AMD Extends Performance Lead With New Athlon and Duron Processor – THG.RU

Introduction

It is only 6 weeks ago when two important things happened in the high-end sector of the x86 processor market. On the very same day that AMD was proudly releasing its Athlon 1100 MHz, Intel finally retracted its Pentium III 1.13 GHz 4 weeks after its release. Since this time AMD is the clear clock frequency leader, offering a product that’s 100 MHz beyond Intel’s fastest Pentium III 1000, as well as the performance leader in the vast majority of benchmarks, as our article about the Athlon 1100 was able to show.

Is Clock Speed Everything?

Considering this fact brings up the old question. Is clock speed the only thing the average customer understands? The educated reader of Tom’s Hardware will know that there’s a lot more. Although clock frequency is obviously a measure for performance within the same group of processors, it doesn’t really say anything about performance if you compare different processor designs. Theoretically, a 300 MHz processor of one design can be faster than a 1 GHz processor of another design. Still clock speed is the one thing that marketing people are just about able to comprehend and if you read the recent CNET-article about AMD’s expected release of the Athlon 1200 and Duron 800 you will notice that even the average ‘computer journalist’ doesn’t seem to know a whole lot more than that. In this write-up you find a comment stating that Duron is now available in higher speeds than Celeron. What the editor of this article doesn’t seem to be aware of is the fact that AMD’s Duron is able to outperform any Celeron at way higher clock frequencies, simply because it benefits from the better design. I wonder if people will ever learn that. Intel is still taking advantage of this ‘MHZ-trap’ in the low-end sector and Pentium 4 will see Intel using the good old clock frequency hype in the high-end sector as well.

The ‘Official Release’ Of Athlon 1200 and Duron 800

Today AMD is finally making the release of Athlon 1200 and Duron 800 official. I am putting it this way, because this time it was easy to buy either of the two processors several days, if not even weeks ahead of their official release. We journalists don’t really like that, because we are supposed to stick to a release date and consider it as rather irritating if everybody is able to buy a product before we are ‘allowed’ to write about it.

AMD’s New Reliability vs. Intel’s New Unreliability

On the other hand this situation marks a really serious shift between AMD and Intel. I don’t know how many of you can remember it, but in the past Intel had this very ‘problem’. I remember that e.g. the first Pentium MMX or the Pentium II 333 were both available in the market weeks before Intel had officially launched the processors. It is an actual sign for a good supply. The manufacturer is ramping up the product, delivers it to its tier-one customers and those guys distribute it to their customers. The more product gets shipped ahead of time, the more likely it is that a dealer is trying to make a catch by offering the new product ahead of the official schedule. You will never see that happen in case of a shortage of a new product.

Interestingly, I haven’t seen any Pentium III processor ahead of its release for more than a year. Intel’s latest high-end processors were mostly not even available after their release for quite a while. Finally AMD, which used to have the bad reputation of not being able to deliver, is now showing the signs we only knew from Intel so far. The situation has indeed shifted. A release of an AMD processor means you can buy it in your store latest by the time of its official launch. In case of the release of an Intel high-end processor you might be lucky and get one of the rare pieces weeks or months after their official launch.

AMD Processors Vs. Intel Processors – Facts and Lies

Before getting into the details of Athlon 1200 and boring you with new super-high benchmarks scores, I’d like to say a few things about the situation we are actually seeing in the processor market. Looking at the marketing of both companies, reading reader’s comments to CNET-processor articles and realizing the current market shares makes you seriously wonder how uneducated the average computer user actually is. Yes, buddies, the majority of PC-users seems utterly unable to realize the facts.

AMD Processors are significantly less expensive than Intel processors although they are at least on par in terms of performance. – FACT

Yeah, my friends, it’s a fact! Have you just bought a Pentium III 933 for almost the same money you’d have to pay for a brand new Athlon 1200 that’s way faster? Congratulations!

