ARMv8 Neoverse-V2 testing with a Quanta Cloud QuantaGrid S74G-2U 1S7GZ9Z0000 S7G MB (CG1) (3A06 BIOS) and NVIDIA GH200 480GB on Ubuntu 22.04 via the Phoronix Test Suite.
Compare your own system(s) to this result file with the
Phoronix Test Suite by running the command:
phoronix-test-suite benchmark 2403013-NE-NGCSMOKER54
ngc smoke run,
"VkFFT 1.3.4 - Test: FFT + iFFT R2C / C2R",
Higher Results Are Better
"a",42782,42600,41808
"b",42355,40894,42178
"c",43390,42827,41525
"d",45046,41892,42267,43019,45265,42575,42576,43002,42396,43137,44419,41306,43309,43352,42153
"VkFFT 1.3.4 - Test: FFT + iFFT C2C 1D batched in half precision",
Higher Results Are Better
"a",151751,152292,151694
"b",151755,151120,152855
"c",153123,152820,152656
"d",151288,152592,152027
"VkFFT 1.3.4 - Test: FFT + iFFT C2C Bluestein in single precision",
Higher Results Are Better
"a",17709,17764,18129
"b",18374,17356,17839,18414,17852
"c",17629,17890,18139
"d",17836,17107,18437,17868,17857,18391,18097
"VkFFT 1.3.4 - Test: FFT + iFFT C2C 1D batched in double precision",
Higher Results Are Better
"a",58242,58705,58268
"b",58285,58161,58314
"c",58240,58291,58238
"d",58327,58256,58314
"VkFFT 1.3.4 - Test: FFT + iFFT C2C 1D batched in single precision",
Higher Results Are Better
"a",188506,185705,183111
"b",183922,186843,187482
"c",186621,191398,191814
"d",189450,191109,190370
"VkFFT 1.3.4 - Test: FFT + iFFT C2C multidimensional in single precision",
Higher Results Are Better
"a",44389,47864,43165,46231,42459,46336,46223,43923,41950,47005,43116,44033,44050,42020,44577
"b",42911,43858,44424
"c",45948,44313,44952
"d",44744,44176,46102
"VkFFT 1.3.4 - Test: FFT + iFFT C2C Bluestein benchmark in double precision",
Higher Results Are Better
"a",20705,21177,20549
"b",20612,21152,21236
"c",21649,20722,20911
"d",20732,21976,20860,20342,22039,21472,21526,20721,22013,20912,20527,21316,21909,21835,21616
"VkFFT 1.3.4 - Test: FFT + iFFT C2C 1D batched in single precision, no reshuffling",
Higher Results Are Better
"a",190747,198561,194184
"b",189020,190349,190742
"c",190892,191928,189906
"d",191337,193821,192363
"cl-mem 2017-01-13 - Benchmark: Copy",
Higher Results Are Better
"a",308.6,308.5,308.6
"b",308.7,308.6,308.3
"c",308.5,308.7,308.7
"d",308.5,308.5,308.6
"cl-mem 2017-01-13 - Benchmark: Read",
Higher Results Are Better
"a",1045.9,1045.9,1045.9
"b",1046.1,1045.5,1046.1
"c",1046.1,1046,1046.1
"d",1046,1046,1045.9
"cl-mem 2017-01-13 - Benchmark: Write",
Higher Results Are Better
"a",2352.3,2355.7,2356.6
"b",2354,2354.6,2351.7
"c",2353.8,2356.9,2354.1
"d",2344.5,2356.1,2355.7
"VkResample 1.0 - Upscale: 2x - Precision: Double",
Lower Results Are Better
"a",24.294,24.297,24.298
"b",24.293,24.295,24.295
"c",24.284,24.292,24.295
"d",24.294,24.295,24.301
"VkResample 1.0 - Upscale: 2x - Precision: Single",
Lower Results Are Better
"a",5.23,5.23,5.229
"b",5.232,5.23,5.23
"c",5.229,5.232,5.23
"d",5.228,5.229,5.233
"clpeak 1.1.2 - OpenCL Test: Integer Compute INT",
Higher Results Are Better
"a",33114.88,33118.75,33123.66
"b",33144.83,33144.83,33144.57
"c",33146.38,33146.38,33145.61
"d",33144.05,33115.92,33128.04
"clpeak 1.1.2 - OpenCL Test: Single-Precision Float",
Higher Results Are Better
"a",64544.8,64546.27,64545.78
"b",64546.27,64549.21,64547.73
"c",64546.27,64548.22,64547.25
"d",64522.77,64519.82,64520.31
"clpeak 1.1.2 - OpenCL Test: Double-Precision Double",
Higher Results Are Better
"a",32957.8,32960.36,32959.34
"b",32962.4,32959.84,32961.38
"c",32904.75,32960.36,32960.86
"d",32934.32,32935.84,32930.74
