Black Viper
Black Viper
May 122004
 

Introduction

After originally posting my AMD and Intel Rant on July 30, 2002, many items touched on have become extremely out of date. I decided to take a “second look” at what both AMD and Intel has accomplished recently and hopefully provide a vision for the future.

This Rant, as with the other ones, may contain some rather random and fragmented thoughts, and many of the “facts” are my observations and beliefs. Do not be discouraged. Read on.

Why no AMD?

Some people have asked me, “Even though you have a few computers, why are none of them AMD based?!?” Previously, I stated that “I do not like AMD.” The reasons behind that statement is outlined in my previous Rant. Before posting my view on AMD and since that time, I have watched AMD closely as to the products they are currently providing as well as what direction Intel is heading.

Each of the points I noted almost two years ago will be discussed further and I will reevaluate my feelings with currently available information.

Stability

Before July 30, 2002, I observed numerous E-Mails from users of AMD products. Almost all of them stemmed from “stability issues.” For example, an underpowered power supply or “cheap” components. No one can get anything done if a computer crashes all of the time.

Since that time, both AMD, VIA, nVidia and Intel chipsets have became much more stable and I feel computer manufactures are now taking steps to provide better quality components to reduce support costs. The problems that I read in my E-Mail daily no longer sound as much like it stems from a poor chipsets choice anymore. In fact, almost all of them are caused by virus ridden systems and “spyware” of some description. However, that is a topic all on its own.

Power Use and Heat

As I outlined before, AMD has a history of high power usage and heat dissipation problems, which in turn contribute to stability issues. However, the tables have turned drastically since July of 2002.

Intel’s release of the Pentium 4 Extreme Edition set a new record for power consumption and heat generation. Yes, it does perform well, but at what cost? According to this document: , the AMD 64 FX-53 2.4 GHz hits around 89W of power dissipation. In contrast,
Intel’s documentation on the P4 3.4 GHz EE, page 69: lists 102.9W. That is a pretty large difference when considering how small each of the chips are. However, in that same document on page 69, Intel’s “normal” P4 3.4 GHz with only 512 KB cache is comparable with AMD’s CPU at 89W.

Cost

Cost of a new computer purchase has been a large factor in the decision making process of what components are included and will remain that way for the foreseeable future. Only a small percentage of people can afford (or are willing to fork out) $4000 for the latest computer components. Many people have E-mailed me stating the cost vs. performance factor while attempting to reverse my opinion. Even though AMD’s cost vs. performance ratio has been good in the past, the AMD 64 changed that upon release. Limited production due to a new fabrication process of the 64-bit CPU’s increased AMD’s cost and therefore increased the price for the end user.

As of this writing (12MAY2004), AMD’s FX-53 CPU at 2.4 GHz is listed on newegg.com at $765 in a retail box format. In contrast, the Intel P4 3.2 GHz EE is $915. Yes, Intel’s offering is much more money and all things equal (both CPU’s mentioned perform well against each other), AMD should be considered as a viable option. But there is more…

The AMD 64 3400+ is listed at $416 and Intel P4 3.4 GHz is $415. One dollar is relatively insignificant when it comes down to total system cost. However, with this information on hand, it reduces the “cost vs. performance” argument to a moot point.

As a side note: I have no affiliation with newegg.com, but have purchased several products from them and been completely satisfied in each case.

Performance

Almost all current benchmarks available on various technical web sites and reviews of products place AMD and Intel neck-and-neck. AMD is better at some things, Intel is better at others. This is how it should be.

GHz is not everything. That is very true. Previously the single most important reason I have not chosen AMD to power my latest gaming system is because of their “PR Rating.”

“But it looks like it can go 300 miles per hour!”

With the Athlon XP, the “blah+” rating was marketing hype to attempt to compete with the GHz battle. Think about this: If a car company placed “V8+” on a four-banger, even though it “performs like a V8,” wouldn’t that raise some eyebrows? People have E-mailed me stating that the “PR Rating” actually is not comparing it to P4’s at all, but previous architecture AMD CPU’s. Funny thing is that type of comparison is not in any QuantaSpeed white paper or FAQ. In fact, I could not find out “what” the actual meaning of the model number spec is!

Model Number Madness

In a strange twist of the plot, Intel announced it will provide a “Model Number” comparing each CPU in its own class and discontinue using clock speed as an indicator, even though it will still be listed on the box. More information is here:

In light of this information, I have to give AMD credit for at least coming up with their PR Rating in relation to some performance figure. 3200+ gives some easy indication of what you are getting. However, I was not happy with them placing the same “Model Number” on CPU’s that had slightly different configurations, such as bus speed, cache or even socket. With that in mind, a model number, I feel, should be an “at a glance” indication of what you are getting. No one should be required to dig deep into the manufactures web site to figure out exactly what they want or what they are getting.

