Jump to content
3000AD Forums
Sign in to follow this  
Guest

Intel P4 Vs. AMD Athlon T-Bird/Palomino

Recommended Posts

Guest

Here's some snippets from the research I did before buying my computer. Seen several faces around here thinking to buy a new rig, hope this helps.

A layman's term explanation that convinced me, amongst other sources, of the P4's crapiness.

"The AMD Athlon is a processor made by Advanced Micro Devices originally intended to compete with the P3. One of its main features is its EV6 derived bus protocol borrowed from the popular server processor, the Alpha. (The bus is basically the way the processor communicates with the rest of the computer.) AMD has not yet utilized one of the main strong points of this interface, but not for much longer. The EV6 bus is ideal for use in multiple processor systems. Basically it allows multiple processors to work independantly without sharing available bus bandwidth between multiple processors. Other features have been added as well, the most important being the ability for one processor to directly access the data contained in the cache memory of the other processor. AMD will be releasing its 760MP chipset very soon, which will allow dual processor Athlon systems to be possible. All current Thunderbirds are already capable of being used in a dual processor setup.

The Athlon also uses a DDR bus interface operating at 100 Mhz which due to the DDR gives an effective speed of 200Mhz. The newer Athlon 'C' processors are 133 -> 266 Mhz processors.

The Athlon basically caught and surpassed the famous 'P6' Intel design first used in the pentium pro, p2, and then p3. It did this by remedying many of the faults with the p6 design. To totally understand why the Athlon is so good, and why the p4 sucks so much you need to understand the concept of a "pipeline." Without taking several years of classes this is hard to understand, but the basic principles aren't too hard to grasp. If you are interested, read on.

First of all let me start by mentioning programming languages. Many of you have probably heard of basic, cobal, fortran, c, c++, java, etc. These are all high level languages used to write the programs we use. (Ah is written in C.) When you write a program in one of these languages it is later processed by what is known as a "compiler." The compiler's job is to translate the high level language instructions into a number of very basic instuctions the processor can understand. Some programmers even write their programs using these instructions, called "assembly language." (I've done this a few times now, it is very time consuming and very hard to debug. ) Assembly instructions are then further decoded into the binary instructions that the processor can work with. All Pc processors to date understand what is known as the x86 instruction set. Basically this is a set of instructions, like add (add), multiply (mul), etc. To expand the functionality of x86 processors manufacturers have recently been adding more instructions as a way to speed up a number of often repeated tasks within the computer. The first example was the "MMX" instructions first added in the original pentium. The next was AMD's 3d now. Then came Intel's SSE, and now with the Pentium 4, SSE2.

Even though these special instructions can greatly speed up execution of a program, they are not yet widely used. Only MMX has really begun to be used in any frequency. (3d now and SSE1 enhancements are occasionally used now.) Why, you ask, if these instructions are so great, aren't they added right away? The simple reason is that to do so would require all the compilers to be updated with newer versions and all computer programmers to be taught how to properly use them. This takes time and money. Usually when money is on the line, retraining all your programmers and buying new software and getting no immediate profit boosts results in no real interest from the software companies. (If people have to buy their products anyway, why should they spend money to make their programs faster?) Eventually though these instructions do come into use, but that can take several years.

This brings me to the P4. Intel is agressively touting the P4 as being future ready, largely as a result of SSE 2 intructions. It is true that if these instructions are used that performance of the P4 gets better. Intel tried this back in the days of the Pentium 1 with MMX. (You guys might remember all the ads then touting how you'll now be able to do voice dictation and play your games faster.) If history repeats itself, and it almost always does, by the time the intructions come into real use the P4 will be long obsolete! Who really cares if 3 years from now your Quake 3 game gained 5 fps, if there are now newer processors, also using the instructions that are 3 times as fast. MMX backfired horribly on Intel. They even went so far as to remove the MMX from the Intel Inside logo for the later p2s and p3s. As if this wasn't enough, AMDs next processor will also understand the SSE2 intructions. Basically what I'm saying is that when programmers finally get around to using SSE2 every processor on the market will understand them.

