo testing involves the required sections of the game program with an input control file- all loaded into cache and all nicely predictable. In an operator driven test, the true measure of the CPU is largest MINIMUM frame rate and the true measure of the whole system is smallest differential bet. maximum and minimum frame rates. If you took these as your tests and measure you'd find your celerons (and Duron) doing well for their price.
    Reply
  • JungleMan1 - Saturday, December 06, 2003 - link

    Arejeje whatever your name is

    Desktop processors: AMD wins by a mile with Athlon FX

    Mobile: I might have to give the upper hand to Intel on this one, Pentium M is a nice chip

    Low end: HAHAHAHAHAHAH!! Celeron loses flat-out, as you can see in this article! Please, someone ban this tard!
    Reply
  • arejerjejjerjre - Saturday, December 06, 2003 - link

    Did I mention that amds heat problems are the last straw the stock cooler is terrible! Only whith water can you cool your Amd cpu enough!!

    Intel provides a very good heatsink and fan that doesnt need to be replaced only the thermal paste should be removed!

    Now I know a way you amd folks could benefit from your machine you could use it to heat the house think about it! Garbage could be but in to use!!!!!!
    Reply
  • arejerjejjerjre - Saturday, December 06, 2003 - link

    It seems that you amd folks live in the biggest denial of all time!! LOL

    desktop processors amd loses(P4 EE vs crappy name p4 wins)

    laptop processors amd loses (1,6ghz Pentium M better than 2400+)

    low end duron vs celeron amd loses(intel wins but poorly :(, but thats going to change when the newer celerons come)

    I just wonder if amd is going to survive with new factory being built :) Lately they had had so big losses that its just a miracle they have even survived the competition!! Business is business and theres no room for amd there!
    Reply
  • Shinei - Saturday, December 06, 2003 - link

    It's not the size of your stick, it's how well you use it that counts. Develop a CPU that only does 1.4GHz but processes 20 instructions per clock and it doesn't MATTER if you have a 3.2GHz chip, it just can't compete in pure computational strength. Megahertz myth hard at work. Reply
  • AnonymouseUser - Saturday, December 06, 2003 - link

    novice said: "Definitely proving once again that clockspeed doesn't really mean much and AMD's "Performance Rating System" is not just a marketing tool."

    Wow! You are only a NOVICE and figured that out! Many "Pros" still can't seem to grasp that simple concept. Makes me wonder who verified them as "Pros". Then again, maybe they've just become brain-dead from trying to figure out which Intel CPU is the faster version.

    Reply
  • CRAMITPAL - Saturday, December 06, 2003 - link

    The Intel fanboys just live in DENIAL even when their favorite hardware review sites show them hard data time and time again, that AMD's Duron/Athlon/A64/FX/Opteron are faster than Intel's best and AMD's products cost less, run cooler and are available NOW. Ya gotta wonder how long these folks can survive in DENIAL??? Reply
  • arejerjejjerjre - Saturday, December 06, 2003 - link

    MoronBasher buy a celeron machine and compare the results to anandtechs review you'll see the difference yourself! (By the way i didnt say that a celeron could ever beat anything else then amds low end!)

    I wonder why they used so high latencies in the tests? Of course because they noticed a celeron would perform poorly with that kind of settings!
    Reply
  • arejerjejjerjre - Saturday, December 06, 2003 - link

    Now I know why the internet has been so slow lately!!!!!!! Its because Intel based servers have been replaced with amds crappy systems!!!!!!! Reply
  • MoronBasher - Saturday, December 06, 2003 - link

    People, keep returning your "dead" procs back to the store, cuz guys like me get them for free. LOL!!!

