ds AMD Ryzen 9 7950X 16-Core testing with a ASUS ROG STRIX X670E-E GAMING WIFI (1416 BIOS) and NVIDIA NV174 8GB on Ubuntu 23.10 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2312110-PTS-DS58174320&gru&sor .
ds Processor Motherboard Chipset Memory Disk Graphics Audio Monitor Network OS Kernel Desktop Display Server Display Driver OpenGL Compiler File-System Screen Resolution a b c d AMD Ryzen 9 7950X 16-Core @ 5.88GHz (16 Cores / 32 Threads) ASUS ROG STRIX X670E-E GAMING WIFI (1416 BIOS) AMD Device 14d8 32GB 2000GB Samsung SSD 980 PRO 2TB + 4001GB Western Digital WD_BLACK SN850X 4000GB NVIDIA NV174 8GB NVIDIA GA104 HD Audio DELL U2723QE Intel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411 Ubuntu 23.10 6.7.0-060700rc2daily20231127-generic (x86_64) GNOME Shell 45.1 X Server 1.21.1.7 + Wayland nouveau 4.3 Mesa 24.0~git2311260600.945288~oibaf~m (git-945288f 2023-11-26 mantic-oibaf-ppa) GCC 13.2.0 + LLVM 16.0.6 ext4 3840x2160 OpenBenchmarking.org Kernel Details - Transparent Huge Pages: madvise Processor Details - Scaling Governor: amd-pstate-epp powersave (EPP: balance_performance) - CPU Microcode: 0xa601203 Python Details - Python 3.11.6 Security Details - gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Vulnerable: Safe RET no microcode + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
ds deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Asynchronous Multi-Stream deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Synchronous Single-Stream deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Stream deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Synchronous Single-Stream deepsparse: ResNet-50, Baseline - Asynchronous Multi-Stream deepsparse: ResNet-50, Baseline - Synchronous Single-Stream deepsparse: ResNet-50, Sparse INT8 - Asynchronous Multi-Stream deepsparse: ResNet-50, Sparse INT8 - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO - Synchronous Single-Stream deepsparse: BERT-Large, NLP Question Answering - Asynchronous Multi-Stream deepsparse: BERT-Large, NLP Question Answering - Synchronous Single-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Synchronous Single-Stream deepsparse: NLP Text Classification, DistilBERT mnli - Asynchronous Multi-Stream deepsparse: NLP Text Classification, DistilBERT mnli - Synchronous Single-Stream deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Asynchronous Multi-Stream deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Synchronous Single-Stream deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Asynchronous Multi-Stream deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Synchronous Single-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Asynchronous Multi-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Synchronous Single-Stream deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Asynchronous Multi-Stream deepsparse: NLP Document Classification, oBERT base uncased on IMDB - Synchronous Single-Stream deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Asynchronous Multi-Stream deepsparse: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Synchronous Single-Stream deepsparse: ResNet-50, Baseline - Asynchronous Multi-Stream deepsparse: ResNet-50, Baseline - Synchronous Single-Stream deepsparse: ResNet-50, Sparse INT8 - Asynchronous Multi-Stream deepsparse: ResNet-50, Sparse INT8 - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO - Synchronous Single-Stream deepsparse: BERT-Large, NLP Question Answering - Asynchronous Multi-Stream deepsparse: BERT-Large, NLP Question Answering - Synchronous Single-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Asynchronous Multi-Stream deepsparse: CV Classification, ResNet-50 ImageNet - Synchronous Single-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Asynchronous Multi-Stream deepsparse: CV Detection, YOLOv5s COCO, Sparse INT8 - Synchronous Single-Stream deepsparse: NLP Text Classification, DistilBERT mnli - Asynchronous Multi-Stream deepsparse: NLP Text