AMD processors are incompatible. – LIE

Not that the average guy who just heard that phrase would know what the heck ‘incompatible’ is, but it sounds really bad, doesn’t it? Well, even the people who do know that ‘incompatible’ means that a product wouldn’t work reliably with other components (which of course is bad) are wrong if they accuse AMD’s Athlon or Duron processors of it. In our labs we are testing all kinds of Athlon platforms with all kinds of different components and I can definitely say that I cannot see any difference between the compatibility of AMD products and platforms compared to the same from Intel.

Duron is 20-40% faster than Celeron at the same clock speed. – FACT

Yes, that is true. There’s two reasons for that. One; Duron is using a bus interface that has a three times higher bandwidth than Celeron’s bus interface. Duron is connected to the system via a 100 MHz interface that is double-pumped, while Celeron uses only a 66 MHz and ‘single-pumped’ interface. Two; the core of Duron is simply more advanced, which makes it even faster when comparing the two processors at the same bus frequency. Basically, Celeron is simply no competition to Duron and I hope that even CNET will finally find out about it.

Quote from my initial Duron article: “You can see that Duron is much closer to the performance of its big brothers than Celeron, because AMD is using the same bus clock for Duron as well as Athlon. Intel is using the little FSB-limitation trick to make people fall into the MHz-trap. Inexperienced users might think that a Celeron 700 is just as fast as a Pentium III 700. Both sell at the same price, but Celeron is significantly slower.”

Find more details about this topic in the following articles:

• The Celeron Killer – AMD’s New Duron Processor
• Overclocking AMD’s Thunderbird and Duron Processor

AMD Processors Vs. Intel Processors – Facts and Lies, Continued

Chipsets for AMD processors are inferior to Intel chipsets. – LIE

Yeah, sure, the earth is flat and politicians are honest … I am still amused when I see people posting the above message in news groups or as their response to articles. How many more times does Intel need to screw up their chipsets (i820, MTH, …) until you guys get the message? Right now Intel is facing its demise in the chipset market. This has a good reason. Today you can find pretty much the same features in VIA’s Ahlon/Duron chipsets that you can find in Intel’s latest offerings. Even ATA100 has been included in the latest VIA south bridge as well now. Another guy wrote something about ACPI. I am sure he doesn’t know what that is, but ACPI is supported by VIA chipsets as well and to be honest, who is using ACPI in desktops? It’s most important for notebooks. No, you don’t need it to ‘run WindowsME and 2000’. In my systems ACPI is better not interfering with anything, since it actually costs resources that I am not willing to share. Bottom line is: there may have been some problems in the first days of the Athlon chipsets, but they are solved. Incompatibilities are more a problem of the motherboard BIOS than of the chipset right now. Thus both chipset makers, Intel as well as VIA, are actually in the same situation.

Pentium III performs better than Athlon at the same clock speed in 3D-games. – FACT

Yes, this is correct. The difference isn’t huge and easily equalized by the price of the two processors, but e.g. a Pentium III 800EB scores a few percentage points higher than a SocketA Athlon 800 in Quake3 and Unreal Tournament. However, as long as you can get a much faster clocked Athlon for the same price as a Pentium III the advantage is still on AMD’s side. You will see that the Athlon 1200 is still scoring the overall best results in any 3D-game, simply because it only has to compete against a Pentium III 1000.

Athlon outperforms Pentium III in scientific and floating-point intensive applications. – FACT

This fact is well known, but it gets easily forgotten. Athlon’s FPU is way superior to the FPU of Pentium III. It is expected that even Pentium 4’s FPU can’t compete against Athlon’s. Thus every software that requires double-precision floating-point calculations will run 40-50% faster on Athlon than on Pentium III. This difference can’t even be equalized with clock speed.

