Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Memory Compiler restrictions

Status
Not open for further replies.

ivlsi

Advanced Member level 3
Advanced Member level 3
Joined
Feb 17, 2012
Messages
883
Helped
17
Reputation
32
Reaction score
16
Trophy points
1,298
Visit site
Activity points
6,868
Hi All,

What restrictions for Memory Compiler do exist? Why is not possible to generate memories of any size?

Should a vendor provide a catalog of the available memories?

Thank you!
 

The memory compiler documentation indicate the possible range.
And usually when you start the compiler memory in GUI mode, the range is also displayed.
limitation occur on:
number of word width
number of address
mux
mask write enable (on 32bits you could writte byte by byte or bit by bit dependant of the compiler)
 
  • Like
Reactions: ivlsi

    ivlsi

    Points: 2
    Helpful Answer Positive Rating
Thanks. As for the Memory Compiler, does it generate the memories, which already exist in the vendor libraries? There are STD cells library, should also be the Memory libraries?
Why the Memories are restricted to a minimum size? What about the single row memories?
As for the STD cells, there are masks for each of them. Do the ASIC vendor have masks for all the memory combinations or it should produce them each time according to the Memory Compiler results.
Thank you again
 

depending of ram type, eeprom typer, flash type, OTP, the layer needed could be as for the std cell or need specific one, that are really dependant.

you could see on tsmc website what are available for each technologies.

one single row memory could be implemented with flop for less area than a ram instance.

do you know on which techno?
 
  • Like
Reactions: ivlsi

    ivlsi

    Points: 2
    Helpful Answer Positive Rating
Yes, as for the tech process, it's 40LP (40nm low, power).

BTW, when it's worth to choose flops and when memory? let's say for 256*4 array, is it better to implement it as flops or memory?

- - - Updated - - -

you could see on tsmc website what are available for each technologies" - searched the site, but did not find an answer. Could you provide a relative link please? Thank you
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top