Difference between cadence's assuar , mentor's calibre and synopsys' Hercules

I am looking for real technical differences. Things like format and in different situation, what kind of error each one of these gives.

Our foundry advises provides rule decks for all three tools. As per their notes, Calibre is better at detecting width/space violations and Assura is better for density checks. Nothing specific about Hercules.

Assura and Calibre are comparable when it comes to DRC errors in general. Both have extensive commands to develop rule decks. What they catch is more dependent on how well the rule decks are written. LVS is slightly different. Assura relies more on graphical highlighting and I find that its LVS error output to be harder to understand than Calibre's.

I don't have any experience with Hercules but as it is a Synopsys tool, I suspect it is geared more towards digital verification. From things I have heard, it is not as technically advanced as either Calibre or Assura.

It is hard to give you more of an evaluation without more specifics about what you are verifying.

Re: Difference between cadence's assura , mentor's calibre and synopsys' Hercules

jm3395 said:

... Assura ... I find that its LVS error output to be harder to understand than Calibre's.

Click to expand...

Aren't these error messages coded, too, in the LVS rules deck?

On the other hand, I've seen these (initially obscure) "multiple stamped connections" error messages from various foundries' rules decks.

Re: Difference between cadence's assura , mentor's calibre and synopsys' Hercules

erikl said:

Aren't these error messages coded, too, in the LVS rules deck?

On the other hand, I've seen these (initially obscure) "multiple stamped connections" error messages from various foundries' rules decks.

Click to expand...

Yes, the LVS deck is also written by the foundry or users, however, the way the output is presented is not. Cadence's older tool, Dracula, did side by side comparison on nets and devices. Assura has sections like "Bad matched nets" and "nets that don't really match" and it does not do a side by side comparison (schematic vs. layout) so I find it harder to interpret. Also, we have seen a lot more bugs with Assura than I ever saw with Calibre (or Dracula for that matter).

As far as multistamped connections, those are written by the foundry or user as well. That is something I am currently working on for one of the rule decks I have responsibility for.

Re: error messages from the rules' decks

jm3395 said:

... As far as multistamped connections, those are written by the foundry or user as well. That is something I am currently working on for one of the rule decks I have responsibility for.

Click to expand...

Thank you for your answer, jm,

formerly I'd asked our team which was responsible for the ERC, DRC, XTR, & LVS decks to write understandable error messages, because all the junior designers used to ask the same questions, e.g. what these "multistamped connections" would mean. Unfortunately I couldn't persuade them to do so, hence the questions continued :-( Perhaps you get this opportunity now?

Good luck! erikl

burmanpankaj said:

I am looking for real technical differences. Things like format and in different situation, what kind of error each one of these gives.

Click to expand...

As you might have already read above, no matter how good the tool is, you won't be able to make effective use of it unless the foundry-provided rules are sufficiently mature (ie. bug-free). This is especially true for parasitic extraction. Base your choice of a physical verification tool on the process you are going to use, reading release notes from foundry and/or design house. E.g. Calibre is a wonderful tool once you master it but we had to deal with a major foundry process were the LVS/xRC results in Calibre were so spotty that the xRC flow broke down in most top-level extractions. We will try Hercules/StarRC in the next iteration, in the past rules and tool UI were quite poor but we hope that integration into Custom Designer changed all that...

I am satisfied with jm3395. I am looking for these kind of answers.