Athlon requires more power and cooling than Pentium III. – FACT

It’s true and actually one of the most annoying facts about Athlon. The power dissipation of Athlon is indeed much higher than what Pentium III radiates. Thus you require a good heat sink, a good (300-400W) power supply and a well-ventilated case. This is obviously even more important if you are overclocking Athlon. Additionally, Athlon/Duron are not secured against overheating in a way that Intel processors are. If you mount the heat sink wrong and the AMD chip doesn’t get its proper cooling it can die within seconds. AMD has, actually secret, plans to incorporate a thermal diode into the die in the next revision of the processors. It’s about time indeed; because there’s quite a number of people whose Athlon had a life span of less than a minute because of a missing thermal protection.

AMD Processors Vs. Intel Processors – Facts and Lies, Continued

AMD processors have a significantly lower clock speed margin than Intel processors. – LIE

This statement used to be true in the good old days of AMD’s K6-line of processors, but today it’s actually the other way around. The best example is the new Athlon 1200. It uses a voltage of only 1.75 V, which is the same voltage used by its brothers, the Athlon 1000 and Athlon 1100. AMD doesn’t need to use ‘overclocking-tricks’ to get higher clock speeds out of their processors. Intel however tried exactly those well-known overclocking measures with the release of the Pentium III 1.13 GHz. Today we know that Intel failed with this product, but even Pentium III 1 GHz is an ‘overvoltaged’ processor already. Today the overclocking margins of Athlon are way higher than those of high-end Pentium III processors.

Intel Processors are better for overclockers than AMD processors. – LIE

I think that this comment is pretty much known to be a lie today. AMD’s processors are in fact easier to overclock, because you can unlock the multiplier with a simple graphite pencil and then use a nice Athlon overclocking-motherboard such as the Asus A7V or the Abit KT7. In case of the Duron processor you can increase the clock speed by up to 40% successfully, in case of the high-end Athlons it’s still at least 5-10%. Intel’s high-end Pentium III processors won’t stand a chance to run 10% faster and you’ve got to take the risk of out-of-spec bus-frequencies. Celerons however are still a good target for overclocking and clock speed increases of up to 33% are certainly possible.

AMD platforms are way more expensive than Intel platforms. – LIE

This one used to be true in the first month after the release of AMD’s SocketA-processors, but today Athlon-platforms with VIA’s KT133 chipset are pretty much in the same price range as comparable Intel platforms.

AMD-platforms are inferior to Intel platforms because AMD doesn’t support RDRAM. – LIE

That’s a funny one. Recently we got in touch with Kingston, asking them for DDR-SDRAM samples. We were told that Kingston doesn’t really like us, because we are writing against RDRAM, which would be the super-fast memory of the future. God bless this, unfortunately rather bad informed marketing lady who told us that, and God bless us for not consulting her superiors. RDRAM may be the memory of the future. I am not able to tell what the future will bring. However, RDRAM is NOT the memory of the PRESENT! It’s not even the memory of the close future either. Maybe space travel is our future too, but that will still not make you use a space shuttle to get to work in the morning, will it? AMD can simply not be bothered to support RDRAM right now. In fact, for AMD it would be close to a suicidal act if it came up with the same policy as Intel, making RDRAM the only memory for its next processor. So let’s get this clear. AMD could support RDRAM if it wanted to, but it’s not crazy enough to ruin its constantly improving image in the market by getting unnecessarily close to one of the most disregarded companies in the whole wide world, RambusInc.

Intel’s Pricing Not Justified

If you summarize this all up you’ll find that AMD’s processors may not be the greatest things on earth. They are suffering from the missing thermal protection and the Athlon doesn’t perform quite as well in 3D-games as Pentium III when both processors are running at the same clock speed. However, all in all the Duron and Athlon are either better or at least just as good as their Intel pedants. At the same time AMD is able to supply both at higher clock speeds, which ensures that the best performers in the high-end as well as the low-cost segments are coming from AMD and not Intel.