Now, back to the pipeline. (Remember that from WAY up in my post? ) The pipeline is the real heart of a processor. When a processor wants to do some work it needs to do 3 things: Get the next instruction from memory, figure out what the instruction is, and finally execute the instruction. The formal names for these three processes are "fetch", "decode", and "execute." The pipeline is bascially a series of stages the processor follows to do these three things.

You are probably wondering why this is important? Hang on, I'm getting there. Hopefully you have made it this far without falling asleep. At this point you are now going to learn the main reason why the p4 sucks so much compared to the Athlon. The reason is it's pipeline! (Bet you didn't see that coming.) The original pentium processor had a 5 stage pipeline. The pentium 2 and 3 use a 10 stage pipeline. The Athlon and Duron is 11 stages. The p4 has a 20 stage pipeline. Why, you ask, is that bad? Soon you will understand. Remember those three things I mentioned that a processor must do, "fetch", "decode", "execute." The first stage in the pipeline is always going to be a "fetch." The processor will get the next program instruction from memory. Many stages in the pipeline may be "fetch" as well. In the middle there will be "decode" stages. This is where the processor figures out what it is supposed to do. At the end of the pipeline there are the "execute" stages, where the processor actually does the instruction. (For example, add 2 to 5.) Here's why a longer pipeline is not better, each stage takes 1 clock cycle. This means that while an Athlon can take up to 11 clock cycles to perform a simple instruction, the P4 can take up to 20! This means that the Athlon at 1 GHz is almost as fast in doing work as a P4 at almost 2 GHz!!! If this were 100% true the P4 at 2 Ghz would be the same speed as a 1 gig p3. Fortunately processor makers have found ways to speed up the process of moving through the pipeline by a process called "branch prediction." Bascially what this does is try to guess what the program is going to do earlier in the pipeline and then skip a few stages. This does work in practice, but it also carries the potential for disaster. What do you think would happen if the branch predictor guessed wrong? Well, what happens is the processor has to start all over from the beginning, regardless of how far along the instruction was in the pipeline. The longer the pipeline, the greater the potential for misprediction. Guess which processor has the longest pipeline? Hmm, P4. Fortunately the branch predictors in processors have been much improved in the Athlon/duron and p4 series processors. Both currently "guess" correctly about 90 - 95% of the time. One of the biggest faults of the p2 and p3s was their poor branch prediction and resulting performance hit. The Athlon has an extremely advanced predictor unit compared to the p3. This accounts for much of it's performance improvement when compared to that processor. The p4s branch predictor has so far been even a bit better, but considering the pipeline is almost twice as long, it better be.

At this point I'll bet you are wondering why anyone would ever want to make the pipeline longer? That's actually very simple, clock speed. Most consumers associate raw clock speed alone as how fast a processor is. If you make the pipeline longer the processor does less per clock cycle, making it much easier to run it at a higher clock speed. If all things were ideal the pipeline would be 3 stages long, eliminating the need for the predictor all together. Unfortuately it is VERY hard to get the whole processor to work at gigahertz speeds without things breaking down. If only small pieces at a time have to work at high speed that makes it a lot easier. It also reduces power consumption.

Do you now see a problem with the P4s we currently have? You should. Compare for the moment an Athlon 1.33 Ghz and P4 at say 1.5 Ghz. From what you've just learned above tell me which processor is now going to be the obvious winner in performance tests. That's right, the Athlon.

Oh, but it gets even worse for Intel. The AMD Athlon has 3 more things going for it. One is it's far superior floating point unit. The next two are its larger instruction and data caches. The caches basically store data and instructions in very high speed memory on the CPU itself waiting to be used. These are what are known as L1, or level 1, caches. The Athlon has 64kb of data cache and 64kb of instruction cache. This is compared to 16kb in the p3 and only 8kb ,YES 8kb, in the p4. Why would intel only put 8 kb into the p4? The simple reason is cost. Nothing takes up more space on the CPU die as memory caches. The p4 die is already 4 times as large as the Athlon and prices were spinning out of control. Intel was forced to make cuts somewhere, and this is one thing they cut.

