AMD In Wonderland


First, let me thank all of you who wrote AMD when I asked, and who sent me copies of their “answer.”

It sure took them long enough, but AMD finally came up with an “answer.” Does it answer the question?
Not at all.

Out Of Touch With Reality

If this message actually reflected reality in any sort of way, this would be something to get upset about.
The people who wrote this are effectively saying, “to hell with overclockers.”

But it doesn’t reflect reality, which just makes this amusing.

The reality is these “locked” processors are no
more difficult/expensive to overclock than the earlier generations of “unlocked” Athlon processors
, and arguably easier.
Connecting the bridges is a little bit of a pain, but that’s sure to improve as people tinker a little.

If AMD were really out to do what they say, they botched the operation. They certainly lost the war when mobo manufacturers
included changing the multiplier in their designs.

Indeed, the “fix” is so simple that you could argue that at least certain
parts of AMD (probably the people actually making the chips) had no intention of stopping the practice in the first place, and what was done was done for show. I cannot believe the people responsible for designing
this “protection” believed for a second this would stop anybody with more than casual interest in the idea.

Nonetheless, regardless of reality, the propaganda machine moves on.

It’s not that AMD is being particularly evil doing this (Intel’s record, for example, is as least as bad). What is noteworthy is that this is so common nowadays.

You have to wonder just what
is going through the heads of the people who do things like this. No doubt most are just following orders, and grumbling about the stupidity of this.

However, some people, and probably some really responsible people, must think they are the Masters of Reality, and whatever they say is, is.

No. Whatever is, is.

The Watered Down But Improved Big Lie

As an amateur historian, what I find a little sinister about this denial of obvious reality is its ancestry.

What the corporate world is practicing is simply a greatly watered-down version
of Joseph Goebbel’s Big Lie: Truth doesn’t really matter. If you say something often enough, and loud enough, people start believing you.

What corporations do is obviously nowhere near as bad or as evil as the actions of the Nazis, but it’s the same principle.

The problem with the Big Lie is that it often works. Say “the multiplier is locked to stop remarkers” often enough (and particularly in front of the gullible media), and people start believing it. It happened here. As we and many
others have shown, it is nonsense.

The current chips don’t even have the protection we thought they had. You change that multiplier, there’s no indication of the rated speed, the screen just shows the overclocked speed.

The New Twist: Truthful Dishonesty

Out-and-out lying isn’t used all that much anymore by the sophisticated. Instead, what you get are what I call “technical accuracies.” What a techical accuracy does is use the truth to be dishonest and mislead. The most famous example is
“I did not have sexual relations with that woman,” but Bill Clinton hardly invented it.

In this particular case, this was used on Tom Pabst. Somebody at AMD tried to use him. What AMD told him was technically accurate, but misleading. While he didn’t swallow it whole, his
comments about “AMD has yet to supply me with false information” served to lend credence to the propaganda.

The fault does not really lie with Mr. Pabst perhaps being a little too trusting and not sufficiently skeptical; it does lie with the AMD person who showed no respect for Mr. Pabst, the customers, and ultimately, the truth.

The problem with the approach is that it is ultimately self-defeating. How many lies does it take to make a liar? Do you think Mr. Pabst is going to believe AMD so readily the next time? Will any of us?

This has become such a common tactic that it promotes paranoia among the wary. Fool me once, shame on you; fool me twice, shame on me.

You see some less-than-clear wording and you start looking for loopholes. Many times, it’s a false alarm, and just poor phrasing. Many times, though, it is not.

What ever happened to “A Man’s Word Is His Bond?”

Not so long ago, honesty was important. Many corporate leaders not only preached, but actually practiced the doctrine that “we must tell the truth.” Times have changed, and not for the better.

Of course, we’ve always had dishonest people and dishonest businesses, and always will. However, what may be different today is the belief that truth doesn’t matter, or truth is whatever you can get other people to believe, or even denying that there is any such thing as the truth.

If you don’t want to be truthful, objective reality is so annoying. It completely ignores lies and dishonesties; it can’t be persuaded at all.

Yes, there are objective realities. Maybe not as many as people thought 100 years ago, but still plenty. The sun rises in the east and sets in the west. Cyanide is bad for you. You can overclock Durons and Thunderbirds. No verbal skills can change any of that, just as none of Mr. Goebbel’s propaganda could
change the Russian armies taking Berlin and driving him to suicide.

Playing Ostrich

The AMD response is not so much lying, or being technically accurate but dishonest, but rather refusing to acknowledge (what seems to be in some eyes an unpleasant) reality. All we asked AMD to do was to clear the air on announced features of their CPU. We didn’t design the chip with the feature; they did. We just asked if they had changed it, and they played ostrich.

