ddf AMD EPYC 8534PN 64-Core testing with a AMD Cinnabar (RCB1009C BIOS) and ASPEED on Ubuntu 23.10 via the Phoronix Test Suite. a: Processor: AMD EPYC 8534PN 64-Core @ 2.00GHz (64 Cores / 128 Threads), Motherboard: AMD Cinnabar (RCB1009C BIOS), Chipset: AMD Device 14a4, Memory: 192GB, Disk: 3201GB Micron_7450_MTFDKCB3T2TFS, Graphics: ASPEED, Network: 2 x Broadcom NetXtreme BCM5720 PCIe OS: Ubuntu 23.10, Kernel: 6.5.0-5-generic (x86_64), Desktop: GNOME Shell, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 640x480 b: Processor: AMD EPYC 8534PN 64-Core @ 2.00GHz (64 Cores / 128 Threads), Motherboard: AMD Cinnabar (RCB1009C BIOS), Chipset: AMD Device 14a4, Memory: 192GB, Disk: 3201GB Micron_7450_MTFDKCB3T2TFS, Graphics: ASPEED, Network: 2 x Broadcom NetXtreme BCM5720 PCIe OS: Ubuntu 23.10, Kernel: 6.5.0-5-generic (x86_64), Desktop: GNOME Shell, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 640x480 c: Processor: AMD EPYC 8534PN 64-Core @ 2.00GHz (64 Cores / 128 Threads), Motherboard: AMD Cinnabar (RCB1009C BIOS), Chipset: AMD Device 14a4, Memory: 192GB, Disk: 3201GB Micron_7450_MTFDKCB3T2TFS, Graphics: ASPEED, Network: 2 x Broadcom NetXtreme BCM5720 PCIe OS: Ubuntu 23.10, Kernel: 6.5.0-5-generic (x86_64), Desktop: GNOME Shell, Display Server: X Server 1.21.1.7, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 640x480 LeelaChessZero 0.30 Backend: BLAS Nodes Per Second > Higher Is Better a . 315 |============================================================== b . 354 |====================================================================== c . 319 |=============================================================== WebP2 Image Encode 20220823 Encode Settings: Quality 75, Compression Effort 7 MP/s > Higher Is Better a . 0.54 |===================================================================== b . 0.51 |================================================================= c . 0.51 |================================================================= Timed Gem5 Compilation 23.0.1 Time To Compile Seconds < Lower Is Better a . 211.56 |================================================================ b . 223.06 |=================================================================== LeelaChessZero 0.30 Backend: Eigen Nodes Per Second > Higher Is Better a . 272 |==================================================================== b . 282 |====================================================================== c . 269 |=================================================================== SVT-AV1 1.8 Encoder Mode: Preset 13 - Input: Bosphorus 4K Frames Per Second > Higher Is Better a . 187.08 |================================================================ b . 194.59 |=================================================================== WebP2 Image Encode 20220823 Encode Settings: Quality 95, Compression Effort 7 MP/s > Higher Is Better a . 0.27 |===================================================================== b . 0.26 |================================================================== QuantLib 1.32 Configuration: Multi-Threaded MFLOPS > Higher Is Better a . 176928.1 |================================================================= b . 170381.3 |=============================================================== c . 176318.2 |================================================================= PyTorch 2.1 Device: CPU - Batch Size: 16 - Model: ResNet-152 batches/sec > Higher Is Better a . 14.70 |==================================================================== b . 14.18 |================================================================== CloverLeaf 1.3 Input: clover_bm Seconds < Lower Is Better a . 13.67 |=================================================================== b . 13.93 |==================================================================== c . 13.46 |================================================================== Xmrig 6.21 Variant: CryptoNight-Heavy - Hash Count: 1M H/s > Higher Is Better a . 20666.7 |================================================================== b . 20071.9 |================================================================ c . 20700.5 |================================================================== Speedb 2.7 Test: Update Random Op/s > Higher Is Better a . 350612 |================================================================= b . 361192 |=================================================================== SVT-AV1 1.8 Encoder Mode: Preset 4 - Input: Bosphorus 1080p Frames Per Second > Higher Is Better a . 17.08 |================================================================== b . 17.47 |==================================================================== OpenRadioss 2023.09.15 Model: Rubber O-Ring Seal Installation Seconds < Lower Is Better a . 76.78 |=================================================================== b . 76.10 |=================================================================== c . 77.76 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 195.60 |=================================================================== b . 191.48 |================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 5.1074 |================================================================== b . 5.2170 |=================================================================== Speedb 2.7 Test: Random Fill Sync Op/s > Higher Is Better a . 239469 |================================================================== b . 244535 |=================================================================== SVT-AV1 1.8 Encoder Mode: Preset 8 - Input: Bosphorus 1080p Frames Per Second > Higher Is Better a . 131.90 |=================================================================== b . 129.18 |================================================================== WebP2 Image Encode 20220823 Encode Settings: Default MP/s > Higher Is Better a . 7.44 |==================================================================== b . 7.49 |==================================================================== c . 7.59 |===================================================================== PyTorch 2.1 Device: CPU - Batch Size: 64 - Model: ResNet-152 batches/sec > Higher Is Better a . 14.89 |==================================================================== b . 14.61 |=================================================================== SVT-AV1 1.8 Encoder Mode: Preset 12 - Input: Bosphorus 4K Frames Per Second > Higher Is Better a . 190.95 |================================================================== b . 194.49 |=================================================================== SVT-AV1 1.8 Encoder Mode: Preset 4 - Input: Bosphorus 4K Frames Per Second > Higher Is Better a . 