    What i found amzing was the fact that anandtech used ddr400 and they also used pretty high latency timings
    Reply
  • novice - Saturday, December 06, 2003 - link

    Wow, while the socket 370 Celerons also trailed the Durons, at least they were close. The current crop is really sad, compared to the AMD products. Definitely proving once again that clockspeed doesn't really mean much and AMD's "Performance Rating System" is not just a marketing tool. Reply
  • MoronBasher - Saturday, December 06, 2003 - link

    arejerjejjerjre, you are a moron. Do you honestly believe amd chips just die? I am a computer maintenance technician and from the returns we get from stupid customers, it's not the proc that's the problem, it was the mobo... i could guarantee 99% of the time, the proc is not to blame. i have a 2500+ and a 2.8C and i have no problems with either. Errors? usually comes from drivers. or a moron like you who doesn't know how to set a computer up properly. Reply
  • DerekWilson - Friday, December 05, 2003 - link

    from #51:

    "This latency is seen in some of the tests where the Barton performs worse than lesser clocked A-XPs despite a larger L2 cache."

    The first XP chip clocked lower than the Barton is the 2200+ (1.800 GHz) which doesn't ever come close to touching the 1.833 GHz 2500+ Barton. AFAIK, all Athlon XP L2 caches (including Duron) have the same latency. I don't know this for a fact (though I highly suspect it), but it would be an easy test (just need to plug in sisoft sandra and look at cache latency).

    Honestly, the fact that there are only a very few benchmarks where the HIGHER clocked 2400+ (2 GHz) can touch the Barton shows how important large cache size is in increasing overall system performance. Even the high latency L3 cache (which is still much lower latency than main memory) on the P4 EE helps to push performance much higher than on similarly clocked P4 CPUs.

    The whole point of any ondie cache is to reduce latency between main memory and the processor. Having a large L1 L2 and even L3 cache doesn't increase latency, it decreases it overall. Without a cache, every single memory access takes a large number of cpu cycles to get to the processor, and much time is spent waiting for data.

    Things get complicated when looking between Intel and AMD. AMD has a larger low latency L1 cache, but Intel's L2 cache is lower latency than AMD's L2 cache. But I think I'll save that analysis for another day :-)

    Really the only advantage (aside from the cost savings) of cutting out a large ammount of cache is that you can more easily clock the chip higher. But that only really gives you a performance advantage if you can increase the clock enough to overcome the performance loss due to the lack of cache. The original Celerons could actually exceed this performance and that's why they were so sought after. The difference here is that P4 architecture is so much more sensitive to memory latency that we really can't hope for these kinds of performance gains.

    I don't think that even overclocking a P4Celeron to 3.6GHz would help enough to matter. But hopefully we'll find out in an upcoming article ;-)
    Reply
  • Pumpkinierre - Friday, December 05, 2003 - link

    What people forget about the original celeron 300 (no L2 cache) was that it was lousy in benchmarks but gamers loved it because of its low latency and overclockability both of which are hindered by cache addition. The same applied to K6-2(no L2 cache) vs K6-3(256Mb L2). This is why these celerons are still out there as the demo benchmarks dont reflect the true gaming experience ie spontaneous response by the user. The other requirement is raw grunt(floating point calculation), province of the K7, which explains the Duron's longevity despite low clock speed. This CPU also has a very small 64K exclusive L2 cache, ideal for low system latency (and 64Kb is the memory unit of the X-86 based systems). This latency is seen in some of the tests where the Barton performs worse than lesser clocked A-XPs despite a larger L2 cache. And as I stated above real world gaming accentuates this quality further.
    Its true that the vast gain in cpu speed cf. memory speed has required a middle man ie cache. But if that middle man doesnt have the goods then re ordering takes time. My favorite theoretical gaming CPU would be a K8 without L2 cache (perhaps with 256Kb L1 cache) which in conjunction with the on die memory controller and optimised fpu would have very low system latency.
    Reply
  • sonyboy851 - Friday, December 05, 2003 - link

    all of you AMD haters need to get a clue. 48: what nforce 2 board did you use? And AMD systems are just as stable as Intel ones. Why have IBM and Sun chosen AMD Optorons for their servers? I doubt its because they are unstable. Maybe its because Intel cant offer such a great product.

    Anyways, this is a budget review, so I shouldnt even mention that. So you think Celerons are good? Is that what your saying?
    Reply
  • CRAMITPAL - Friday, December 05, 2003 - link

    Same old shit from the criminally insane Intel fanboys. Must really bruise your ego to have your face punched in every day by online reviews showing how pathetic Intel products are! Get use to it as things are only going to get worse for Intel and it's fanboys.

