Re: Intel has a powerful defense in this typical lawyer lawsuit: it says its machines run at 2.2 GHz, 2.4 GHz, 2.6 GHz, etc
But, they don't run at those speeds. The fsb runs at 100mhz or 133mhz (x4). The cache runs at half that speed, some of the execution units run at twice that speed.
By your logic, Intel has already lost its case.
What will be most convincing to a jury, in my opinion, is the obvious misrepresentation of the source of the information. It's one thing for Intel (or AMD) to present some tests as performance indicators. Buyers will consider the benchmarks - and consider the source. No one was surprised that Intel's official iComp benchmark favored Intel processors of the time. It's a very different thing to present, as independent, conclusions that are secretly controlled and determined by your company.
That's pure, obvious, deliberate, incontestable, fraud.
Somebody at Intel really screwed up on this one. It caused AMD several quarters of pain, but has put Intel Incorporated at some risk. These are not the right times to be convicted of creating a secretly controlled organization to fake tests - and then get a judgement entered against you based upon such a finding of fact. Especially if the findings of fact were to include words like "in order to hide the performance failings of the P4 design."
I'm not claiming that's sure to happen, but there's a good chance of it, and the chances seem to increase drastically with each new revelation of Intel/Bapco's blatant fixing of the the "independent" benchmarks they were able to make the centerpiece of almost every chip and system review of the past few years.
Did Intel distribute samples to reviewers only if the reviewers agreed to use BAPCO benchmarks in their reviews? That's the next question that will be interesting to have answered. |
