TIOVX User Guide
J742S2 Linux Performance Report
SoC J742S2
Test EVM J742S2 TI EVM
HOST ARM A72 running Linux at default frequency
Target Standalone C7x DSP running Free RTOS at default frequency, L1/L2 Cache Enabled (Cache-only based performance.)



TIOVX Kernel Performance
Index Kernel Variant Frame Size (Pixels) Graph Performance (msec) Node Performance (msec)
1 Accumulate 640x480 (307200) 1.335000 1.234000
2 AccumulateSquare 640x480 (307200) 1.917000 1.822000
3 AccumulateWeighted 640x480 (307200) 0.941000 0.847000
4 Addition U8+U8=U8 640x480 (307200) 0.729000 0.633000
5 Addition U8+U8=S16 640x480 (307200) 0.838000 0.744000
6 Addition U8+S16=S16 640x480 (307200) 1.303000 1.208000
7 Addition S16+S16=S16 640x480 (307200) 1.338000 1.243000
8 Subtraction U8-U8=U8 640x480 (307200) 0.738000 0.646000
9 Subtraction U8-U8=S16 640x480 (307200) 0.793000 0.698000
10 Subtraction S16-U8=S16 640x480 (307200) 1.931000 1.835000
11 Subtraction S16-S16=S16 640x480 (307200) 1.337000 1.239000
12 Not Operation U8 640x480 (307200) 0.643000 0.554000
13 And Operation U8 640x480 (307200) 0.725000 0.632000
14 OR Operation U8 640x480 (307200) 0.701000 0.610000
15 XOR Operation U8 640x480 (307200) 0.710000 0.620000
16 Abs Diff U8 640x480 (307200) 0.758000 0.670000
17 Abs Diff S16 640x480 (307200) 1.549000 1.454000
18 Box 3x3 640x480 (307200) 1.439000 1.350000
19 Dilate 3x3 640x480 (307200) 0.818000 0.716000
20 Erode 3x3 640x480 (307200) 0.813000 0.722000
21 Fast Corners No NMS 640x480 (307200) 13.696000 13.586000
22 Canny 640x480 (307200) 7.662000 7.561000
23 Channel Combine 3 to 1 (RGB) 640x480 (307200) 1.156000 1.059000
24 Channel Combine 4 to 1 (RGBX) 640x480 (307200) 1.286000 1.170000
25 Channel Combine YUYV 640x480 (307200) 1.158000 1.063000
26 Channel Combine 2 to 1 (NV12) 640x480 (307200) 1.025000 0.935000
27 Channel Extract 1 of 3 (RGB) 640x480 (307200) 1.738000 1.643000
28 Channel Extract 1 of 4 (RGBX) 640x480 (307200) 2.233000 2.140000
29 Color Convert RGB=>RGBX 640x480 (307200) 3.519000 3.378000
30 Color Convert RGB=>NV12 640x480 (307200) 3.102000 3.004000
31 Color Convert RGB=>IYUV 640x480 (307200) 3.113000 3.020000
32 Color Convert RGB=>YUV4 640x480 (307200) 3.067000 2.965000
33 Color Convert RGBX=>RGB 640x480 (307200) 3.747000 3.646000
34 Color Convert RGBX=>NV12 640x480 (307200) 3.592000 3.493000
35 Color Convert RGBX=>IYUV 640x480 (307200) 3.535000 3.436000
36 Color Convert RGBX=>YUV4 640x480 (307200) 3.667000 3.566000
37 Color Convert NVXX=>RGB 640x480 (307200) 2.932000 2.829000
38 Color Convert NVXX=>RGBX 640x480 (307200) 1.899000 1.799000
39 Color Convert NVXX=>IYUV 640x480 (307200) 0.990000 0.895000
40 Color Convert NVXX=>YUV4 640x480 (307200) 1.174000 1.046000
41 Color Convert YUVX=>RGB 640x480 (307200) 3.299000 3.198000
42 Color Convert YUVX=>RGBX 640x480 (307200) 3.610000 3.508000
43 Color Convert YUVX=>NV12 640x480 (307200) 1.798000 1.695000
44 Color Convert YUVX=>IYUV 640x480 (307200) 1.880000 1.780000
45 Color Convert IYUV=>RGB 640x480 (307200) 2.569000 2.470000
46 Color Convert IYUV=>RGBX 640x480 (307200) 2.982000 2.873000
47 Color Convert IYUV=>NV12 640x480 (307200) 1.098000 0.999000
48 Color Convert IYUV=>YUV4 640x480 (307200) 0.925000 0.826000