    Xbit Labs is reporting Intel can't even produce EE's and that the Flame Throwing Prescotts are a disaster. Intel has so many production and design problems that they can't resolve, they may need to release Tejas as an even bigger FLAME THROWER than Prescott. And by all accounts the Xeon even with L3 is dead.

    I'd suggest anyone with a clue, buy stock in water-cooling companies and liquid nitrogen producers.
    Reply
  • arejerjejjerjre - Friday, December 05, 2003 - link

    And for those amd folk out there! You will never have the stability and flexibility what intel based machine can offer! I've seen too many amds aka too many problems,errors,etc!
    Theres allways something that doesnt work on an amd system!!! :)

    By the way if anyone hasnt noticed NFORCE 2 is the worse I've ever seen! Nothing works as it should! :) LOL!
    Reply
  • arejerjejjerjre - Friday, December 05, 2003 - link

    The celeron test i mentioned earlier was done with far worse system than i posses now so something is defienetly wrong with anandtechs benchmarking method or how they are reported!!

    They seem to have "lost" some points in making the articles!
    Reply
  • arejerjejjerjre - Friday, December 05, 2003 - link

    Most of amd processors have been flukes i dont think they can make anything else!! Like the thouroughbred!! There was a significant amount of processors wich suddenly just died!

    When you clone enough you get these kind of flukes :) LOL FOR AMD!!!!!!1
    Reply
  • arejerjejjerjre - Friday, December 05, 2003 - link

    Amds cpu really suck they dont work correctly and the life span is quite sort if you have a amd cpu dont be surprised if someday your computer wont start! ITS JUST AMD QUALITY!!!!! Reply
  • arejerjejjerjre - Friday, December 05, 2003 - link

    cheater site doesnt even work correctly!
    Reply
  • arejerjejjerjre - Friday, December 05, 2003 - link

    How the hell is that even possible to get 150 fps in quake 3 and with 640x480 ??????
    Anandtech sure knows how to cheat in benchmarks!!

    I got sometime ago a Celeron 2ghz and TI4200 and I scored 170 fps with 1024 resolution!!!!!!!(Every other option to the best grahics mode!)
    so how is it possible that they could get such bad results??!!!(AND they even had a RADEON 9800!!!) CHEATERS!!! Trying to mock Intel!!

    Now my system is Abit IC7-G,P4 2.4C (800fsb),TI4800(Gainward as was my TI4200),Kingston hyperx 3000(370mhz@400mhz),maxtor sata 120gt in INTEL RAID(It is the fastest no doubt about it!)
    and of course the greatest device of all time 56k modem!! :)

    now with that machine I score about 305 fps(it varys in range of 300-310) in quake 3 with 1024x768 and other options to best graphics!
    Reply
  • DrFreeze - Friday, December 05, 2003 - link

    Great Article!

    I would LOVE to see you add in another lesser known cheap CPU alternative though. It is the $20 Slot-T CPU Upgrade Adapter with a $37 Intel Celeron 1.4GHz 100MHz 256K CPU OEM. It is only running 1.4GHz but yet it is built on the PIII core so it is not hurt as much as the PIV is by branch mispredicts. It might be surprising at how well it performs, and then again, it is only using SDRAM so it might not be. =)

    Dr. Ffreeze

    $20 Slot-T CPU Upgrade Adapter
    http://store.yahoo.com/strattoncomputer/slot1-fcpg...

    $37 Intel Celeron 1.4GHz 100MHz 256K CPU
    http://www.ewiz.com/detail.php?p=CL1.4GFCPG&c=...
    Reply
  • BlackShrike - Friday, December 05, 2003 - link

    We can at least conclude one thing. Anyone who bought the AMD 2500+ got an awesome CPU, whether you overcloacked it or not, it has the performance of a INTEL 2.4 B, but the price of sub $90. And you know, with the extra money you saved, you can get a radeon 9700 Pro or 9800 Pro instead of a radeon 9600 pro or gefroce 5700 Ultra. For once I think I can say this definitively, THE IMPLICATIONS ARE CLEAR AMD ROCKS THE MID TO CHEAP MARKET ANYWAY YOU LOOK AT IT.

