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.

Why should we place hard macros manually in PnR?

Status
Not open for further replies.

biju4u90

Full Member level 3
Joined
Dec 10, 2014
Messages
172
Helped
3
Reputation
6
Reaction score
3
Trophy points
18
Activity points
1,437
While doing PnR, why should we place hard macros manually and let the tool place standard cells only? Isn't the tool intelligent enught to place the hard macros too?
 

Macro has a custom specification about Layout, signal intergrity, timing and spacing....
Therefore, if you let the tool do placement for the macro ramdomly like standard cells, there will be a big number of cases because of the marco charateristics.
Instead, if you fix the macro, the tool only care about standard cells. That would be a reasonable approach to reduce running time, constraint complixity ....
 

for the macro placement first do the flyline analysis, that is analyze the connections between modelnpins and macro and macro to the combo logic, and macro to macro. this will give you the idea about the placement of macro with optimal routing resources (So we are doing all the steps manually).
 

It is a good question. Infact macro placement can be either automatic or manual. In general manual placement is very efficient if you have few macro's to place and engineer is well aware of the connectivity and data flow diagram. If you have hundreds of macro's to place and not aware of data flow then we can rely on the automatic macro placement solution by the placer. Most of the placement tools even offer automatic macro placement solutions. All these solutions use a flavour of global placer solution which is present in the PnR tool and as you are aware that global placement solution main objective is to place the cells based on connectivity and reduce wire length. The same approach is used in the automatic macro placement solution and in fact the wire length optimisation is the most promising approach in automatic macro placement solution.

Automatic macro placement solution is an iterative process, I call it as semi-automatic macro placement. Simple automatic macro placement solution is that any macro is free to be moved until min cost of wire length solution is achieved. If you do not have any connectivity between macro's then it would be difficult for the algorithm to achieve any solution. In such case we need to concurrently place the standard cells and refine macro placement based on the standard cell connectivity.

Just to answer your question in short to achieve automatic macro placement solution is a time and space problem which needs to be solved simultaneously.

My recommendation is use the automatic macro placement solution for the initial macro placement and refine the placement and fix the macro's which you finished and again re-run the auto macro placer until you finish the placement of all macros considering all the constraints w.r.t to macro.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top