News & Analysis
Editor's Notebook: Cooley Demands a Recount
Peggy Aycinena, Editor
12/19/2000 6:11 PM EST
I hate clichs as much as the rest of you — particularly over-used, trendy ones. But, the truth of the matter is that EDA ran into a problem this past week and it's related to polling practices. We need a recount, but there's no way to get one — given the circumstances, the partisan politics, the dollars, and the positioning that are at stake.
Synopsys held a press conference/webcast on Tuesday, December 5th. They were announcing Version 2.0 of Physical Compiler. The package has been beefed up from last Fall's initial offering by the addition of power, test, and datapath capabilities. Okay — nothing earth-shatteringly surprising here. Anybody could have predicted these enhancements would be here eventually.
The problem came at the point in the event when Synopsys put up the slide displaying the company's claim to overwhelming dominance in the physical design space (see Figure 1). Only two other competitors were listed on the slide — Cadence and Magma. Synopsys was shown to have 53 tape-outs, Cadence with 3, and Magma with 1. Interesting stuff and pretty impressive — if it's true. From the looks of the slide, you'd think it was time for Cadence and Magma to pick up their toys and go home, not to mention Avanti and Monterey and others.
|
Figure 1 - The Synopsys Slide
|
|
|
But the story didn't end there. Everybody's favorite EDA BadBoy, John Cooley, was in the audience and got to ask the toughest question: "Where do you guys get your stats? Can you prove 53 tape-outs? To date, I've only been able to confirm 13."
Synopsys' answer? We've been crucial in many, many tape-outs for companies who are unwilling to publicly acknowledge our involvement. We have to honor their confidentially and we can't tell you.
The press conference ended and the fallout began.
By Thursday, Cooley had sent out a survey to all of his 11,000 e-mail intimates asking for a recount. He boldly challenged anybody to tell him about a successful tape-out that could be attributed to Physical Compiler and to site alternate successes related to Cadence or Monterey or other vendors. Cooley's intending to get to the bottom of this, with or without the help of Synopsys. If Synopsys won't back up their claim, Cooley will do it for them.
That's the problem with a society awash in free speech. You better be careful, who's got the bullhorn before you declare potentially unsupportable facts.
So, let's assume that Cooley can get an accurate count. So what? What if he discovers that the number is only 43, or 33, or 13? Does it really matter? I'd say, "No." If the tool is a good one, offering true value, then the tool deserves to be the pre-eminent presence in the physical design space. In fact, even if the tool's only got one user right now — but the tool is robust and thorough and valuable — it deserves to garner respect and overwhelming market share. But, is that the case?
Here's a portion of an e-mail that I received from a very non-Cooley kind of guy that lists complaints far more critical than the winner/loser scoreboard stats.
"Synopsys is touting the number of tape-outs its customers have done using Physical Compiler and John Cooley's tape-out call to the designer community [is not a surprise]. I've been hearing a lot as a result of Synopsys' Webcast, as maybe you have. For example, it's claiming tape-out victories, which is stretching the truth a bit. Synopsys' Physical Compiler is only able to deliver a placement. That leaves the hard work of detailed routing, clock, and power for a true physical design tool or system such as those offered by Cadence, Avanti, or Magma."
This is the real problem, in my mind. Is Physical Compiler all that it should be or could be? Are there other tools out there that are just downright better? How are designers supposed to effectively evaluate their upcoming purchases of these increasingly critical physical design tools if it sounds like Synopsys has got a lock on the market AND the technology? How are designers supposed to avoid the overwhelming sense that they need to buy into the enlarged Synopsys flow just to simplify their vendor-interface issues and to avoid being at odds with the majority?
Two of the three Physical Compiler customers providing testimonial at last Tuesday's event had never even evaluated the competitions' offerings. They acknowledged that it was just easier to continue to work through their existing relationship with Synopsys to put enhanced physical design capability into their design flow than to look at other options.
Meanwhile, Cooley's recount is continuing. He's set a deadline for the final numbers by the end of this coming week and it's an interesting story — one worth following. Will Cooley succeed in having the voice of the majority heard? Or will an Electoral College nominated by various Marketing Departments have the final say?
By far, the more important question, however, is whether the designing/voting public can sort through this political firestorm — or tempest in a tea pot — and be allowed to make intelligent, long-term decisions that will guarantee the quality of their designs and insure their life, liberty, and pursuit of design closure.
Let's give the designers the respect and the voice they deserve. It's a free country — with free speech and a free market. In the long run, the tools worthy of success should succeed on their merit. Convenience is no substitute for excellence.
|
|
|
|
|