6.628 |=================================================================== b . 6.750 |==================================================================== PyTorch 2.1 Device: CPU - Batch Size: 32 - Model: ResNet-152 batches/sec > Higher Is Better a . 14.37 |==================================================================== b . 14.13 |=================================================================== SVT-AV1 1.8 Encoder Mode: Preset 12 - Input: Bosphorus 1080p Frames Per Second > Higher Is Better a . 511.85 |=================================================================== b . 503.62 |================================================================== PyTorch 2.1 Device: CPU - Batch Size: 64 - Model: ResNet-50 batches/sec > Higher Is Better a . 36.87 |==================================================================== b . 36.29 |=================================================================== OpenRadioss 2023.09.15 Model: Bumper Beam Seconds < Lower Is Better a . 88.11 |=================================================================== b . 88.16 |=================================================================== c . 89.19 |==================================================================== Embree 4.3 Binary: Pathtracer - Model: Crown Frames Per Second > Higher Is Better a . 67.93 |==================================================================== b . 67.11 |=================================================================== Speedb 2.7 Test: Read While Writing Op/s > Higher Is Better a . 15411684 |================================================================= b . 15231425 |================================================================ OpenRadioss 2023.09.15 Model: Bird Strike on Windshield Seconds < Lower Is Better a . 143.85 |=================================================================== b . 142.38 |================================================================== c . 142.76 |================================================================== Embree 4.3 Binary: Pathtracer ISPC - Model: Crown Frames Per Second > Higher Is Better a . 69.20 |==================================================================== b . 68.58 |=================================================================== FFmpeg 6.1 Encoder: libx265 - Scenario: Live FPS > Higher Is Better a . 114.74 |================================================================== b . 115.73 |=================================================================== c . 115.48 |=================================================================== PyTorch 2.1 Device: CPU - Batch Size: 1 - Model: Efficientnet_v2_l batches/sec > Higher Is Better a . 9.53 |==================================================================== b . 9.61 |===================================================================== rav1e 0.7 Speed: 6 Frames Per Second > Higher Is Better a . 4.851 |=================================================================== b . 4.891 |==================================================================== Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 5.3704 |================================================================== b . 5.4136 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 185.99 |=================================================================== b . 184.51 |================================================================== SVT-AV1 1.8 Encoder Mode: Preset 8 - Input: Bosphorus 4K Frames Per Second > Higher Is Better a . 67.98 |=================================================================== b . 68.52 |==================================================================== PyTorch 2.1 Device: CPU - Batch Size: 32 - Model: ResNet-50 batches/sec > Higher Is Better a . 36.66 |==================================================================== b . 36.38 |=================================================================== OpenRadioss 2023.09.15 Model: INIVOL and Fluid Structure Interaction Drop Container Seconds < Lower Is Better a . 164.67 |=================================================================== b . 163.46 |=================================================================== c . 163.42 |================================================================== SVT-AV1 1.8 Encoder Mode: Preset 13 - Input: Bosphorus 1080p Frames Per Second > Higher Is Better a . 597.06 |================================================================== b . 601.57 |=================================================================== Speedb 2.7 Test: Sequential Fill Op/s > Higher Is Better a . 371857 |=================================================================== b . 369178 |=================================================================== Embree 4.3 Binary: Pathtracer - Model: Asian Dragon Obj Frames Per Second > Higher Is Better a . 69.09 |==================================================================== b . 68.59 |==================================================================== Blender 4.0 Blend File: Classroom - Compute: CPU-Only Seconds < Lower Is Better a . 67.06 |==================================================================== b . 67.51 |==================================================================== PyTorch 2.1 Device: CPU - Batch Size: 16 - Model: ResNet-50 batches/sec > Higher Is Better a . 36.12 |==================================================================== b . 36.35 |==================================================================== Blender 4.0 Blend File: Fishy Cat - Compute: CPU-Only Seconds < Lower Is Better a . 35.54 |==================================================================== b . 35.33 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 36.81 |==================================================================== b . 37.02 |==================================================================== Timed FFmpeg Compilation 6.1 Time To Compile Seconds < Lower Is Better a . 18.15 |==================================================================== b . 18.04 |==================================================================== easyWave r34 Input: e2Asean Grid + BengkuluSept2007 Source - Time: 240 Seconds < Lower Is Better a . 1.947 |==================================================================== b . 1.958 |==================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 32.97 |==================================================================== b . 33.14 |==================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 30.32 |==================================================================== b . 