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.

ASIC flow for power estimation

Status
Not open for further replies.

inputoutput

Member level 1
Joined
Mar 18, 2017
Messages
32
Helped
2
Reputation
4
Reaction score
2
Trophy points
8
Activity points
231
Presently I'm following the power optimization flow described in the Power Compiler (Version D-2010.03-SP2) described on page 2-3. Essentially the flow is:

- Simulate RTL and generate SAIF
- Synthesize in DC compiler to get gate-level netlist
- Place and route design
- Simulate gate-level or placed netlist to generate new SAIF
- Re-synthesize in DC/Power compiler to perform power optimization and get power consumption of design

My question is if a placed netlist is used for the power optimization (last step), does the place and route step need to be performed again? Is it typical to route a design twice or more? I imagine for very large designs this might not be viable?
 

Hi,

Once you change the netlist for whatever reason, it should be a bunch of things to do afterward.
- Formal verification
- DFT insertion
- CTS, Place & Route.

In case of your flow, there MUST have some guidance for p&n steps from the last step.
So that, the final design would have better power consumption.
 

Presently I'm following the power optimization flow described in the Power Compiler (Version D-2010.03-SP2) described on page 2-3. Essentially the flow is:

- Simulate RTL and generate SAIF
- Synthesize in DC compiler to get gate-level netlist
- Place and route design
- Simulate gate-level or placed netlist to generate new SAIF
- Re-synthesize in DC/Power compiler to perform power optimization and get power consumption of design

My question is if a placed netlist is used for the power optimization (last step), does the place and route step need to be performed again? Is it typical to route a design twice or more? I imagine for very large designs this might not be viable?

This is one flow, it is not the only flow. Here is my view on the problem:

- Simulate RTL and generate SAIF/VCD
- Synthesize in DC compiler to get gate-level netlist of a block
- Place and route the block design
- Simulate gate-level placed netlist to generate new SAIF/VCD
- Use SAIF/VCD as input for the backend P&R tools
- Measure power with the backend tool. This is as accurate as you can get.
If power budget violated, trigger resynthesis with new targets, start over.
 

Thanks for your suggestions @Sam and @slutarius
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top