onednn tgl Intel Core i5-1145G7 testing with a LENOVO 20XW004AUS (N32ET71W 1.47 BIOS) and Intel Xe TGL GT2 3GB on Ubuntu 20.04 via the Phoronix Test Suite. A: Processor: Intel Core i5-1145G7 @ 4.40GHz (4 Cores / 8 Threads), Motherboard: LENOVO 20XW004AUS (N32ET71W 1.47 BIOS), Chipset: Intel Device a0ef, Memory: 16GB, Disk: 1024GB SAMSUNG MZVLB1T0HBLR-000H1, Graphics: Intel Xe TGL GT2 3GB (1300MHz), Audio: Realtek ALC287, Network: Intel Device a0f0 OS: Ubuntu 20.04, Kernel: 5.14.0-1027-oem (x86_64), Desktop: GNOME Shell 3.36.9, Display Server: X Server 1.20.13, OpenGL: 4.6 Mesa 21.2.6, Vulkan: 1.2.182, Compiler: GCC 9.4.0, File-System: ext4, Screen Resolution: 1920x1200 B: Processor: Intel Core i5-1145G7 @ 4.40GHz (4 Cores / 8 Threads), Motherboard: LENOVO 20XW004AUS (N32ET71W 1.47 BIOS), Chipset: Intel Device a0ef, Memory: 16GB, Disk: 1024GB SAMSUNG MZVLB1T0HBLR-000H1, Graphics: Intel Xe TGL GT2 3GB (1300MHz), Audio: Realtek ALC287, Network: Intel Device a0f0 OS: Ubuntu 20.04, Kernel: 5.14.0-1027-oem (x86_64), Desktop: GNOME Shell 3.36.9, Display Server: X Server 1.20.13, OpenGL: 4.6 Mesa 21.2.6, Vulkan: 1.2.182, Compiler: GCC 9.4.0, File-System: ext4, Screen Resolution: 1920x1200 C: Processor: Intel Core i5-1145G7 @ 4.40GHz (4 Cores / 8 Threads), Motherboard: LENOVO 20XW004AUS (N32ET71W 1.47 BIOS), Chipset: Intel Device a0ef, Memory: 16GB, Disk: 1024GB SAMSUNG MZVLB1T0HBLR-000H1, Graphics: Intel Xe TGL GT2 3GB (1300MHz), Audio: Realtek ALC287, Network: Intel Device a0f0 OS: Ubuntu 20.04, Kernel: 5.14.0-1027-oem (x86_64), Desktop: GNOME Shell 3.36.9, Display Server: X Server 1.20.13, OpenGL: 4.6 Mesa 21.2.6, Vulkan: 1.2.182, Compiler: GCC 9.4.0, File-System: ext4, Screen Resolution: 1920x1200 oneDNN 2.6 Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 8.46208 |=========================================================== B . 8.95669 |============================================================== C . 9.51734 |================================================================== oneDNN 2.6 Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 5.40966 |================================================================= B . 5.40629 |================================================================= C . 5.45216 |================================================================== oneDNN 2.6 Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 1.95301 |=========================================================== B . 2.08277 |=============================================================== C . 2.16909 |================================================================== oneDNN 2.6 Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 2.11018 |============================================================= B . 2.28842 |================================================================== C . 2.29786 |================================================================== oneDNN 2.6 Harness: IP Shapes 1D - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 19.84 |==================================================================== B . 19.84 |==================================================================== C . 19.80 |==================================================================== oneDNN 2.6 Harness: IP Shapes 3D - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 6.11673 |================================================================ B . 6.31136 |================================================================== C . 6.35160 |================================================================== oneDNN 2.6 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 9.41308 |========================================================== B . 9.99736 |============================================================== C . 10.51531 |================================================================= oneDNN 2.6 Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 17.70 |================================================================== B . 18.30 |==================================================================== C . 18.21 |==================================================================== oneDNN 2.6 Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 10.17120 |================================================================= B . 9.93060 |=============================================================== C . 9.82487 |=============================================================== oneDNN 2.6 Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 7.00896 |============================================================= B . 7.61626 |================================================================== C . 7.63750 |================================================================== oneDNN 2.6 Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 3.00906 |================================================================= B . 3.03781 |================================================================= C . 3.07591 |================================================================== oneDNN 2.6 Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 2.41032 |================================================================== B . 2.35766 |================================================================= C . 2.34011 |================================================================ oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 9343.97 |================================================================== B . 9342.43 |================================================================== C . 9341.05 |================================================================== oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 4767.33 |================================================================== B . 4768.46 |================================================================== C . 4759.63 |================================================================== oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 9344.99 |================================================================== B . 9347.91 |================================================================== C . 9346.82 |================================================================== oneDNN 2.6 Harness: Convolution Batch Shapes Auto - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 38.04 |==================================================================== B . 37.97 |==================================================================== C . 37.96 |==================================================================== oneDNN 2.6 Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 63.56 |=================================================================== B . 63.93 |=================================================================== C . 64.46 |==================================================================== oneDNN 2.6 Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 38.27 |==================================================================== B . 38.31 |==================================================================== C . 38.40 |==================================================================== oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 4762.61 |================================================================== B . 4762.94 |================================================================== C . 4763.90 |================================================================== oneDNN 2.6 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU ms < Lower Is Better A . 3.97598 |================================================================== B . 3.97577 |================================================================== C . 3.98095 |================================================================== oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 9342.56 |================================================================== B . 9340.03 |================================================================== C . 9341.19 |================================================================== oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 4760.20 |================================================================== B . 4762.01 |================================================================== C . 4764.79 |================================================================== oneDNN 2.6 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better A . 1.70174 |================================================================== B . 1.70897 |================================================================== C . 1.71077 |================================================================== oneDNN 2.6 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better A . 12.27 |==================================================================== B . 12.20 |==================================================================== C . 12.20 |====================================================================