Sm90 mega moe on sgl dev

[qiushixiaoyu]

Batch	Fused avg us	Baseline avg us	Baseline / Fused	Fused TFLOPS	Baseline TFLOPS	Fused HBM GB/s	Baseline HBM GB/s
1	183.4	327.6	1.787	1.6	1	755.1	422.8
2	263	380.4	1.446	2.1	1.5	1005.5	695.6
4	406.1	497.4	1.225	3	2.4	1070.5	873.6
8	497.1	546.1	1.099	4.8	4.5	1293.1	1177.2
16	566	641.2	1.133	8.4	7.4	1376.8	1214.6
32	576	651	1.13	16.8	14.8	1404.6	1242.4
64	592.5	653.2	1.103	32.8	29.6	1371.9	1242.5
128	597.9	680.1	1.138	64.9	56.9	1370.9	1202.9
512	1144	1220.9	1.067	135.9	126.6	752.1	702
1024	1989.5	2189.1	1.1	156	141.1	458.8	415
4096	6949.8	6913.9	0.995	179	179	176	176
8192	13514.9	13343.6	0.987	184.2	185.4	121.2	122.2