    Thank you, I just had to say that.
    Reply
  • DerekWilson - Friday, December 05, 2003 - link

    to update what I just mentioned, that prescott number would have been a little higher if we had had the 1.02 patch for halo at that point (it removed needless memory usage checks in the timedemo mode). Reply
  • skiboysteve - Friday, December 05, 2003 - link

    35 just got dick slaped

    go derek

    die 35
    Reply
  • DerekWilson - Friday, December 05, 2003 - link

    #35:

    We wanted to test the upper limit of performance on these processors, so we eliminated as many other bottlenecks in the system as possible.

    This is very useful, because it will let you know that you will absolutely not (with current high end technology) be able to acheive more than 34 fps with a celeron 2.6 under Halo at 10x7. When you start adding more bottlenecks to the system (like slower and less RAM and a budget video card) you will end up with an even lower frame rate.

    If you take a look at our article with the 256MB 9800 Pro (benched on Prescott 2.8GHz) you will see that we only hit 43.8 fps (slower than with the barton 2500+), and with the Athlon64 FX51 we were able to get 60.5 fps out of a 9800XT card. The barton hit 51.5 fps with the 9800Pro256.

    What that says to me is that if you buy a barton 2500+, you are very close to elminiating the processor as a bottleneck in Halo compared to the current fastest gaming system on the market.

    That's not bad for less than $100 if you ask me.
    Reply
  • tonyp1023 - Friday, December 05, 2003 - link

    I repair and build pc's in a small college town in upstate NY so I get a good sampling of whats out there in bargain PC land. My off the top of my head statistic is about five to one on dead Athlon/Duron systems compared to Intel Celeron P3/P4 systems. I have read and seen that AMD cpus are more prone to thermal failures than Intel and my repair data bears this out. Or cheap motherboards die quicker because of poorer QC on their assembly lines. Whatever, I won't build clones out of AMD's because I don't want them coming back to haunt me. A warranty is a terrible thing to have to honor (and I always do) but the Intel machines seldom come back to haunt me. Reply
  • LoneWolf15 - Friday, December 05, 2003 - link

    (Quoted)

    "These Sub-$100 CPUs serve as decent upgrades for aging systems (e.g. the P3-800 that is barely chugging along) when combined with a new motherboard, but they are also the heart and soul of many of today's sub-$1000 PCs that you'd find in the retail market."

    OT a little, but "barely chugging along"? A P3-800 still runs stuff pretty well with enough RAM (which is pretty darn cheap), a 7200rpm hard disk, and if you want to game a bit, a good budget 3D graphics card. Yeah, the Duron is a great chip, so is the Athlon XP 2500+, but for those locked into a case/motherboard setup, an old P3-800 can be made to do a lot of newer tricks.
    Reply
  • INTC - Friday, December 05, 2003 - link

    So who would put a "Budget" chip into a high end platform that included:

    ATI Radeon 9800Pro 256MB
    2 x 256MB DDR400 at 2-3-3-6 (frequency chosen by the BIOS)
    2 x Western Digital Special Edition Hard Disk Drive

    Even for enthusiasts, if you can afford a Radeon 9800Pro then you probably can afford a high-end CPU. Even if someone bought an Athlon XP 2500+ to overclock, they probably shelled out extra money somewhere (heatsink, power supply, high-end RAM, extra fans, etc.) in order to get a successful overclock so their system would not be considered "budget" any longer.

    Budget CPUs fall mostly into the ultra cheap integrated video all-in-1 systems sold at retail stores. eMachines and HP typically have Celeron systems on sale in every sale flyer almost every Sunday. Their cheapest systems feature Celeron processors and their AMD systems are typically several hundred dollars more.

    I'm not sure what percentage of the market on Celerons is through boxed processors or individual OEM chips but I've heard that Intel is able to sell every single Celeron chip that they are currently making. So, someone is buying them - a whole bunch of them!