Now we come to the floating point unit. This is best explained in an example: What is the simplest way to multiply any number by 10 or divide by 10? That's easy, just move the decimal point to the left or right one position. (Same as adding a zero or removing one.) This is what a floating point unit in the CPU does, it just moves the decimal point around to do multiplication and division. The difference is that in binary moving the decmial point is like multiplying by 2, not 10. It's time for another example: What is the easiest way for a computer to multiply by 3? The answer is to move the decimal point to the right one position and then add the original number to this. The Athlon can do this MUCH faster than the p3 or p4. Basically it can do multiplication and division faster, by about 30%, than the p3 and about 33% faster than the p4. This means that for software that basically do a lot of math, like engineering software that I use, the Athlon just blows the doors off the p4. Guess what, AH does a lot of math in it's flight model calculations. This gives the Athlon an additional edge over the P4 in AH and most direct x games.

Now you understand why the Athlon tends to be overall faster than the P4. The P4 is a processor designed for one thing, clock speed. It does this at the expense of performance per clock. Only if Intel can release p4s running at double the clock speed of the Athlon will it actually outperform it. This will never happen. Intel, for all it's faults, is very smart when it comes to marketing. All they have to do is release P4s at a few hundred Mhz above the competion and all but the extremely CPU savvy people will ooh and awe over it and fork over $500

Intel has even gone one step farther this time. In an attempt to do what can only an attempt at deception, it named a set of features in the P4 "Netburst." They are actively marketing the processor as being the "center of your digital world" and being able to make your internet experience come to life. Let me be the first to tell you that "Netburst" will enhance your "Internet" about as much as forgetting to take baths will enhance your social life. "Netburst" is just Intelese for a Quad pumped 100 Mhz bus "400 Mhz" along with a few other "features" of the p4 design.

As if you need another reason at this point not to buy a P4, there is one more, and it's a big one. Intel will be changing the P4 later this year, and as a result the socket it fits in will be different. This means if you buy a P4 now you won't be able to upgrade it to a faster one in as little as 3 months. Intel only released the current P4s because their market share was rapidly being lost to AMD. In doing so they have basically screwed the people who are buying the current P4s for upgrading.

If you want to read even more in depth why the P4 is not the best processor out there, I suggest: www.emulators.com/pentium4.htm

Guys, please don't fall for clever marketing and just assume the P4 is the better processor. The truth is it's not. So far most people haven't been swayed by Intel's ads, and the P4 only has about a 1% market share in it's first quarter of release. Recently though Intel has milked out most of the uneducated consumers with lots of cash to burn. They have slashed the prices of the P4s by as much as 60%, but the AMD is still cheaper AND Faster!

All is not lost for Intel though, the p4 does have some very good features, but unfortunately it has a number of flaws that prevent it from being the performer it should be. Most of those flaws are caused by Intel cutting costs by disabling or removing parts of the processor to save money. Intel had better hurry though, in 2 months AMD will release it's next version of the Athlon core. The codename of this is "Palomino." Among it's improvements are much reduced power consumption, possibly even more cache memory, and the ability to go to higher clock speeds. The newer Kt133a and AMD 761 ddr Motherboards will be able to use these chips with no more than a simple bios flash. The Palomino core should easy allow Athlon to reach 2 Gig by the first part of next year. It already exists, but AMD is waiting to release it because it's current Athlons are more than enough to stand up to the P4. The next 6 months should see a massive ramp up in clock speeds by both Intel and AMD just like last year.

I hope those of you who have made it to the end of this post have learned something about CPUs. My hope is that you now have the ability to recognize that clock speed comparisons between 2 different types of processors are almost totally meaningless.

Anyone still want a P4?"

Also look into http://www.tech-report.com/reviews/2001q1/p4-vs-athlon/

for an AMD Athlon Vs. Intel P4 chipset tech review.

Also, Intel is releasing the P4/G2 next month. What's this? Simply put, a P4 mark2 which is hyped (hype AGAIN! all intel does) to be "improved" P4... does that mean they put some of the stuff they originally neutered from the original P4? Or is it another marketing scam? I'll wait for the reviews.