Classification, DistilBERT mnli - Synchronous Single-Stream deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Asynchronous Multi-Stream deepsparse: CV Segmentation, 90% Pruned YOLACT Pruned - Synchronous Single-Stream deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Asynchronous Multi-Stream deepsparse: BERT-Large, NLP Question Answering, Sparse INT8 - Synchronous Single-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Asynchronous Multi-Stream deepsparse: NLP Token Classification, BERT base uncased conll2003 - Synchronous Single-Stream a b c d 22.7389 19.5685 983.3126 299.9092 296.0378 197.0798 2314.2272 1295.8078 127.7126 101.8661 29.3517 21.1279 296.0322 198.8967 133.5376 103.1665 203.9666 116.3673 38.9835 32.3888 462.4186 110.6228 22.7592 19.5417 350.7492 51.0955 8.1243 3.3312 27.0091 5.0670 3.4471 0.7695 62.6288 9.8076 272.1665 47.3214 27.0107 5.0217 59.8641 9.6870 39.2065 8.5884 204.8572 30.8613 17.2878 9.0307 349.9757 51.1656 22.6899 19.5391 988.2981 301.5141 295.5210 198.4779 2315.5240 1295.2310 127.2462 101.7672 29.3894 21.1400 295.7606 198.4507 133.1707 103.1635 203.4893 116.5816 38.8700 32.4189 462.6802 110.0314 22.7193 19.5292 351.5559 51.1690 8.0835 3.3138 27.0574 5.0320 3.4442 0.7698 62.8602 9.8180 271.9143 47.2956 27.0344 5.0331 60.0283 9.6883 39.2941 8.5729 205.5998 30.8326 17.2779 9.0802 351.5076 51.1975 22.7000 19.5454 986.6769 301.1738 295.7285 198.2479 2317.2555 1308.7670 127.4662 101.8841 29.3969 21.1302 295.2366 198.7653 132.0251 103.3168 203.9185 116.7136 39.0428 32.4022 461.5344 110.0550 22.6850 19.5308 351.4011 51.1551 8.0963 3.3173 27.0386 5.0382 3.4420 0.7620 62.7323 9.8068 271.7780 47.3165 27.0809 5.0250 60.5733 9.6736 39.2164 8.5631 204.5755 30.8484 17.3189 9.0770 351.4431 51.1932 22.6176 19.5678 987.4091 301.6484 295.3355 198.3620 2293.2480 1299.0909 128.0169 101.4802 29.3591 21.2410 295.5731 198.9679 132.5933 103.2328 203.6193 117.2571 39.1305 32.4054 462.4074 110.4956 22.7096 19.5097 352.4834 51.0971 8.0908 3.3119 27.0742 5.0356 3.4780 0.7672 62.4691 9.8453 272.1527 47.0684 27.0539 5.0198 60.2946 9.6809 39.2783 8.5234 204.2205 30.8462 17.2882 9.0414 351.4079 51.2514 OpenBenchmarking.org
Neural Magic DeepSparse Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream a c b d 5 10 15 20 25 SE +/- 0.03, N = 3 SE +/- 0.08, N = 3 SE +/- 0.03, N = 3 SE +/- 0.03, N = 3 22.74 22.70 22.69 22.62
Neural Magic DeepSparse Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream a d c b 5 10 15 20 25 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 19.57 19.57 19.55 19.54
Neural Magic DeepSparse Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream b d c a 200 400 600 800 1000 SE +/- 0.99, N = 3 SE +/- 1.44, N = 3 SE +/- 1.16, N = 3 SE +/- 0.48, N = 3 988.30 987.41 986.68 983.31
Neural Magic DeepSparse Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream d b c a 70 140 210 280 350 SE +/- 0.32, N = 3 SE +/- 0.53, N = 3 SE +/- 1.04, N = 3 SE +/- 0.35, N = 3 301.65 301.51 301.17 299.91
Neural Magic DeepSparse Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream a c b d 60 120 180 240 300 SE +/- 0.11, N = 3 SE +/- 0.15, N = 3 SE +/- 0.12, N = 3 SE +/- 0.10, N = 3 296.04 295.73 295.52 295.34
Neural Magic DeepSparse Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream b d c a 40 80 120 160 200 SE +/- 0.91, N = 3 SE +/- 0.35, N = 3 SE +/- 0.03, N = 3 SE +/- 0.82, N = 3 198.48 198.36 198.25 197.08
Neural Magic DeepSparse Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream c b a d 500 1000 1500 2000 2500 SE +/- 4.26, N = 3 SE +/- 6.99, N = 3 SE +/- 1.66, N = 3 SE +/- 8.04, N = 3 2317.26 2315.52 2314.23 2293.25
Neural Magic DeepSparse Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream c d a b 300 600 900 1200 1500 SE +/- 1.41, N = 3 SE +/- 0.19, N = 3 SE +/- 10.93, N = 3 SE +/- 7.10, N = 3 1308.77 1299.09 1295.81 1295.23
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream d a c b 30 60 90 120 150 SE +/- 0.24, N = 3 SE +/- 0.61, N = 3 SE +/- 0.72, N = 3 SE +/- 0.56, N = 3 128.02 127.71 127.47 127.25
