A Novel CSR-Based Sparse Matrix-Vector Multiplication on GPUs

Table 4

Comparison of PCSRI and PCSRII without communication on four GPUs.

Matrix	ET (GPU)	PCSRI (4 GPUs)			PCSRII (4 GPUs)
		ET	SD	PE	ET	SD	PE
2cubes_sphere	0.4444	0.1560	0.0132	71.23	0.1527	0.0111	72.78
scircuit	0.3484	0.1453	0.0262	59.94	0.1357	0.0130	64.17
Ga41As41H72	4.2387	1.6123	0.7268	65.72	1.3410	0.1846	79.02
F1	6.5544	2.5240	0.6827	64.92	1.9121	0.1900	85.69
ASIC_680ks	0.8196	0.2944	0.0298	69.59	0.2887	0.0264	70.98
ecology2	1.2321	0.3593	0.0160	85.72	0.3554	0.0141	86.67
Hamrle3	1.7684	0.5114	0.0307	86.45	0.4775	0.0125	92.59
thermal2	2.0708	0.5553	0.0271	93.22	0.5546	0.0255	93.33
cage14	5.9177	1.8126	0.3334	81.62	1.5386	0.0188	96.15
Transport	4.7305	1.2292	0.0270	96.21	1.2275	0.0158	96.35
G3_circuit	1.9731	0.5804	0.0489	84.99	0.6195	0.0790	79.63
kkt_power	4.3465	1.4974	0.5147	72.57	1.1584	0.0418	93.80
CurlCurl_4	5.1605	1.3554	0.0153	95.18	1.3501	0.0111	95.56
memchip	3.8257	1.1439	0.1741	83.61	1.1175	0.1223	85.59
Freescale1	5.0524	1.7588	0.4039	71.81	1.4806	0.1843	85.31