"clpeak 1.1.2 - OpenCL Test: Global Memory Bandwidth",
Higher Results Are Better
"a",3483.35,3484.48,3484.14
"b",3484.14,3483.69,3484.36
"c",3484.48,3483.57,3483.8
"d",3484.36,3484.36,3484.25
"ArrayFire 3.9 - Test: Conjugate Gradient OpenCL",
Lower Results Are Better
"a",2.992,2.997,3.003
"b",2.988,2.988,2.972
"c",3.003,2.988,3.002
"d",3.002,2.997,2.992
"FinanceBench 2016-07-25 - Benchmark: Black-Scholes OpenCL",
Lower Results Are Better
"a",4.33,4.347,4.364
"b",4.368,4.369,4.381
"c",4.34,4.342,4.37
"d",4.324,4.372,4.322
"ViennaCL 1.7.1 - Test: CPU BLAS - sCOPY",
Higher Results Are Better
"a",2880,2940,2940
"b",2960,2810,2950,2840,2900
"c",2910,2900,2910
"d",2810,2880,2880
"ViennaCL 1.7.1 - Test: CPU BLAS - sAXPY",
Higher Results Are Better
"a",3920,3970,3940
"b",3930,3930,3950,3890,3920
"c",3900,3950,3900
"d",3930,3940,3890
"ViennaCL 1.7.1 - Test: CPU BLAS - sDOT",
Higher Results Are Better
"a",675,662,663
"b",685,652,665,660,656
"c",675,662,662
"d",667,666,656
"ViennaCL 1.7.1 - Test: CPU BLAS - dCOPY",
Higher Results Are Better
"a",1940,2050,2090
"b",1890,1960,1980,1980,1930
"c",1840,1940,1980
"d",1920,1890,1940
"ViennaCL 1.7.1 - Test: CPU BLAS - dAXPY",
Higher Results Are Better
"a",1840,1760,1810
"b",1700,1780,1850,1860,1840
"c",1840,1830,1840
"d",1820,1820,1850
"ViennaCL 1.7.1 - Test: CPU BLAS - dDOT",
Higher Results Are Better
"a",1240,1250,1250
"b",1240,1230,1240,1240,1240
"c",1250,1240,1240
"d",1240,1250,1250
"ViennaCL 1.7.1 - Test: CPU BLAS - dGEMV-N",
Higher Results Are Better
"a",411,411,410
"b",406,408,418,408,400
"c",404,403,409
"d",413,435,407
"ViennaCL 1.7.1 - Test: CPU BLAS - dGEMV-T",
Higher Results Are Better
"a",670,720,667
"b",699,702,711,691,692
"c",689,693,690
"d",686,684,717
"ViennaCL 1.7.1 - Test: CPU BLAS - dGEMM-NN",
Higher Results Are Better
"a",137,134,135
"b",141,139,138,135,134
"c",136,144,136
"d",142,136,145
"ViennaCL 1.7.1 - Test: CPU BLAS - dGEMM-NT",
Higher Results Are Better
"a",126,125,123
"b",128,124,124,124,125
"c",125,124,124
"d",125,124,124
"ViennaCL 1.7.1 - Test: CPU BLAS - dGEMM-TN",
Higher Results Are Better
"a",143,140,139
"b",144,140,139,140,139
"c",142,140,141
"d",141,139,140
"ViennaCL 1.7.1 - Test: CPU BLAS - dGEMM-TT",
Higher Results Are Better
"a",138,134,139
"b",140,135,146,135,135
"c",138,149,133
"d",137,134,138
"ViennaCL 1.7.1 - Test: OpenCL BLAS - sCOPY",
Higher Results Are Better
"a",316,315,316
"b",316,315,316
"c",315,316,316
"d",315,316,316
"ViennaCL 1.7.1 - Test: OpenCL BLAS - sAXPY",
Higher Results Are Better
"a",416,420,425
"b",430,426,426
"c",431,423,426
"d",421,424,434
"ViennaCL 1.7.1 - Test: OpenCL BLAS - sDOT",
Higher Results Are Better
"a",282,282,282
"b",283,282,282
"c",283,283,284
"d",284,282,284
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dCOPY",
Higher Results Are Better
"a",603,603,604
"b",605,603,604
"c",605,602,604
"d",603,604,605
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dAXPY",
Higher Results Are Better
"a",799,798,799
"b",797,798,798
"c",801,798,797
"d",797,799,800
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dDOT",
Higher Results Are Better
"a",549,552,550
"b",553,552,552
"c",554,550,552
"d",554,551,554
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dGEMV-N",
Higher Results Are Better
"a",81.1,81.1,81.5
"b",81.4,81.3,81.7
"c",81.7,80.8,81.2
"d",81,81.7,81.5
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dGEMV-T",
Higher Results Are Better
"a",308,308,308
"b",308,308,308
"c",308,308,307
"d",308,307,307
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dGEMM-NN",