49 Convert Depth U8 to S16 640x480 (307200) 0.723000 0.623000
50 Convert Depth S16 to U8 640x480 (307200) 1.314000 1.220000
51 Convolution 3x3, O: U8 640x480 (307200) 1.357000 1.263000
52 Convolution 3x3, O: S16 640x480 (307200) 1.302000 1.203000
53 Convolution 9x3, O: U8 640x480 (307200) 47.407000 47.274000
54 Convolution 9x3, O: S16 640x480 (307200) 48.306000 48.174000
55 Convolution 3x9, O: U8 640x480 (307200) 62.992000 62.888000
56 Convolution 3x9, O: S16 640x480 (307200) 63.799000 63.683000
57 Convolution 5x5, O: U8 640x480 (307200) 2.396000 2.303000
58 Convolution 5x5, O: S16 640x480 (307200) 2.411000 2.316000
59 Convolution 7x7, O: U8 640x480 (307200) 3.812000 3.721000
60 Convolution 7x7, O: S16 640x480 (307200) 4.077000 3.977000
61 Convolution 9x9, O: U8 640x480 (307200) 121.919000 121.817000
62 Convolution 9x9, O: S16 640x480 (307200) 122.840000 122.731000
63 Equalize Histogram 640x480 (307200) 2.011000 1.917000
64 Gaussian 3x3 640x480 (307200) 1.013000 0.903000
65 Gaussian Pyramid 640x480 (307200) 7.008000 6.902000
66 Half Scale Gaussian Pyramid 640x480 (307200) 0.733000 0.640000
67 Harris Corners 640x480 (307200) 16.692000 16.581000
68 Histogram U8 640x480 (307200) 1.302000 1.211000
69 Integral Image 640x480 (307200) 1.296000 1.194000
70 Laplacian Pyramid U8; Levels = 6 640x480 (307200) 3.420000 3.322000
71 Laplacian Reconstruct U8 640x480 (307200) 4.261000 4.158000
72 LookUpTable S16 640x480 (307200) 1.963000 1.861000
73 LookUpTable U8 640x480 (307200) 0.943000 0.852000
74 Magnitude S16 640x480 (307200) 3.929000 3.811000
75 Mean/Standard Deviation U8 640x480 (307200) 0.694000 0.603000
76 Median 3x3 640x480 (307200) 1.490000 1.399000
77 Min Max Loc U8 640x480 (307200) 1.168000 1.075000
78 Min Max Loc S16 640x480 (307200) 2.265000 2.171000
79 Multiply U8 x U8 = U8 640x480 (307200) 0.838000 0.745000
80 Multiply U8 x U8 = S16 640x480 (307200) 0.900000 0.809000
81 Multiply U8 x S16 = S16 640x480 (307200) 2.144000 2.047000
82 Multiply S16 x S16 = S16 640x480 (307200) 2.674000 2.576000
83 NonLinear Filter mask=5x5; Erosion 640x480 (307200) 2.860000 2.766000
84 NonLinear Filter mask=5x5; Dilation 640x480 (307200) 2.864000 2.775000
85 NonLinear Filter mask=5x5; Median 640x480 (307200) 60.361000 60.255000
86 Optical Flow 5x5 640x480 (307200) 4.853000 4.746000
87 Phase 640x480 (307200) 4.836000 4.742000
88 Remap Bilinear Interpolation *640x480 (307200) 9.310000 9.203000
89 Remap Nearest Neighbor *640x480 (307200) 7.724000 7.618000
90 Scale Image 1 to 2; Nearest Neighbor *640x480 (307200) 1.419000 1.322000
91 Scale Image 1 to 2; Bilinear Interpolation *640x480 (307200) 5.240000 5.150000
92 Sobel 3x3 640x480 (307200) 1.729000 1.622000
93 Threshold Binary 640x480 (307200) 0.745000 0.655000
94 Warp Affine Nearest Neighbor *256x256 (65536) 0.380000 0.288000
95 Warp Affine Bilinear Interpolation *256x256 (65536) 1.153000 1.061000
96 Warp Pespective Nearest Neighbor *256x256 (65536) 0.627000 0.538000
97 Warp Pespective Bilinear Interpolation *256x256 (65536) 1.748000 1.655000

Footnote 1: All optional parameters for relevant kernels are used

Footnote 2: All image sizes are for input parameters unless noted with an asterisk (*) in which case these are the output parameter sizes