His latest SNUG report is full of SNPS cheerleading. His latest report at deepchip is constantly harping on Physical Compiler "tape-outs" again and again when it is clear that PC has neither the clock tree synthesis engine nor routing engine to complete the tapeouts. Cadence/avanti Backend flow is still required for the tape-outs!! with PC. It is clear the so called testimonials is full of shit by front-end folks who never saw a layout in their lives cheering as if its the next thing to sliced bread. None of them gave details about the iterations that the team may have gone thro' to meet timing closure and the contribution of back-end (timing driven) tools to the sucess of meeting timing. His FUD is becoming unbearable and he has the audacity to accuse likes of CDN for FUD.
The original comment had nothing to do with Cooley trashing CDN, CNPS or anybody else. It was a disagreement with the term "tape out" as applied to Physical Compiler. If the output from Physical Compiler is not fed directly to the mask house to make a masks, then it is not a tape out.
Since before I started in the IC industry in 1979, "tape out" has always referred to the act of sending the mask design to the mask shop to have a reticle made for the step and repeat machine to make production masks. We have more sophisticated techniques than reticles and step and repeat machines for making masks but "tape out" has remained the same.
Cooley should only use tape out if there is no other tool involved between Physical Compiler and the tape/disk/FTP that moves the design to the mask shop.