Share this post


Link to post
Share on other sites

I also saw a cpu comparison over at videogames.com (the pc hardware section), and in performance tests, the Athlon 1.2ghz beat the P4 1.4ghz, for some games.

I recently bought the Athlon, and that baby is pretty sweet. One thing I noticed, which I think is cool, is that the Althlons are still one big chip, rather than the big cartridge thingy that the P3's are. (I don't know about P4's, but I'm assumeing it's the same)

Capt. HM.

Share this post


Link to post
Share on other sites

Good post TAC! Nice explanation of why I always suggest T-Birds to customers. And just to hear another rant, what do you think of RAMbus vs. DDR?

Share this post


Link to post
Share on other sites

hehehe, Tac beat me to the emulators' P4 article. If you haven't yet read that whole article, go read and educate yourself about intel crap!!!

Yep, my 1.2ghz T-bird has been up and running over 3 weeks now. Once I get a more capable Vidcard, I will truely be in BCM heaven, as BC3K screams on that thing in SOFTWEAR mode.

Current important specs are:

Case: Antec KS188 FULL tower

Mobo: Tyan Trinity KT-A S2390B

PSU: Antec 400w

RAM: 512mb Crucial PC133 CAS2 (2x256)

CPU: AMD Athlon T-Bird 1200mhz 266fsb

I am waiting to see what the ATI R200 brings to the video graphics table, but I have a feeling I will be getting a Gforce3.

Share this post


Link to post
Share on other sites

Tac, you did your homework and then some!

I have always used the Intel cpu's, but as of late I have not been too impressed with Intel(I run a P3-933).

In terms of a gaming system, I think that you are 100% correct, go AMD!

Only plus I can plug for Intel, is in my experiences, Intel was better for Suite/productivity/Payroll&Accounting software. Might be time to look into a AMD system.

Share this post


Link to post
Share on other sites
Guest

Absolutely correct, I bought my wife a ready made system and it had a P-2, took me 2 days to replace it with an AMD, and of course ALL the machines I build are AMD, My next door neighbor built a P-4, and when he decided to build another system, I talked him into getting the AMD, the AMD system SMOKES the Intel!! And I mean SMOKES IT, he's trying to sucker a friend of his into buying the P-4 system so he can build another AMD.

I am an AMD fan, so it's preaching to the choir, but gotta love those things!!!

Share this post


Link to post
Share on other sites

...and most of all, The Tbirds cost less than p3's... and much lesser than p4's!

In fact, for my new computer, I'll be haeding toward the 850mhz Tbird.

Share this post


Link to post
Share on other sites

By the way, if anyone needs any computer parts cheap, I've got my own company and live right by alot of wholesalers. Just give me an email.

Share this post


Link to post
Share on other sites

When I get round to buying (or building, if I've got the balls) a new PC (ie when my wife let's me) it'll definately be an AMD. A T-bird 1.2GHz is cheaper than a PIII 1 Ghz and all the benchmarks I've seen are no contest. My 3 year old Dell has been upgraded almost as much as is humanly possible, the poor old soul. Anyway, what do I use my PC for anyway? Gaming, email, internet.

Oh, and I'll probaly be writing my thesis on it too (hopefully).

Share this post


Link to post
Share on other sites
Guest

RDRAM? Mwaahhahaha. You mean that $$$$ Ram which only works with $$$$$ P4's and that still ends up AMD 1.4 > Intel 1.8? If AMD ever goes into the RDRAM speed range, ill buy it definetely.. but not until its paired with something that will actually make a difference in.

64 Meg RDRAM=$30 256 PC2100 DDR=$45

MidnightGreen: You can always return it if inside a 30 day period..if you bought it from a dealer or store. I would. Get a refund, get an AMD. Shop local stores NEVER BUY from big companies like Gateway or Dell...you will be paying for the brand name and getting crappy components. Be sure you know what you want when you shop, you wont be dissapointed. And there's also the karmic rightness that if you screw up, it will be entirely your fault, not someone else screwing it up for you and getting your money for it.

"In terms of a gaming system, I think that you are 100% correct, go AMD!

Only plus I can plug for Intel, is in my experiences, Intel was better for Suite/productivity/Payroll&Accounting software. Might be time to look into a AMD system."