Now you might be aware of all that, or you may have realized it only just now. However, don’t you wonder how it can possibly be that Intel can still afford selling their processors at prices that are ridiculous compared to what you’ve got to pay for AMD parts? This is not the same situation as a debatable price difference between a Corvette and a Ferrari. Microprocessors are no status symbols! You won’t impress any woman by saying ‘I’ve got Intel inside’. As long as Intel asks almost triple the price for a Pentium III 1000 than what AMD wants for an Athlon 1000 there is something SERIOUSLY wrong my dear readers. If Intel is able to get away with that then I am right, as tough as it may sound, ‘the average computer user has got to be very stupid indeed’. Now please spare with emails saying ‘you are obviously biased against Intel’. Common sense and simple math is all it takes to follow me.

Processor Specs

There’s not much to say about the specs of AMD’s new Duron 800 and Athlon 1200 processors. Both are manufactured using the 0.18-micron process known from their predecessors and all the other specs are also identical to the next slower AMD-processor. This also means that AMD didn’t have to increase the voltage of either of the two. Athlon 1200 is using 1.75 V and Duron 800 runs at 1.6 V. This is good to know for overclockers, since it is leaving some voltage-headroom.

Benchmark Setup

Benchmarking Considerations

I guess the expectations are rather clear. The new Athlon 1200 will obviously rule all benchmarks just as its 7-weeks older brother Athlon 1100 has exercised it before. No new competitor from Intel has joined the battle yet, so that AMD’s performance bastion remains even safer than before.

I refrained from benchmarking Duron 800 at its defined clock frequency, because I have published numbers of Duron 800 already a few months ago and they obviously haven’t changed since. Please have a look at “The Celeron Killer -AMD’s New Duron Processor” and “Overclocking AMD’s Thunderbird and Duron Processor” for more details. I still did benchmark the Duron 800 sample, but only at 1000 MHz. The Duron 800 seems to be the first Duron processor that is able to reach 1 GHz reliably and that is what makes it particularly interesting. I also overclocked Athlon 1200. You might know from AMD’s data sheets, that no current Athlon and Duron processors can be adjusted to a higher clock divider (multiplier) than x12.5. Therefore I was limited to 1250 MHz as long as I stuck to the official 100 MHz bus clock of Athlon. To reach more I had to increase the bus clock and the overclocked Athlon 1200 was ‘only’ able to finish the touchy Linux kernel compilation successfully at up to 1287 MHz (103 MHz x 12.5).

If you look at the results of Duron 1000 and Athlon 1287 you might notice something very interesting. The benchmark score in the majority of the tests doesn’t increase anymore as much as expected. This is particularly obvious in case of Duron at 1 GHz. There’s only one explanation for this behavior. A clock multiplier of x9.5 seems to be the sweet spot of Duron and its 64 kB second level cache. Increasing Duron’s clock frequency to 10 times the bus clock or more makes the clock difference between bus and processor core too big for Duron’s small L2-cache to buffer the supply with code and data from main memory. Basically Duron is simply not getting any faster anymore, because the core is starving for code and data at a core-to-bus-clock ratio of beyond 9.5. The same seems to be valid for Athlon at a multiplier of beyond 12.5, which might be the reason why AMD didn’t incorporate any higher multiplier into Athlon’s core.

This brings us to the simple conclusion that AMD will need to increase the bus clock in order to get more than 950 MHz out of Duron and more than 1250 MHz out of Athlon. The next Athlon processor will probably be an Athlon 1330, running at 133 MHz bus clock x 10. This could also mean that you will require new platforms for the next batch of Athlons, since I doubt that the current ones are able to support 133 MHz EV6-bus.

Of course that’s wild speculation on my part, but it sounds pretty reasonable. It also reminds me of the fact that Intel will supply the next generation Pentium III core by the name of ‘Tualatin’ with 512 kB L2-cache. ‘Tualatin’ is supposed to run at 1.13 GHz and beyond, with a bus clock of 133 MHz.