30.16 |==================================================================== TensorFlow 2.12 Device: CPU - Batch Size: 16 - Model: ResNet-50 images/sec > Higher Is Better a . 51.06 |==================================================================== b . 50.80 |==================================================================== PyTorch 2.1 Device: CPU - Batch Size: 1 - Model: ResNet-50 batches/sec > Higher Is Better a . 45.19 |==================================================================== b . 45.42 |==================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 185.30 |=================================================================== b . 186.24 |=================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 5.3901 |=================================================================== b . 5.3631 |=================================================================== OpenRadioss 2023.09.15 Model: Chrysler Neon 1M Seconds < Lower Is Better a . 297.08 |=================================================================== b . 295.72 |=================================================================== c . 296.96 |=================================================================== Xmrig 6.21 Variant: Wownero - Hash Count: 1M H/s > Higher Is Better a . 40330.7 |================================================================== b . 40146.1 |================================================================== c . 40251.2 |================================================================== TensorFlow 2.12 Device: CPU - Batch Size: 16 - Model: AlexNet images/sec > Higher Is Better a . 299.15 |=================================================================== b . 297.84 |=================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 698.36 |=================================================================== b . 695.35 |=================================================================== Quicksilver 20230818 Input: CTS2 Figure Of Merit > Higher Is Better a . 16240000 |================================================================= b . 16310000 |================================================================= c . 16240000 |================================================================= WebP2 Image Encode 20220823 Encode Settings: Quality 100, Compression Effort 5 MP/s > Higher Is Better a . 11.63 |==================================================================== b . 11.68 |==================================================================== Y-Cruncher 0.8.3 Pi Digits To Calculate: 1B Seconds < Lower Is Better a . 10.25 |==================================================================== b . 10.20 |==================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 45.76 |==================================================================== b . 45.95 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 130.92 |=================================================================== b . 131.46 |=================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 7.6305 |=================================================================== b . 7.5994 |=================================================================== easyWave r34 Input: e2Asean Grid + BengkuluSept2007 Source - Time: 1200 Seconds < Lower Is Better a . 39.64 |==================================================================== b . 39.48 |==================================================================== FFmpeg 6.1 Encoder: libx265 - Scenario: Upload FPS > Higher Is Better a . 23.20 |==================================================================== b . 23.29 |==================================================================== c . 23.23 |==================================================================== Speedb 2.7 Test: Random Fill Op/s > Higher Is Better a . 369023 |=================================================================== b . 367684 |=================================================================== Embree 4.3 Binary: Pathtracer ISPC - Model: Asian Dragon Frames Per Second > Higher Is Better a . 83.88 |==================================================================== b . 83.58 |==================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 3813.64 |================================================================== b . 3800.68 |================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 8.3709 |=================================================================== b . 8.3990 |=================================================================== rav1e 0.7 Speed: 5 Frames Per Second > Higher Is Better a . 3.584 |==================================================================== b . 3.596 |==================================================================== PyTorch 2.1 Device: CPU - Batch Size: 16 - Model: Efficientnet_v2_l batches/sec > Higher Is Better a . 6.00 |===================================================================== b . 5.98 |===================================================================== PyTorch 2.1 Device: CPU - Batch Size: 32 - Model: Efficientnet_v2_l batches/sec > Higher Is Better a . 6.06 |===================================================================== b . 6.08 |===================================================================== Quicksilver 20230818 Input: CORAL2 P1 Figure Of Merit > Higher Is Better a . 21350000 |================================================================= b . 21280000 |================================================================= c . 21290000 |================================================================= Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 145.62 |=================================================================== b . 145.15 |=================================================================== rav1e 0.7 Speed: 10 Frames Per Second > Higher Is Better a . 12.32 |==================================================================== b . 12.36 |==================================================================== Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 6.8554 |=================================================================== b . 6.8775 |=================================================================== OpenRadioss 2023.09.15 Model: Cell Phone Drop Test Seconds < Lower Is Better a . 31.94 |==================================================================== b . 31.84 |==================================================================== c . 31.91 |==================================================================== Xmrig 6.21 Variant: CryptoNight-Femto UPX2 - Hash Count: 1M H/s > Higher Is Better a . 