AMD has some stability problems in earlier processors, but they got it right on their T-bird. One plug for Intel is that they have spread their legs for good ol' Gates. P4's rant the enchanced performance of DX8 stuff hardwired into the chip (read my 1st post). Again, it seems Microsoft is giving it hard, and Intel is learning... Marketing IS THE WAY, not quality.

Share this post


Link to post
Share on other sites

Don't EVEN bring Microsoft up. That company p**ses me of daily. RAMbus is great IF you have a prog made to take advantage of it's high bandwidth. Which by the way, almost nothing does. For the same reason you don't see many companies putting out software on DVD. Though on a different note even though DDR says it transfers at twice the rate of SDRAM you really only see about a 15% increase over SDRAM. But hey it's cheap, expecially compared to RAMbus.

Share this post


Link to post
Share on other sites

Wow, Tac, that hearkens back to the days of my undergrad and graduate comp sci classes...

Just thought I'd add my $0.02 to the article for any tech-heads who're interested; if you aren't one just skip the rest of the post. BTW, the article was 100% correct on the impact of architectural decisions. Definitely gives you something to think about AMD and Intel.

quote:


At this point I'll bet you are wondering why anyone would ever want to make the pipeline longer? That's actually very simple, clock speed. Most consumers associate raw clock speed alone as how fast a processor is. If you make the pipeline longer the processor does less per clock cycle, making it much easier to run it at a higher clock speed. If all things were ideal the pipeline would be 3 stages long, eliminating the need for the predictor all together. Unfortuately it is VERY hard to get the whole processor to work at gigahertz speeds without things breaking down. If only small pieces at a time have to work at high speed that makes it a lot easier. It also reduces power consumption.

Also, it increases parallelism in the processor. Rather than having it work on 2 or three instructions at a time, each stage of the pipeline is doing something. So a 20 stage pipeline means it has parts of (up to) 20 instructions being executed in parallel simultaneously.

But there is a tradeoff: each stage of the pipeline has to move data to the next stage. (In fact, due to physical locations of pipeline stages on the chip, sometimes an entire clock cycle is needed just to move the bits from one pipeline unit to the next - a "dummy" stage.) At any rate, this movement imposes a small but constant overhead. The tradeoff occurs because the more pipeline stages you have, the shorter they (usually) get in time, plus you have to add the communication overhead for each additional stage. Beyond a certain point, adding pipeline stages yields diminishing returns because the the time to communicate data from one stage to the next starts to become significant.

Share this post


Link to post
Share on other sites
Guest

you mean its the equivalent of designing a faster car but having to put it in a longer stretch of road because otherwise it wont work, and the gas mileage is shot (P4's $$$) as well. mwahahahah. Go Intel!

Malleus: You got any heatsinks or cooling solutions under $100 that may bring my T-bird 1.4ghz to about 30C? Its currently at 45-50, thinking on o'c'ing the sucker.

"Though on a different note even though DDR says it transfers at twice the rate of SDRAM you really only see about a 15% increase over SDRAM. But hey it's cheap, expecially compared to RAMbus."

Yep, but for DDR being cheaper than SDRAM AND gettin 15% more of it.. sounds good to me. AMD Mobo's dont use the RAMbus tech (havent heard or seen one that does, if you do, let me know!). And again, programs still dont take advantage of it. Maybe Windows 2004 (prolly gonna be marketed as MS Win "EWPS" (Enchanced Windows Proccessing System)

[ 08-09-2001: Message edited by: Tac ]

Share this post


Link to post
Share on other sites

For all of you who seem all too quick to bash Intel: Intel's design philosophy simply emphasizes clock speed. For example, having advanced (read- very fast) clock speeds is just as valid a way of speeding up a cpu as having a more efficient pipeline/cache/you name it. The fact that a 1.7 Ghz P4 may be roughly equivilant 1.4 T-Bird in benchmarks does not make the T-Bird more advanced. Intel simply chose to emphasize clock speed, whereas AMD focused on streamlning their microarchitecture.