Price – Performance

If you’ve read my article about Athlon 1100 you may remember that I introduced price-performance charts to make it a bit easier for you to see what you are actually getting for the money spent on your processor. I received a lot of response to this new system and many people suggested that I should also add the motherboard price to the price of each processor. I took this to heart so that this time the price-performance charts will be based on the prices of both, the processor plus the required motherboard.

Here’s the pricing I was able to find for all the processors and motherboards:

Sysmark 2000

While in Sysmark 2000 an overclocked Pentium III 1000 still used to be a bit faster than Athlon 1100, Athlon 1200 is finally able to destroy any doubts that AMD has got the fastest processor in this application benchmark.

The price-performance chart proves once more that Intel processors are still ridiculously overpriced.

3D-Games

You can already get used to this picture. Athlon 1200 is the untouched performance leader in Quake 3 Arena as well.

We will see this all over again as well; Intel’s price-performance index proves that their processors are too expensive, unless you think that AMD’s processors are too cheap.

3D-Games, Continued

Under Unreal Tournament Intel’s Pentium III is scoring exceptionally well, but the overclocked 1072 MHz P3 is still not quite able to reach the overclocked Athlon 1287.

Once more the price-performance ratio of AMD’s processors leaves Intel behind.

Linux Kernel Compilation

Athlon beats Pentium III even at the same clock frequency in this test, so that Athlon 1200 is able to look particularly good. Duron suffers quite a bit from its reduced L2-cache.

The Linux community is very aware of the fact that with Athlon you get a lot more processing power for your hard earned money than with Pentium III.

FPU-Performance 3D Studio Max 2

It’s an old story that Athlon really destroys Pentium III in this benchmark. It used to be the other way around in the days of AMD’s K6-family of processor.

The difference of price-performance index between Athlon and Pentium III is particularly impressive here. You get more than double the floating point processing power for your money if you chose Athlon over Pentium III.

SPECviewperf 6.1.2

I only picked one test out of the six that make SPECviewperf’s professional OpenGL benchmark. In Light-04 the difference between the processors is most obvious. Athlon looks better here thanks to its better FPU.

Once more AMD’s processors produce much better price-performance results than Intel’s CPUs.

Conclusion

It is certainly far from surprising that Athlon 1200 has extended AMD’s lead in the market of X86-processors even further. We have to salute AMD for not trying to take the lead in terms of pricing as well, although AMD would actually deserve top-dollars for the top-processor in the PC-arena. Duron 800 is even way better in terms of how much you get for your money. You were never before able to get so much computing power for that little money. Additionally AMD is still leaving plenty of room for tweakers to squeeze even more performance out of Athlon and Duron. Especially the latter can be run at 1 GHz only with a graphite pencil and the right motherboard. Right now there’s simply nothing better out there.

Although the last 13 months saw Athlon go from 500 to 1200 MHz, there is still plenty of headroom left for this processor. Soon the performance of AMD-systems will be improved by the introduction of DDR-SDRAM support. It won’t be long until AMD will introduce a new chipset that finally enables multi-processing for its processors, at the same time when Intel releases a new processor that won’t be able to do dual or multi-processing for quite a while.

Still there seems to be the need for other improvements that might not be welcomed by current SocketA-platform owners. If my speculation is right then Athlon will soon require the 133 MHz processor bus to successfully compete with Intel’s upcoming Pentium 4. This might also require new motherboards to accommodate the new processor versions.

After looking at Intel’s current pricing it is pretty obvious that only a minority can be crazy enough to sink major amounts of money in high-end Pentium III processors. It is not surprising that the majority of sold beyond-900-MHz processors are actually carrying the brand ‘AMD’. We will have to see if Intel’s next-generation Pentium 4 will be able to change this trend. I personally think that Intel’s major justification for still being reasonably successful in the processor market can only be the uneducated average PC-user, who’s simply too blind to see the writing on the wall. I wouldn’t know why else Intel would be able to get away with a pricing that’s simply appalling. I am not saying that Intel’s processors are bad at all, but the pricing is just not justifiable.