    As for the Athlon XP and Durons, I'm not an expert in economics, but if demand surpasses supply then prices would increase for a product and the inverse is true which says that if supply is greater than demand then prices decrease in order to spark more demand. So if Celerons and P4s sell enough to get more than 80% of the market for CPUs, did AMD make too many Athlon XPs and Durons or is nobody bying them or a little of both? Maybe they should limit the quantity that they produce and their products would sell at a higher price premium - case and point are the Athlon FX and 64 chips - they haven't made very many of those so they can keep the prices high as long as demand is greater than supply.
    Reply
  • CRAMITPAL - Friday, December 05, 2003 - link

    Well for the non-believers... Ace's Hardware has a good review of Opteron vs. Xeon that should (but won't) end the bickering regarding server performance. Between Anand's low end CPU review, Ace's Server review, and any number of A64/EE reviews, it should be obvious where the future of PC X86 computing lies. No it's not the G5 Mac as Ace's demonstrated.

    Enjoy !
    Reply
  • JeremiahTheGreat - Friday, December 05, 2003 - link

    If AMD bothered to update its ageing MP chipset, it would indeed be! And with the rumoured dual-mp chipset from nVidia down the drain.. there's not much left in this area.

    Save for dual-opterons :)
    Reply
  • jawg - Friday, December 05, 2003 - link

    Has anyone done any research on dual Duron setups? I think that if they are so cost effective and you wanted to have a dual processor box it would be a most noble pursuit. Reply
  • AnonymouseUser - Friday, December 05, 2003 - link

    "$40 for a Celeron 1.2 (which overclocks easily to 1.6 on a 133 mhz bus). I'd be curious to see how it stacked up against the P4 Celerons."

    Why? The performance would still be abysmal at any price point. What's the point?

    "Intel can afford to charge an arm and a leg for CPUs that perform on this level, why can't AMD?"

    It's called MARKETING + ID10Ts. The world is full of uninformed consumers who are pushed to Intel by the fucking ID10T geeks (who apparently outnumber the non-ID10T geeks)...
    Reply
  • HammerFan - Friday, December 05, 2003 - link

    Support you AMDjihad? never. While I do like AMD, one has to wonder why such high-performing CPUs are priced so low. Intel can afford to charge an arm and a leg for CPUs that perform on this level, why can't AMD? Reply
  • AMDjihad - Friday, December 05, 2003 - link

    see intel is being beat on all fronts zserdw you faatty pugly thing. how can you suppooot intek huh you fatty. you thing. everyone suoport me dumbas Reply
  • KGB - Thursday, December 04, 2003 - link

    /Pet's his 1.3 Tualatin Celeron box @ 1.73 :D Reply
  • srue - Thursday, December 04, 2003 - link

    #16 Kristopher:
    That's probably what I'm going to do, but it would be nice if I didn't have to.
    Reply
  • Spacecomber - Thursday, December 04, 2003 - link

    Speaking of Tualatins (#19), where are the the Tualatin Celerons? ;-) They have 256kb of cache, making them very similiar to coppermine PIIIs

    $40 for a Celeron 1.2 (which overclocks easily to 1.6 on a 133 mhz bus). I'd be curious to see how it stacked up against the P4 Celerons.

    Space
    Reply
  • HammerFan - Thursday, December 04, 2003 - link

    lol @ Kyler's comment, I'm sure he's right :D Reply
  • tfranzese - Thursday, December 04, 2003 - link

    good comparison Derek, I knew AMD would be faster but the margin was surprising to say the least. Reply
  • Kyler - Thursday, December 04, 2003 - link

    #22 here:Ack sorry guys was just testing my login from a few months again.

    My comment to #20, you're just pissed cause you wanted to show her your sprocket :p
    Reply
  • Kyler - Thursday, December 04, 2003 - link

    Reply
  • DerekWilson - Thursday, December 04, 2003 - link

    So, I know I didn't explain this, but we used 2x256MB memory modules in each system, and both the AMD and Intel systems were running in Dual Channel mode.

    In other words, The Intel CPU was supplied with plenty of memory bandwidth. There may have been some small issues with the clocks not matching, but we made everything run as fast as we could, and if it made a difference at all, it would be negligable.
    Reply
  • sprockkets - Thursday, December 04, 2003 - link

    Or in other words she bought the Compaq and not the computer I was going to build. Reply
  • sprockkets - Thursday, December 04, 2003 - link

    Without cache and fast FSB and memory the P4/selloutron are crap. I thought though that some of the bottlenecks were removed, but I guess not, a simple 1.6ghz processor kills most if not all Intel's low end all the time.

