A Universal Standard

As we mentioned on the front page, the Inquirer has a poll about coming up with a universal standard for measuring processors.

While we like the idea, we doubt we’ll ever see a voluntary standard for the following reasons:

Intel has a racket going on Intel is probably stunned at how successful this little exercise has been. They devalue the performance worth of a MHz, and who gets blamed for the end results? AMD! This is like shortchanging somebody and having the victim blame the cop.

Intel must have figured that the average person would never catch on to this. What had to be a surprise was the
alleged “experts” putting the onus on the victim rather than the perpetrator.

I have been absolutely amazed at this con job. Even those who at the least on the surface seem to understand what’s going on still put whatever blame on AMD as in “I don’t like this, but . . .”

Either this is blatant pandering to the MHz-blindered or they don’t really understand this, either. If they did, they’d be blaming Intel, not AMD, for this situation.

(Yes, I am calling anybody and everybody who is blaming AMD for trying to level the playing field Intel unleveled wrong on this issue. Intel devalued the standard, and AMD is just compensating for Intel’s action. If you blame AMD, Intel is playing you for a sucker, and you need to be told that. If you blame AMD,
you’re wrong on this one. If you don’t believe me, just look at all the benchmarks between the two for concrete proof a PIV doesn’t do as much per clock cycle as an Athlon for most things.)

So why would Intel want this wonderful state of affairs to end and let AMD off this meathook?

If I’m Jerry Sanders, I’m telling the Hammer design team, “I don’t care what you do, even if it actually lowers the CPU’s performance, make sure you come up with the highest stupid number possible.”

We Already Have Such A Number It’s called Spec, and that gets argued ad infinitum in certain circles. Imagine what will happen if it (or anything else) ever becomes really important.

Any standard that gets suggested will get nitpicked to death by the anal-compulsive. Some will have an agenda, some will do so for purely anal reasons.

The reality is CPU performance is a complex issue, and getting more complex all the time. There is no simple raw formula that works well all the time.

Even something like PR stills boils down to a rough average. It’s a better measurement than raw MHz is nowadays, but it’s not perfect. Neither it nor anything else can be perfect, simply because no single number can represent opposing results.

The Athlon uncontestably does certain things better than the current PIV. On the other hand, the PIV uncontestably does a few things better than the Athlon. How can you indicate both in one number?

Eventually, The Government

I really don’t see any consensus on this issue voluntarily developing.

What I could see happening sometime in the future is the government stepping in and imposing a solution for the benefit of Joe SixPack.

In America, that would probably be the Federal Trade Commission, though I suspect the European Union is likelier to get the ball rolling.

I’m sure some are kneejerking, “Keep government out of this,” but when a section of the private sector can’t govern itself, and customers get misled as a result, that’s an open invitation for the bureaucrats to expand their turf.

Probably wouldn’t happen for a few years, but if the situation doesn’t fix itself, and this industry can’t settle this among themselves, it’s a definite possibility.

Email Ed

Be the first to comment

Leave a Reply