dfhj Apple M2 testing with a Apple MacBook Air (13 h M2 2022) and llvmpipe on Arch rolling via the Phoronix Test Suite. a: Processor: Apple M2 @ 2.42GHz (4 Cores / 8 Threads), Motherboard: Apple MacBook Air (13 h M2 2022), Chipset: Apple Silicon, Memory: 8GB, Disk: 251GB APPLE SSD AP0256Z + 2 x 0GB APPLE SSD AP0256Z, Graphics: llvmpipe, Network: Broadcom Device 4433 + Broadcom BRCM4387 Bluetooth OS: Arch rolling, Kernel: 6.3.0-asahi-13-1-ARCH (aarch64), Desktop: KDE Plasma 5.27.6, Display Server: X Server 1.21.1.8, OpenGL: 4.5 Mesa 23.1.3 (LLVM 15.0.7 128 bits), Compiler: GCC 12.1.0 + Clang 15.0.7, File-System: ext4, Screen Resolution: 2560x1600 b: Processor: Apple M2 @ 2.42GHz (4 Cores / 8 Threads), Motherboard: Apple MacBook Air (13 h M2 2022), Chipset: Apple Silicon, Memory: 8GB, Disk: 251GB APPLE SSD AP0256Z + 2 x 0GB APPLE SSD AP0256Z, Graphics: llvmpipe, Network: Broadcom Device 4433 + Broadcom BRCM4387 Bluetooth OS: Arch rolling, Kernel: 6.3.0-asahi-13-1-ARCH (aarch64), Desktop: KDE Plasma 5.27.6, Display Server: X Server 1.21.1.8, OpenGL: 4.5 Mesa 23.1.3 (LLVM 15.0.7 128 bits), Compiler: GCC 12.1.0 + Clang 15.0.7, File-System: ext4, Screen Resolution: 2560x1600 c: Processor: Apple M2 @ 2.42GHz (4 Cores / 8 Threads), Motherboard: Apple MacBook Air (13 h M2 2022), Chipset: Apple Silicon, Memory: 8GB, Disk: 251GB APPLE SSD AP0256Z + 2 x 0GB APPLE SSD AP0256Z, Graphics: llvmpipe, Network: Broadcom Device 4433 + Broadcom BRCM4387 Bluetooth OS: Arch rolling, Kernel: 6.3.0-asahi-13-1-ARCH (aarch64), Desktop: KDE Plasma 5.27.6, Display Server: X Server 1.21.1.8, OpenGL: 4.5 Mesa 23.1.3 (LLVM 15.0.7 128 bits), Compiler: GCC 12.1.0 + Clang 15.0.7, File-System: ext4, Screen Resolution: 2560x1600 PyTorch 2.1 Device: CPU - Batch Size: 1 - Model: ResNet-50 batches/sec > Higher Is Better a . 6.30 |===================================================================== b . 6.26 |===================================================================== c . 6.29 |===================================================================== PyTorch 2.1 Device: CPU - Batch Size: 1 - Model: ResNet-152 batches/sec > Higher Is Better a . 3.10 |==================================================================== b . 3.15 |===================================================================== c . 3.11 |==================================================================== PyTorch 2.1 Device: CPU - Batch Size: 16 - Model: ResNet-50 batches/sec > Higher Is Better a . 4.79 |==================================================================== b . 4.82 |===================================================================== c . 4.85 |===================================================================== PyTorch 2.1 Device: CPU - Batch Size: 16 - Model: ResNet-152 batches/sec > Higher Is Better a . 2.32 |=================================================================== b . 2.39 |===================================================================== c . 2.28 |================================================================== PyTorch 2.1 Device: CPU - Batch Size: 1 - Model: Efficientnet_v2_l batches/sec > Higher Is Better a . 0.56 |==================================================================== b . 0.57 |===================================================================== c . 0.56 |==================================================================== PyTorch 2.1 Device: CPU - Batch Size: 16 - Model: Efficientnet_v2_l batches/sec > Higher Is Better a . 1.51 |===================================================================== b . 1.50 |===================================================================== c . 1.51 |===================================================================== Quicksilver 20230818 Input: CTS2 Figure Of Merit > Higher Is Better a . 4330000 |================================================================ b . 4310000 |=============================================================== c . 4500000 |================================================================== Quicksilver 20230818 Input: CORAL2 P1 Figure Of Merit > Higher Is Better a . 4832000 |=============================================================== b . 4791000 |============================================================== c . 5094000 |================================================================== Quicksilver 20230818 Input: CORAL2 P2 Figure Of Merit > Higher Is Better a . 7720000 |============================================================ b . 7686000 |============================================================ c . 8440000 |================================================================== rav1e 0.7 Speed: 1 Frames Per Second > Higher Is Better a . 0.553 |============================================================= b . 0.552 |============================================================= c . 0.612 |==================================================================== rav1e 0.7 Speed: 5 Frames Per Second > Higher Is Better a . 3.157 |================================================================ b . 3.140 |=============================================================== c . 3.372 |==================================================================== rav1e 0.7 Speed: 6 Frames Per Second > Higher Is Better a . 4.175 |================================================================ b . 4.197 |================================================================ c . 