ᴡʜʏ ᴏʀᴅᴇʀ ᴏғ ғɪʟʟɪɴɢ ɪs ғᴏʀᴍ ᴍᴏʀᴇ ᴡɪᴅᴛʜ ғɪʟʟᴇʀ ᴄᴇʟʟs ᴛᴏ ʟᴏᴡ ᴡɪᴅᴛʜ ғɪʟʟᴇʀ ᴄᴇʟʟs?

Why order of filling is form more width filler cells to low width filler cells?

It forms the wide metal due to continuous placement of low width filler cells. Creates Metal slotting rule violation. The CMP damascene process also introduces undesirable side effects, including dielectric erosion and metal dishing.

Please always insert fat fillers first and then thin fillers afterwards. To avoid the metal-slot-rule violation, please do not only use thin filler cells to fill large I/O spacing. For example, use one 20um pitch filler cell (PFILLER20) and one 10um pitch filler cell (PFILLER10) instead of using 6 “5um pitch” filler cells (PFILLER5) to fill 30um spacing. In addition, if spacing is larger than one cell pitch, please first insert core power cell and I/O power cell. Then insert filler cells to fill up the rest of spacing for ESD robustness.

Comments

Sanity checks before floorplan in Physical Design

Sanity checks before floorplan in Physical design Sanity checks are an important step for physical design engineers to make sure that the inputs received for physical design are correct and consistent. Any issues in the input may cause problems in the later stages. So it is important to perform the sanity checks in the initial stage that is when the design is loaded in PnR tool and before the start of the floorplan. Here is a list of checks which must be performed before floorplan of design. Figure-1: Sanity checks before floorplan Library Check: In library check, basically, we validate the libraries before starting the physical design by checking the consistency between the physical and logical library. It also checks the quality of both libraries and reports the error if any. The cells used in the design must be present in the logical as well as in the physical library. Innovus commands: check_design -physicalLibrary : This command will check the physical library and repor...

Why we are not checking the hold before CTS?

Why we are not checking the hold before CTS? Ans: Before CTS, clock is ideal that means exact skew is not there. All the clocks reaching the flops at the same Time. So, we dont have skew and transition numbers of the clock path, but this information is sufficient to perform setup analysis since setup violation depends on the data path delay. Clock is propagted only after CTS (actual clock tree is built, clock buffers are added & clock tree hierarchy, clock skew, insertion delay comes into picture) and that's why hold violations are fixed only after CTS.

Search This Blog