CONTANGO: Integrated Optimization of SoC Clock Networks

Table 6

Results from ASPDAC’10 clock routing papers on the ISPD’09 Contest benchmark suite [26–28]. Runtimes may be from different workstations. CLR and skew are reported in ps and CPU time in 𝑠. Only average skew was published for HKPU [28].

Benchmark	Contango (this work)			NTU [26]			NCTU [27]			HKPU [28]
	CLR	Skew	⌚	CLR	Skew	⌚	CLR	Skew		CLR	Skew	⌚
ispd09f11	13.36	2.867	6488	19.71	4.478	4639	18.77	7.12	30787	12.2	—	180
ispd09f12	15.27	2.611	6564	17.46	4.088	4231	15.5	3.06	27622	10.9	—	213
ispd09f21	17.40	2.738	6673	19.92	3.868	4629	17.04	3.02	33056	12.1	—	210
ispd09f22	12.36	2.227	3618	16.47	3.671	3937	16.25	4.11	19136	9.9	—	113
ispd09f31	12.81	3.91	21379	31.13	4.762	11112	22.63	7.58	66588	13.4	—	777
ispd09f32	17.92	4.594	12895	23.04	4.234	7293	20.59	5.52	49907	11.5	—	420
ispd09fnb1	13.40	3.5	778	15.73	6.798	3719	14.32	3.77	7643	13.8	—	82
Average	14.65	3.207	8342	20.49	4.56	5651	17.87	4.88	33534	11.97	7.72	285