Skip to main content

Table 2 OpenCL kernel performance for different number of query sequences

From: Coupling SIMD and SIMT architectures to boost performance of a phylogeny-aware alignment kernel

  Execution times GCUPS Speedup GPU vs
Number of queries Seq SSE(4) GPU Seq SSE(4) GPU Seq SSE(4)
100 39.60 1.14 0.78 0.40 14.10 20.2 50.8 1.46
250 98.31 2.78 1.87 0.41 14.27 20.9 52.6 1.48
500 197.57 5.57 3.7 0.40 14.33 21.1 53.4 1.51
750 296.23 8.40 5.5 0.40 14.24 21.1 53.9 1.53
1000 395.59 11.20 7.4 0.40 14.28 21.2 53.4 1.51
1250 494.65 14.30 9.23 0.40 14.29 21.2 53.6 1.52
  1. Execution times (in seconds) to align the QS to 320 RS with length 5000.