Composition of SESAME stems
PRELIMINARY
“Consider alternately a SoC or hard virtual component having as primary objective to reach some threshold of low dynamic power consumption, say with a stringent figure of maximum dissipation, with the complementary goal to maximize performance in processing speed”
Extract from ‘IDRT-aware-SoC-Integration-Flow’
This quote illustrates well a serious problem that SoC designers have to handle: how to obtain the best “trade off” between 1 optimization and 2 constraints in the same block?
The traditional approach would be to offer one single library providing many cells of mixed performances for either one-pass or two-pass Synthesis…
The SESAME approach is based on the composition in one single circuit of 2 stems optimized for different criteria:one for the optimization, one for the constraint.
Illustration
- Suppose the main objective is to minimize cost => the recommended library stem is SESAME uHD (ultra High Density stem of SESAME)
- The constraint to be met is a clock frequency of 300 MHz => the recommended library stem is SESAME uHS (ultra High Speed stem of SESAME)

Glossary
The right method must be selected based on the information received after the first synthesis using proper analyzers for choosing the right combination method.
Criticity analyzers (pending):
The criticity analyzers are Dolphin EDA solutions to qualify results from the synthesis. Results are expressed in terms of quantity of critical paths (for speed) or consumption on “hot zones” (for power).
Input: reports of STA and from power estimator for localization of the critical areas
Output: criticity index of the design with respect to one constraint
Composition methods:
- Blending: composition of 2 stems with incremental or 1 pass synthesis
- Grafting: composition of 2 stems using a grafting script – 1 pass and a half synthesis
- Grafting script: script for replacement of the cells on the critical paths.
Grafting: why “1 pass and a half synthesis”?
“Classical” blending method | Dolphin innovative “grafting method”

|
Key benefits
- Performance for highly constrained designs
- Flexibility:chose the right couple of stems to achieve the target optimization of each SOC or each island of your SoC
|
Key Features
- strictly homogeneous stems:
- same functions
- same number of cells
- same number of tracks
|
Sesame stems table of compatibility for composition
The following stems are combinable.
Each one can be chosen either to reach the objective or to meet the constraint.

|