发明名称 Optimized buffer placement based on timing and capacitance assertions
摘要 Optimized buffer placement is provided based on timing and capacitance assertions in a functional chip unit including a single source and multiple macros, each having a sink. Placement of the source and macros with the sinks is pre-designed and buffers are placed in branches connecting the source with the multiple sinks. An estimated slack is calculated for each branch, the branches are arranged according to the calculated slack, decoupling buffers are inserted in all branches except the most critical branch(es), the most critical branch(es) are globally routed and slew conditions are fixed within this branch, and at least one next branch is globally routed and slew conditions are fixed therein.
申请公布号 US8930870(B2) 申请公布日期 2015.01.06
申请号 US201314034660 申请日期 2013.09.24
申请人 International Business Machines Corporation 发明人 Daellenbach Lukas;Gaugler Elmar;Richter Ralf
分类号 G06F17/50;G06F13/40 主分类号 G06F17/50
代理机构 Heslin Rothenberg Farley & Mesiti P.C. 代理人 McNamara, Esq. Margaret;Radigan, Esq. Kevin P.;Heslin Rothenberg Farley & Mesiti P.C.
主权项 1. A method for optimized buffer placement based on timing and capacitance assertions in a functional chip unit comprising a single source and multiple macros, each having a sink, wherein the placement of the source and the macros with the sinks is pre-designed and the buffers are placed in branches connecting the source with the multiple sinks, the method comprising: calculating, by a processor, an estimated slack for each branch of the branches; arranging the branches according to the calculated slack to evaluate at least one most critical branch of the branches; inserting decoupling buffers in each branch of the branches except for the at least one most critical branch; globally routing the at least one most critical branch and fixing slew conditions within this branch; subsequently, globally routing at least one next branch of the branches as arranged according to the calculated slack and fixing slew conditions within this at least one next branch, wherein globally routing the at least one next branch comprises: performing an initial routing of the at least one next branch between the source and a sink of the at least one next branch; andbased on a slack of the initial routing of the at least one next branch and a slack of the global routing of the at least one most critical branch, re-routing the at least one next branch between the source and the sink of the at least one next branch, wherein the re-routing of the at least one next branch comprises: removing a portion of the global routing of the at least one most critical branch between the source and an identified point on the global routing;removing the initial routing of the at least one next branch; androuting a replacement routing of the at least one next branch between the source and the sink of the at least one next branch via the identified point, wherein a portion of the replacement routing replaces the removed portion of the global routing of the at least one most critical branch; and routing at least one remaining branch of the branches.
地址 Armonk NY US