Sticking your head in the sand or anyplace else is a useless gesture, and as you’ll see, it was here.

Let’s Get To It

First, here’s the model letter:

Dear AMD,

I am very interested in your new Socket A processors, and had been planning on buying one of them in the near future.

Unfortunately, there have been contradictory messages from what looks to be your public relations division which make it unlikely I will do so unless a certain matter is cleared up.

I understand that the Thunderbird and Duron processors are designed to begin and display the rated speed to discourage remarking. It was also my understanding from your datasheets that after this initial period, the multiplier of the CPU could be changed by motherboards with circuitry to enable this.

There have been several socket A motherboards announced, such as the Asus A7V and the Abit KT7, which claim to have this feature.

Recent comments sent to computer hardware websites such as could lead one to believe that you have made a last-minute change which would render this feature inoperable.

However, the comments are so vague and contradictory that they may only represent a restatement of the original AMD policy: a default initial multiplier lock that could be overridden by hardware such as a motherboard.

This is a feature I find important given the lack of effective FSB range in Athlon motherboards. This will determine whether I buy from you or your major competitor, who, while they also have a multiplier lock, also have chipsets which allow for a much broader range of FSB speeds.

So my question to you is:

Can someone with a motherboard such as the ones mentioned above change the multiplier settings on Durons and Thunderbirds after initial start-up. Has AMD revised, or will AMD revise the CPUs to make this impossible?

My purchasing decision awaits your answer.



The letter clearly asks for clarification on a specific point.

Here’s the response. Judge for yourself if the question were answered.

Comments in bold italics: AMD. Comments in plain print: mine

Thank you for contacting AMD. AMD places a very high value on its reputation as a supplier of PC processors. Our reputation for quality and reliability rests in part on operation of our products within the specified range of performance parameters for each product.

Interesting comment from a company that not too long ago was saying, “Run them ’til they break.” Even more interesting from a company that admits it cranked up K6-2s until they did break. Notice this line includes overclockers as well as remarkers.

AMD is especially concerned that consumers, who purchase or build systems based on AMD’s reputation as a supplier of high-performance PC processors, receive products that have not been altered or modified without AMD’s authorization and therefore are assured of AMD’s high level of quality and reliability.

This particularly annoys me. A number of months ago, AMD used language along these lines, and I hollered about it. Back then, an AMD PR honcho essentially said that AMD didn’t care about people overclocking chips (outside of warrantees, which is legit.) Either they changed their tune, or weren’t being too truthful back then.)

Either computer systems or data may be damaged as a result of unauthorized alteration of AMD processors such as alteration of AMD products to achieve higher clock speeds than specified (i.e., “overclocking”).

Notice there isn’t a single specific reference to remarkers in this? If that had been AMD’s real concern, don’t you think they might have mentioned them as the big reason why they did that? Sure, remarkers can be lumped in together with overclockers, but doesn’t equating remarkers with overclockers say something about AMD’s attitude?

Additionally, AMD’s limited warranty specifically excludes unauthorized alterations among other actions and does not cover damages due to external causes such as improper use or operation outside of the data sheet specifications for the product, abuse, negligence, improper installation, accident, loss or damage in transit, or unauthorized repair or alteration by a person or entity other than AMD. (Please refer to AMD’s warranty information at

This is legitimate.

AMD’s goal is to provide our customers with the very best computing experience possible.

Except if you like to overclock, of course. Being untrustworthy doesn’t exactly enhance the old computing experience, either.

AMD contributes to the achievement of this goal by providing high-performance, high-quality, reliable PC processors. To support the performance, quality, and reliability of our PC processors, AMD is committed to supporting our processors when they are configured to maintain the specified operating range that is authorized by AMD.

Ah, ha, that’s what’s wrong! We overclockers just don’t have the right goals. 🙂

We appreciate your understanding, cooperation, and your recognition that AMD does not support the unauthorized alteration of AMD products.

Hmmm. That’s the result of a couple weeks’ work? Notice they refused to answer the question? Notice they didn’t even mention the multiplier lock? I asked you to ask AMD about a documented feature of their processors, and they wouldn’t answer.

Again, it’s not like Intel looks like archangels in comparison; they certainly aren’t. Just recently, they denied problems with the 820 boards and caused people all kinds of misery until they were apparently faced with legal action.

You want to say AMD still has a better record than Intel, you’re probably right, but is it a lack of evil or just a lack of opportunity? You can say one is less of a liar than the other, but they’re both still liars.

The title of the piece I wrote a couple weeks back “No Better Than Intel” says it all. You have to buy something. You buy whatever is best for you, but you verify first.

Don’t get mad, don’t write them. We got what we want. Let them live in their own fantasy world.

Just beware, and be skeptical.

Email Ed

Be the first to comment

Leave a Reply