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream c a b d 20 40 60 80 100 SE +/- 0.04, N = 3 SE +/- 0.06, N = 3 SE +/- 0.06, N = 3 SE +/- 0.19, N = 3 101.88 101.87 101.77 101.48
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream c b d a 7 14 21 28 35 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 29.40 29.39 29.36 29.35
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream d b c a 5 10 15 20 25 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 21.24 21.14 21.13 21.13
Neural Magic DeepSparse Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream a b d c 60 120 180 240 300 SE +/- 0.22, N = 3 SE +/- 0.22, N = 3 SE +/- 0.38, N = 3 SE +/- 0.26, N = 3 296.03 295.76 295.57 295.24
Neural Magic DeepSparse Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream d a c b 40 80 120 160 200 SE +/- 0.45, N = 3 SE +/- 0.11, N = 3 SE +/- 0.17, N = 3 SE +/- 0.31, N = 3 198.97 198.90 198.77 198.45
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream a b d c 30 60 90 120 150 SE +/- 0.30, N = 3 SE +/- 0.59, N = 3 SE +/- 0.43, N = 3 SE +/- 0.45, N = 3 133.54 133.17 132.59 132.03
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream c d a b 20 40 60 80 100 SE +/- 0.00, N = 3 SE +/- 0.14, N = 3 SE +/- 0.06, N = 3 SE +/- 0.06, N = 3 103.32 103.23 103.17 103.16
Neural Magic DeepSparse Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream a c d b 40 80 120 160 200 SE +/- 0.15, N = 3 SE +/- 0.24, N = 3 SE +/- 0.23, N = 3 SE +/- 0.06, N = 3 203.97 203.92 203.62 203.49
Neural Magic DeepSparse Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream d c b a 30 60 90 120 150 SE +/- 0.43, N = 3 SE +/- 0.84, N = 3 SE +/- 0.80, N = 3 SE +/- 0.35, N = 3 117.26 116.71 116.58 116.37
Neural Magic DeepSparse Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream d c a b 9 18 27 36 45 SE +/- 0.13, N = 3 SE +/- 0.07, N = 3 SE +/- 0.08, N = 3 SE +/- 0.36, N = 3 39.13 39.04 38.98 38.87
Neural Magic DeepSparse Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream b d c a 8 16 24 32 40 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 32.42 32.41 32.40 32.39
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream b a d c 100 200 300 400 500 SE +/- 0.43, N = 3 SE +/- 0.23, N = 3 SE +/- 0.35, N = 3 SE +/- 0.34, N = 3 462.68 462.42 462.41 461.53
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream a d c b 20 40 60 80 100 SE +/- 0.42, N = 3 SE +/- 0.78, N = 3 SE +/- 0.22, N = 3 SE +/- 0.94, N = 3 110.62 110.50 110.06 110.03
Neural Magic DeepSparse Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream a b d c 5 10 15 20 25 SE +/- 0.07, N = 3 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 22.76 22.72 22.71 22.69
Neural Magic DeepSparse Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream OpenBenchmarking.org items/sec, More Is Better Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream a c b d 5 10 15 20 25 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 19.54 19.53 19.53 19.51
Neural Magic DeepSparse Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream a c b d 80 160 240 320 400 SE +/- 0.16, N = 3 SE +/- 0.75, N = 3 SE +/- 0.09, N = 3 SE +/- 0.20, N = 3 350.75 351.40 351.56 352.48
Neural Magic DeepSparse Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream a d c b 12 24 36 48 60 SE +/- 0.06, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 51.10 51.10 51.16 51.17
Neural Magic DeepSparse Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream b d c a 2 4 6 8 10 SE +/- 0.0081, N = 3 SE +/- 0.0118, N = 3 SE +/- 0.0093, N = 3 SE +/- 0.0039, N = 3 8.0835 8.0908 8.0963 8.1243
Neural Magic DeepSparse Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream d b c a 0.7495 1.499 2.2485 2.998 3.7475 SE +/- 0.0036, N = 3 SE +/- 0.0056, N = 3 SE +/- 0.0115, N = 3 SE +/- 0.0038, N = 3 3.3119 3.3138 3.3173 3.3312
Neural Magic DeepSparse Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream a c b d 6 12 18 24 30 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 27.01 27.04 27.06 27.07