20698.4 |================================================================== b . 20683.0 |================================================================== c . 20737.0 |================================================================== Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 496.34 |=================================================================== b . 497.62 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 64.07 |==================================================================== b . 63.91 |==================================================================== Xmrig 6.21 Variant: GhostRider - Hash Count: 1M H/s > Higher Is Better a . 4436.2 |=================================================================== b . 4442.9 |=================================================================== c . 4431.9 |=================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 673.33 |=================================================================== b . 674.99 |=================================================================== Xmrig 6.21 Variant: KawPow - Hash Count: 1M H/s > Higher Is Better a . 20682.1 |================================================================== b . 20732.7 |================================================================== c . 20694.5 |================================================================== PyTorch 2.1 Device: CPU - Batch Size: 1 - Model: ResNet-152 batches/sec > Higher Is Better a . 16.54 |==================================================================== b . 16.58 |==================================================================== Y-Cruncher 0.8.3 Pi Digits To Calculate: 10B Seconds < Lower Is Better a . 112.54 |=================================================================== b . 112.81 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 222.25 |=================================================================== b . 221.72 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 143.53 |=================================================================== b . 143.84 |=================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 46.96 |==================================================================== b . 46.86 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 1458.05 |================================================================== b . 1454.99 |================================================================== CloverLeaf 1.3 Input: clover_bm64_short Seconds < Lower Is Better a . 57.21 |==================================================================== b . 57.09 |==================================================================== c . 57.11 |==================================================================== Embree 4.3 Binary: Pathtracer - Model: Asian Dragon Frames Per Second > Higher Is Better a . 77.26 |==================================================================== b . 77.10 |==================================================================== TensorFlow 2.12 Device: CPU - Batch Size: 16 - Model: VGG-16 images/sec > Higher Is Better a . 35.21 |==================================================================== b . 35.14 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 37.09 |==================================================================== b . 37.16 |==================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 65.66 |==================================================================== b . 65.78 |==================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 486.60 |=================================================================== b . 485.73 |=================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 21.92 |==================================================================== b . 21.96 |==================================================================== FFmpeg 6.1 Encoder: libx265 - Scenario: Video On Demand FPS > Higher Is Better a . 47.03 |==================================================================== b . 47.05 |==================================================================== c . 46.97 |==================================================================== FFmpeg 6.1 Encoder: libx265 - Scenario: Platform FPS > Higher Is Better a . 47.15 |==================================================================== b . 47.07 |==================================================================== c . 47.07 |==================================================================== Blender 4.0 Blend File: BMW27 - Compute: CPU-Only Seconds < Lower Is Better a . 26.71 |==================================================================== b . 26.75 |==================================================================== Y-Cruncher 0.8.3 Pi Digits To Calculate: 5B Seconds < Lower Is Better a . 53.07 |==================================================================== b . 53.15 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 852.18 |=================================================================== b . 853.31 |=================================================================== Quicksilver 20230818 Input: CORAL2 P2 Figure Of Merit > Higher Is Better a . 16170000 |================================================================= b . 16190000 |================================================================= c . 16170000 |================================================================= Blender 4.0 Blend File: Barbershop - Compute: CPU-Only Seconds < Lower Is Better a . 239.83 |=================================================================== b . 240.12 |=================================================================== Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 27.57 |==================================================================== b . 27.54 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 36.26 |==================================================================== b . 36.30 |==================================================================== QuantLib 1.32 Configuration: Single-Threaded MFLOPS > Higher Is Better a . 2634.4 |=================================================================== b . 2633.8 |=================================================================== c . 2631.7 |=================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 15.78 |==================================================================== b . 15.77 |==================================================================== Speedb 2.7 Test: Random Read Op/s > Higher Is Better a . 