[ 08-09-2001: Message edited by: Sunanta ]

Share this post


Link to post
Share on other sites

Yeah, and also, you can't shorten the pipeline (T-bird) but you can make a higher clock speed (Intel) So think about the future possible chips from Intel and Athlon.

Share this post


Link to post
Share on other sites
Guest

aye, but castrating a chip and marketing and selling it as if it wasn't is not my idea of good bussiness. Think about other things, you see massive marketing campaigns by intel on the tv and media.. i've yet to see something from AMD of the likes. Yet AMD has a good share of the market. When I see a company getting results through quality and keeping costs down for the costumers (not themselves..unlike the P4), I will bet AMD will in the end always beat Intel... or history may repeat itself and AMD may go down because of massive marketing and market grabbing with major manufacturers (aka Dell). Its like MS vs Apple all over again. *sigh*.

Share this post


Link to post
Share on other sites

IMO technical elegance can always beat brute force. This is why the Radeon series from ATI in 32-bit color started out with lower performance than contemporaneous nVidia cards at low resolutions, but easily outperformed it at higher resolutions. And also why I'm betting on ATI for the long haul (not to say nVidia doesn't have some interesting stuff in the works too, I just favor ATI for now).

Back on topic, AMD appears to have the better pipeline; higher clock speeds tend to come naturally with time so given time, so AMD will get there too (just that Intel will likely get there first).

Higher clock speed = more heat, which tends to shorten CPU life. So given a choice between Intel and AMD of equal performance but different speeds, I choose the lower speed.

Also, higher clock speed = greater disparity between CPU and memory (be it L1 cache, L2 cache or system RAM). Shooting for higher and higher CPU clock speeds is pointless if other bottlenecks (like RAM speed) becomes significant (hence DDR). So IMHO again, AMD is taking the smarter approach by focusing on improving performance for a given clock speed.

Of course, AMD and Intel are taking divergent courses with regard to 64-bit CPU support for 32-bit applications. That will be interesting to watch play out.

Share this post


Link to post
Share on other sites

Luc, I checked and couldn't find a good cooler for you in that price range. Most of them were WAY higher, not to mention special order at that. I did find something though. I found plans for a water cooler/radiator system that could be made for a bit over a hundred. I can send you the plans if you want. They are in a microsoft word document that has been zipped down to 1.03Mb. Just give me an email if you want them. If you want me to make it for you, give me an email and I'll discuss how much I would charge for it. I'll have to check how much the parts, shipping, and labor would cost. Either way from looking at it it seems to be a good design set up specificly for AMD processors.

Share this post


Link to post
Share on other sites
Guest Hellbinder[CE]

Sunanta,

Intel is not just "choosing to emphasize" clock speed. That almost sounds like you are an Intel Marketing agent. Intel has "chosen" to run at 400 to 500 MHZ faster than AMD because the HAVE TO. An evenly clocked P4 would get ambarassingly destroyed by an Athlon. The new Athlon4 will ramp up the clock speed to intel levels. Then there will simply be no marketing manuvers left for intel. The Athlon4 is roughly 84% faster than an equally clocked Tbird. There is simply NO WAY intel can hang with that. At 1.6ghz an Athlon4 will be 20% faster than a 2ghz P4. Even their next gen P4 is just a Non Rambus (slower) solution. Unless Intel pulls a rabbit out of their hat they are in for a very painfull year.

Share this post


Link to post
Share on other sites
Guest

Also, take into account that most AMD users will be able to upgrade to the Palomino with a simple bios flash. The 761 chipset in many AMD mobos is extremely versatile.

Intel... hehe, buy the whole new system all over, no upgrades possible for current P4 owners... gone the way of the dodo.

Now, dont you just wish your old Nintendo DUCK HUNT pistol would let you *shoot* the blue guys on TV?

Share this post


Link to post
Share on other sites

Intel will still go strong. The informed will buy AMD (and are already) everyone else will buy Intel just because of the marketing. And I must say one thing for Intel. Their processors are good for business end servers at least. Any gamer worth his salt will buy an AMD though.

And if anyone wants those plans for the water cooler, just send me an email, and I'll send them on over. I set them up in Word 2000 to make it easier to deal with.

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×