Higher Results Are Better
"a",7120,7030,7020
"b",7020,7050,7210
"c",7100,7000,7010
"d",7120,7000,7040
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dGEMM-NT",
Higher Results Are Better
"a",7530,7520,7530
"b",7550,7540,7520
"c",7540,7530,7540
"d",7540,7540,7540
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dGEMM-TN",
Higher Results Are Better
"a",7060,7020,7000
"b",6990,7060,7150
"c",7030,6990,6980
"d",7160,7020,7030
"ViennaCL 1.7.1 - Test: OpenCL BLAS - dGEMM-TT",
Higher Results Are Better
"a",7070,7070,7070
"b",7070,7090,7050
"c",7060,7050,7060
"d",7070,7070,7070
"NCNN 20230517 - Target: Vulkan GPU - Model: mobilenet",
Lower Results Are Better
"a",4.86,4.89,4.93
"b",4.97,4.89,4.9
"c",4.92,4.87,4.96
"d",4.91,4.89,4.92
"NCNN 20230517 - Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2",
Lower Results Are Better
"a",2.08,2.14,2.17
"b",2.13,2.15,2.19
"c",2.04,2.14,2.19
"d",2.02,2.15,2.2
"NCNN 20230517 - Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3",
Lower Results Are Better
"a",2.2,2.29,2.3
"b",2.26,2.29,2.26
"c",2.19,2.33,2.37
"d",2.19,2.29,2.33
"NCNN 20230517 - Target: Vulkan GPU - Model: shufflenet-v2",
Lower Results Are Better
"a",2.26,2.29,2.31
"b",2.26,2.3,2.31
"c",2.26,2.25,2.31
"d",2.2,2.31,2.3
"NCNN 20230517 - Target: Vulkan GPU - Model: mnasnet",
Lower Results Are Better
"a",2,2.05,2.08
"b",2.04,2.01,2.05
"c",1.98,2.03,2.12
"d",1.98,2.06,2.07
"NCNN 20230517 - Target: Vulkan GPU - Model: efficientnet-b0",
Lower Results Are Better
"a",3.48,3.46,3.52
"b",3.55,3.6,3.5
"c",3.47,3.48,3.6
"d",3.45,3.54,3.59
"NCNN 20230517 - Target: Vulkan GPU - Model: blazeface",
Lower Results Are Better
"a",1.74,1.75,1.76
"b",1.82,1.74,1.77
"c",1.72,1.71,1.8
"d",1.75,1.75,1.81
"NCNN 20230517 - Target: Vulkan GPU - Model: googlenet",
Lower Results Are Better
"a",4.15,4.14,4.2
"b",4.33,4.21,4.16
"c",4.16,4.19,4.34
"d",4.21,4.19,4.23
"NCNN 20230517 - Target: Vulkan GPU - Model: vgg16",
Lower Results Are Better
"a",5.25,5.23,5.29
"b",5.26,5.26,5.23
"c",5.2,5.23,5.31
"d",5.23,5.25,5.29
"NCNN 20230517 - Target: Vulkan GPU - Model: resnet18",
Lower Results Are Better
"a",2.18,2.13,2.18
"b",2.19,2.16,2.19
"c",2.18,2.14,2.2
"d",2.19,2.18,2.22
"NCNN 20230517 - Target: Vulkan GPU - Model: alexnet",
Lower Results Are Better
"a",1.67,1.61,1.6
"b",1.66,1.63,1.6
"c",1.74,1.61,1.61
"d",1.62,1.62,1.63
"NCNN 20230517 - Target: Vulkan GPU - Model: resnet50",
Lower Results Are Better
"a",4.24,4.28,4.29
"b",4.3,4.31,4.22
"c",4.27,4.3,4.39
"d",4.29,4.31,4.35
"NCNN 20230517 - Target: Vulkan GPUv2-yolov3v2-yolov3 - Model: mobilenetv2-yolov3",
Lower Results Are Better
"a",4.86,4.89,4.93
"b",4.97,4.89,4.9
"c",4.92,4.87,4.96
"d",4.91,4.89,4.92
"NCNN 20230517 - Target: Vulkan GPU - Model: yolov4-tiny",
Lower Results Are Better
"a",6.8,6.78,6.79
"b",6.76,6.85,6.79
"c",6.76,6.84,6.84
"d",6.85,6.82,6.79
"NCNN 20230517 - Target: Vulkan GPU - Model: squeezenet_ssd",
Lower Results Are Better
"a",5.38,5.42,5.49
"b",5.61,5.42,5.39
"c",5.34,5.51,5.58
"d",5.36,5.51,5.42
"NCNN 20230517 - Target: Vulkan GPU - Model: regnety_400m",
Lower Results Are Better
"a",14.57,14.74,15.03
"b",15.02,14.71,14.49
"c",14.33,14.82,15.16
"d",14.99,15.07,15.6
"NCNN 20230517 - Target: Vulkan GPU - Model: vision_transformer",
Lower Results Are Better
"a",30.99,31.91,31.67
"b",34.29,31.28,31.4
"c",31.14,33.11,31.52
"d",30.98,31.15,31.25
"NCNN 20230517 - Target: Vulkan GPU - Model: FastestDet",
Lower Results Are Better
"a",3.09,3.06,3.12
"b",3.13,3.15,3.03
"c",3.01,3.04,3.19
"d",3.09,3.09,3.17