303866048 |================================================================ b . 304143127 |================================================================ Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 36.35 |==================================================================== b . 36.32 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 27.50 |==================================================================== b . 27.53 |==================================================================== Neural Magic DeepSparse 1.6 Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 63.29 |==================================================================== b . 63.34 |==================================================================== Xmrig 6.21 Variant: Monero - Hash Count: 1M H/s > Higher Is Better a . 20714.7 |================================================================== b . 20732.3 |================================================================== c . 20722.8 |================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 800.59 |=================================================================== b . 799.96 |=================================================================== Y-Cruncher 0.8.3 Pi Digits To Calculate: 500M Seconds < Lower Is Better a . 5.110 |==================================================================== b . 5.106 |==================================================================== Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 145.02 |=================================================================== b . 144.91 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 65.69 |==================================================================== b . 65.74 |==================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 1.2456 |=================================================================== b . 1.2465 |=================================================================== Embree 4.3 Binary: Pathtracer ISPC - Model: Asian Dragon Obj Frames Per Second > Higher Is Better a . 71.49 |==================================================================== b . 71.44 |==================================================================== Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 6.7843 |=================================================================== b . 6.7891 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 147.25 |=================================================================== b . 147.15 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream ms/batch < Lower Is Better a . 24.36 |==================================================================== b . 24.37 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 854.12 |=================================================================== b . 853.69 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream items/sec > Higher Is Better a . 41.01 |==================================================================== b . 40.99 |==================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 323.07 |=================================================================== b . 322.92 |=================================================================== Neural Magic DeepSparse 1.6 Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 98.82 |==================================================================== b . 98.86 |==================================================================== Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 220.07 |=================================================================== b . 220.15 |=================================================================== TensorFlow 2.12 Device: CPU - Batch Size: 1 - Model: AlexNet images/sec > Higher Is Better a . 30.46 |==================================================================== b . 30.45 |==================================================================== easyWave r34 Input: e2Asean Grid + BengkuluSept2007 Source - Time: 2400 Seconds < Lower Is Better a . 111.04 |=================================================================== b . 111.07 |=================================================================== Neural Magic DeepSparse 1.6 Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 486.00 |=================================================================== b . 485.88 |=================================================================== Blender 4.0 Blend File: Pabellon Barcelona - Compute: CPU-Only Seconds < Lower Is Better a . 86.14 |==================================================================== b . 86.16 |==================================================================== Speedb 2.7 Test: Read Random Write Random Op/s > Higher Is Better a . 2539184 |================================================================== b . 2539687 |================================================================== TensorFlow 2.12 Device: CPU - Batch Size: 16 - Model: GoogLeNet images/sec > Higher Is Better a . 155.77 |=================================================================== b . 155.78 |=================================================================== TensorFlow 2.12 Device: CPU - Batch Size: 1 - Model: ResNet-50 images/sec > Higher Is Better a . 5.97 |===================================================================== b . 5.97 |===================================================================== TensorFlow 2.12 Device: CPU - Batch Size: 1 - Model: GoogLeNet images/sec > Higher Is Better a . 17 |======================================================================= b . 17 |======================================================================= TensorFlow 2.12 Device: CPU - Batch Size: 1 - Model: VGG-16 images/sec > Higher Is Better a . 9.87 |===================================================================== b . 9.87 |===================================================================== PyTorch 2.1 Device: CPU - Batch Size: 64 - Model: Efficientnet_v2_l batches/sec > Higher Is Better a . 6.03 |===================================================================== b . 6.03 |===================================================================== rav1e 0.7 Speed: 1 Frames Per Second > Higher Is Better a . 0.85 |===================================================================== b . 0.85 |===================================================================== WebP2 Image Encode 20220823 Encode Settings: Quality 100, Lossless Compression MP/s > Higher Is Better a . 0.06 |===================================================================== b . 0.06 |=====================================================================