4.456 |==================================================================== rav1e 0.7 Speed: 10 Frames Per Second > Higher Is Better a . 11.80 |=================================================================== b . 11.90 |=================================================================== c . 12.02 |==================================================================== Xmrig 6.21 Variant: KawPow - Hash Count: 1M H/s > Higher Is Better a . 2232.7 |=========================================================== b . 2169.4 |========================================================== c . 2515.6 |=================================================================== Xmrig 6.21 Variant: Monero - Hash Count: 1M H/s > Higher Is Better a . 2168.2 |========================================================== b . 2128.4 |========================================================= c . 2484.2 |=================================================================== Xmrig 6.21 Variant: Wownero - Hash Count: 1M H/s > Higher Is Better a . 2437.7 |=========================================================== b . 2391.3 |========================================================== c . 2746.9 |=================================================================== Xmrig 6.21 Variant: GhostRider - Hash Count: 1M H/s > Higher Is Better a . 505.5 |==================================================================== b . 505.4 |==================================================================== c . 504.3 |==================================================================== Xmrig 6.21 Variant: CryptoNight-Heavy - Hash Count: 1M H/s > Higher Is Better a . 2319.2 |============================================================== b . 2302.6 |============================================================== c . 2494.0 |=================================================================== Xmrig 6.21 Variant: CryptoNight-Femto UPX2 - Hash Count: 1M H/s > Higher Is Better a . 2134.6 |======================================================== b . 2179.5 |========================================================= c . 2553.1 |=================================================================== Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 1.3721 |============================================================ b . 1.5268 |=================================================================== c . 1.5115 |================================================================== 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 . 54.87 |============================================================= b . 61.02 |==================================================================== c . 59.66 |================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 20.99 |============================================================== b . 23.18 |==================================================================== c . 22.71 |=================================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 190.65 |============================================================ b . 212.76 |=================================================================== c . 207.85 |================================================================= Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream items/sec > Higher Is Better a . 9.2739 |============================================================ b . 10.2386 |================================================================== c . 10.0390 |================================================================= LeelaChessZero 0.30 Backend: BLAS Nodes Per Second > Higher Is Better LeelaChessZero 0.30 Backend: Eigen Nodes Per Second > Higher Is Better a . 43 |===================================================================== b . 44 |======================================================================= c . 43 |===================================================================== Neural Magic DeepSparse 1.6 Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 1452.63 |================================================================== b . 1306.11 |=========================================================== c . 1320.49 |============================================================ 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 . 36.43 |==================================================================== b . 32.74 |============================================================= c . 33.50 |=============================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 95.22 |==================================================================== b . 86.21 |============================================================== c . 87.88 |=============================================================== Neural Magic DeepSparse 1.6 Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 10.4731 |================================================================== b . 9.3840 |=========================================================== c . 9.6063 |============================================================= Neural Magic DeepSparse 1.6 Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream ms/batch < Lower Is Better a . 215.41 |=================================================================== b . 194.82 |============================================================= c . 198.87 |==============================================================