    That also pisses me off, I'm pretty sure that 2 years ago a potential customer of mine went for a 1.6 Celeron P4 series processor instead of a 1ghz P3 Tulatin. She said I'm going to pass, she of course didn't know why I was sticking in a "slower" processor.
    Reply
  • EglsFly - Thursday, December 04, 2003 - link

    Some people are blinded by clock speed... Intel knows this and will continue to design chips to sell to the unwise. It wouldn't surprise me if Intel would design a chip that clocked 5GHz, but performed like a 1GHz Pentium III. People would still buy it.

    Its time for the average joe to wake up already!
    Smell the crap Intel is shoveling...
    Reply
  • CRAMITPAL - Thursday, December 04, 2003 - link

    Another day, another reality check. God bless America and AMD !

    :>))
    Reply
  • KristopherKubicki - Thursday, December 04, 2003 - link

    #11 srue: Click "print this article" and you will get all the graphs on one page. :)

    Kristopher
    Reply
  • DAPUNISHER - Thursday, December 04, 2003 - link

    #9, That would defeat the entire purpose behind a budget CPU review. The market segment this addresses is the average consumer on a tight budget looking to get the most for their money. Enthusiasts like ourselves have different perspective than the average user. They aren't likely to ever open the case let alone upgrade just the CPU later. Also, many mainstream consumers are extending their upgrade cycle between computer purchases due to economic reasons combined with things like word processing, solitaire, E-mail and internet use not requiring more power than their P3 system they bought a couple years ago can provide.Many who would buy a AMD or Intel based budget system now might not upgrade again for 2-3years at which time the upgradability of either current platform in the budget catagory will be meaningless.

    Besides, how do you put a price on all those months they had to wait for the P4 3.2 to hit a price they could afford, or they had saved enough money to upgrade the CPU, while they could have had much better performance for those months by using the AMD? Of course the faster HDD is still a better investment for most than the better CPU IMO.
    Reply
  • DAPUNISHER - Thursday, December 04, 2003 - link

    I agree that coupling the Celery with dual channel 3200DDR and overclocking it would get it more competitive with the stock AMD CPUs in this review, but it doesn't seem it's going to give a an AMD in the same price range any trouble once it's overclocked with 3200DDR in dual channel on a nF2 board.

    As mentioned by another already, OEMs still turn out massive numbers of Celery based budget systems simply because Intel is a household name and most of the average consumers usage isn't stressful enough to make any real world performance difference that they notice. A budget system sold with a 7200rpm HDD instead of a 5400rpm model would be more likely to make a difference than the CPU for most folks.

    Anywho, neither the Duron nor Celery appeal to most enthusiasts, but it does help me sell AMD systems to clients when I can direct them to reviews like this to show how much bang for their buck in the ultra-budget catagory AMD offers :D So thanks for contributing to me making a few sells Derek :) Oh, well done on the article too, very few typos and it was fairly straight and to the point.
    Reply
  • Wesley Fink - Thursday, December 04, 2003 - link

    The overclocking story here is very important to determining true value. Pop a 2500+ in a Socket A 400 motherboard, set the FSB to 200 from the stock 166 and in most cases you have a 3200+. Even the multiplier is the same for the 2500+ and 3200+. In most cases you don't even need to up the voltage, though with some CPUs a small boost is necessary. I am hearing from readers that the recent 2500+ are not quite as good at this as those of just a few months ago, but I have had great success with every 2500+ I have used.