Intel will be using a three digit number 3xx, 5xx, or 7xx, to differentiate its product line. Unfortunately, the confusion between different CPU’s, even in the same family may still exist in the future. From Intel on April 2, 2004: (Source page)

A higher number within a processor family can indicate more processor features, more of a specific processor feature, or a change in architecture. Note that in some cases, a higher number processor may potentially have more of one feature and less of another.

Further additions to the confusion factor include (again from Intel): (Source page)

Processor numbers are also not a measurement of performance.

My opinion is that the marketing geniuses on both sides of the processor fence have no clue how to sell to the “average user.” People look for “numbers” (higher is better) or “catchy phrases” like “Extreme Edition.” Metal names have also been used to describe a “better product” than a previous model, such as Platinum or Gold. Everyone should

know that Platinum is “more valuable” than Gold, so the average person would naturally assume that a “CPU Platinum” is better than “CPU Gold.” However, computers have many factors that contribute to overall performance.

I feel that AMD and Intel should both provide different “names” for their family of CPU’s. The “next generation,” whether it be a different fabrication process (130nm, 90nm, etc.), dual core, or even a virtual CPU function (Hyper-threading) should be immediately obvious to the average person. Even though HT on a CPU box does indicate the features, ones without it, a different size cache, and even different sockets are all labeled “Pentium 4.”

Currently, Intel is going to release a different core with the Pentium 4 designation and add a “number” to it. As stated above, straight from Intel, higher is not always “better” but does indicate “different.” This seems like a step in the right direction, but could obviously (to me) be tweaked a little bit better.

Even though I am not fond of AMD keeping the “Athlon” name in there 32/64-bit cross over, at least they added “FX” and even “51/53” to the CPU to indicate the different revisions.

Which ever CPU you choose, do research on each of the manufactures web sites and try to sift through the hype. It could save you some money in the long run and possibly answer some questions you may have later on.

But you can over-clock it!

I do not support over clocking any system component. Thus, this is not an argument “for” AMD, “against” Intel or vice-versa. Intel chooses to not make it “easy” to over clock their CPU’s to reduce their support costs and return/replace instances from people that do not know when “enough is enough.”

I do admit that a “different-from-norm” cooled CPU and components, as in water-cooled, is a great idea and a super geek way to go! The amount of time, effort and money to get a system up and running reliably is quite a task in itself, but obtainable, nonetheless. Oddly enough, the same person that “saves” $100 on a CPU will go and get 4 case fans, a water cooling system and pretty neon lights when they could have spent that money on a “faster” CPU and quality components to wrap around it . If your over clocking efforts fail, you will have to spend more money on replacing your fried components.

64-bits

AMD has made massive improvements in the last couple of years to their products. Taking a big gamble on 64-bit computing really spun Intel around. Intel stated recently (I cannot find the source, now) that 64-bit computing is not required on the desktop. This is the kind of statements that tickle me. Software cannot use the extra functionality until hardware is available, but hardware is difficult to justify when no software will run on it! I still love Mr. Bill Gates statement about 640KB is all the RAM that anyone will ever need. Then came Windows. 🙂 (Yes, it is a joke).

Currently, very little support is available for the 32/64-bit AMD hybrid when it comes to actually using the 64-bit capabilities. According to various information available on the internet, Service Pack 2 for Windows XP may change that. What features will be available and to what extent 64-bit computing will be supported is still very much up in the air and rather vague as to the details. However, the Athlon 64 is backward compatible with Windows XP running at 32-bit.

Do not confuse Intel’s Itanium/Itanium2 server line with AMD’s Athlon 64 or Opteron . They are very different and much more costly.

With AMD’s recent track record and their forward thinking aspect on 64-bit computing, even though little support for it is currently available, they are driving progress and AMD should be commended for it.

Choosing an Athlon 64 is difficult decision to make. Should you also look to the future of desktop applications and take the 64-bit plunge while still being able to use your existing applications? That is entirely up to you and your wallet.

Conclusion

I thank you for making it this far in my second look about AMD and Intel. I am sure that, from the comments and “facts based on opinions” that I have, I am going to still get flames from the AMD supporters, even though most of my previous arguments have been revisited. I am not trying to tell you never to get AMD. I want you to be happy. Whether you get an Apple, AMD or an overly priced SGI, I couldn’t care less. Just remember that if you ask for my opinion or recommendations on a subject, you will get it… whether or not you agree with it is totally up to you.

Has any of this “new” information changed my mind about AMD? Almost. I am not going to run out tomorrow and build up an AMD powered computer, but in the near future, I could very well do so. Even though my harsh “I do not like AMD” statement is no longer valid, both AMD and Intel are currently “on the line” and could fall either way.

Only time will tell how each will end up.

Black Viper

May 12, 2004