Neural Magic DeepSparse Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream b d c a 1.1401 2.2802 3.4203 4.5604 5.7005 SE +/- 0.0228, N = 3 SE +/- 0.0087, N = 3 SE +/- 0.0010, N = 3 SE +/- 0.0211, N = 3 5.0320 5.0356 5.0382 5.0670
Neural Magic DeepSparse Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream c b a d 0.7826 1.5652 2.3478 3.1304 3.913 SE +/- 0.0064, N = 3 SE +/- 0.0106, N = 3 SE +/- 0.0021, N = 3 SE +/- 0.0122, N = 3 3.4420 3.4442 3.4471 3.4780
Neural Magic DeepSparse Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream c d a b 0.1732 0.3464 0.5196 0.6928 0.866 SE +/- 0.0008, N = 3 SE +/- 0.0001, N = 3 SE +/- 0.0065, N = 3 SE +/- 0.0041, N = 3 0.7620 0.7672 0.7695 0.7698
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream d a c b 14 28 42 56 70 SE +/- 0.12, N = 3 SE +/- 0.30, N = 3 SE +/- 0.37, N = 3 SE +/- 0.28, N = 3 62.47 62.63 62.73 62.86
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream c a b d 3 6 9 12 15 SE +/- 0.0035, N = 3 SE +/- 0.0059, N = 3 SE +/- 0.0053, N = 3 SE +/- 0.0182, N = 3 9.8068 9.8076 9.8180 9.8453
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream c b d a 60 120 180 240 300 SE +/- 0.09, N = 3 SE +/- 0.13, N = 3 SE +/- 0.09, N = 3 SE +/- 0.01, N = 3 271.78 271.91 272.15 272.17
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream d b c a 11 22 33 44 55 SE +/- 0.01, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.03, N = 3 47.07 47.30 47.32 47.32
Neural Magic DeepSparse Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream a b d c 6 12 18 24 30 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 27.01 27.03 27.05 27.08
Neural Magic DeepSparse Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream d a c b 1.1324 2.2648 3.3972 4.5296 5.662 SE +/- 0.0116, N = 3 SE +/- 0.0028, N = 3 SE +/- 0.0044, N = 3 SE +/- 0.0078, N = 3 5.0198 5.0217 5.0250 5.0331
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream a b d c 14 28 42 56 70 SE +/- 0.13, N = 3 SE +/- 0.27, N = 3 SE +/- 0.21, N = 3 SE +/- 0.21, N = 3 59.86 60.03 60.29 60.57
Neural Magic DeepSparse Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream c d a b 3 6 9 12 15 SE +/- 0.0004, N = 3 SE +/- 0.0125, N = 3 SE +/- 0.0055, N = 3 SE +/- 0.0056, N = 3 9.6736 9.6809 9.6870 9.6883
Neural Magic DeepSparse Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream a c d b 9 18 27 36 45 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.01, N = 3 39.21 39.22 39.28 39.29
Neural Magic DeepSparse Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream d c b a 2 4 6 8 10 SE +/- 0.0316, N = 3 SE +/- 0.0624, N = 3 SE +/- 0.0593, N = 3 SE +/- 0.0256, N = 3 8.5234 8.5631 8.5729 8.5884
Neural Magic DeepSparse Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream d c a b 50 100 150 200 250 SE +/- 0.62, N = 3 SE +/- 0.32, N = 3 SE +/- 0.47, N = 3 SE +/- 1.77, N = 3 204.22 204.58 204.86 205.60
Neural Magic DeepSparse Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream b d c a 7 14 21 28 35 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 30.83 30.85 30.85 30.86
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream b a d c 4 8 12 16 20 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 17.28 17.29 17.29 17.32
Neural Magic DeepSparse Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream a d c b 3 6 9 12 15 SE +/- 0.0337, N = 3 SE +/- 0.0640, N = 3 SE +/- 0.0179, N = 3 SE +/- 0.0759, N = 3 9.0307 9.0414 9.0770 9.0802
Neural Magic DeepSparse Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream a d c b 80 160 240 320 400 SE +/- 1.12, N = 3 SE +/- 0.17, N = 3 SE +/- 0.11, N = 3 SE +/- 0.23, N = 3 349.98 351.41 351.44 351.51
Neural Magic DeepSparse Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream OpenBenchmarking.org ms/batch, Fewer Is Better Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream a c b d 12 24 36 48 60 SE +/- 0.02, N = 3 SE +/- 0.07, N = 3 SE +/- 0.03, N = 3 SE +/- 0.08, N = 3 51.17 51.19 51.20 51.25
Phoronix Test Suite v10.8.5