    3200+ performance for $86 is a bargain indeed. Certainly, in terms of overclocked performance, the 2500+ is in my mind the bargain chip of 2003.
    Reply
  • jkreese - Thursday, December 04, 2003 - link

    Yo Shinei...I think they wanted to remove the video card as a possible bottleneck to the test. Using the fastest video card gives a better CPU comparison. Although it would have been nice to see what a "value" video card would do with these same processors. Reply
  • srue - Thursday, December 04, 2003 - link

    Enough with the one graph per page, especially when they all say the same thing. You could have easily put 4 or more graphs on a page, but then you wouldn't have got all your banner loads. Come on people, you encourage system builders to do what's best for the customers - the least you can do is make the readers your priority as well. Reply
  • Shinei - Thursday, December 04, 2003 - link

    Um, what budget system drops in a 256Mb Radeon 9800 Pro/XT, again? The card would cost more than the rest of the system!!! ;)
    Would have been more realistic results with an FX5200, since that's what just about every OEM computer comes with these days, except for the extreme high end that come with 9800 Pro's for $2000+.
    Reply
  • jkreese - Thursday, December 04, 2003 - link

    I knew the Athlons would blow away the Celerons. I also know that, given the motherboards you used in this comparison, you have better upgrading choices with the Celeron that the Athlon. I could buy the 2.6 G Celeron and later upgrade to a 3.2 G Pentium 4 which would easily defeat all the Athlons in this comparison. You don't have many upgrading options with the Athlons. Reply
  • Oxonium - Thursday, December 04, 2003 - link

    I have been able to overclock an Celeron 2.4 to 3.4 pretty easily with stock cooling and no voltage increase. But given the performance delta between the Celeron and Athlon/Duron, I'll probably be upgrading to an Athlon (Barton) for my next budget system. I'm actally looking at the mobile Athlons so I can have a cooler system.

    About the ads, I've agree they are getting more annoying lately. There are 19 alone on the home page, including the sponsored links. Plus there are some of the words in all the articles that are linked to ads, although I find this less annoying. I understand that Anandtech is a business and needs to make money to keep operating but this is getting bad.
    Reply
  • Keeksy - Thursday, December 04, 2003 - link

    Great article. I never knew the Celeron was such a bad performer. If I had to build a new machine on budget, I'd definitely go with a Barton Athlon.

    Hey, what's with the freaking huge ads at the top of every page? Really annoying.
    Reply
  • pxc - Thursday, December 04, 2003 - link

    I'm a little happier about the Duron 1.8GHz/motherboard combo I purchased now for $60 on black Friday. :D As a bonus, it's a 1.5v core and came unlocked. With 2 pencil traces and a short wire, it's an 11.5x 333MHz FSB Duron which still performs great without the rest of the cache enabled and has headroom for overclocking.

    The Celerons are really affected by the low speed memory (@ PC2100 on the test system) more than any other processors. Those processors would still lag in this review, the Celerons would have scaled better on other chipsets that allow better memory/FSB ratios. For example, PC3200 on single channel to match the bandwidth of the 400MHz FSB (SiS and VIA) or dual channel PC2100 (SiS648).
    Reply
  • FishTankX - Thursday, December 04, 2003 - link

    ..Why does it say 'It is clear which *card* 'offers the better performance'?? Reply
  • Pandaren - Thursday, December 04, 2003 - link

    Is anyone suprised at all by these test results? THG OC'ed a Celeron to 3 GHz a few months ago and even then, the Pentium 4 2.0A beat the pants off the Celeron.

    Fact is that most consumers don't give a rats @ss if they will get half the FPS in a game they will never buy or play. For email, Microsoft Word, and AOHell, a 2.6 GHz Celeron will seem the same as a 1.6 GHz Duron to the average person.

    I am sure that Compaq and other retail companies are well aware of the current Celeron's shortcomings. I am also sure that they will pay this article no heed and that they really don't care. The Celeron is mediocre enough.

    What I'd like to see is Celeron M (Banias-512) based desktops in the budget segment. This 800 MHz wonder provides all the office firepower most people need, and probably doesn't even need a fan to keep it cool.
    Reply
  • Boonesmi - Thursday, December 04, 2003 - link

    and whats is really pitiful is that the 1.6ghz duron is faster then the P4 1.8A in most tests

    and the duron only costs $41 LOL
    Reply
  • Boonesmi - Thursday, December 04, 2003 - link

    dang!!

    celerons suck more then i though!!
    Reply
  • mattsaccount - Thursday, December 04, 2003 - link

    The message is clear! Celerons are even worse than I thought... Reply

Log in

Don't have an account? Sign up now