deeprec win AMD Ryzen 9 7950X 16-Core testing with a ASUS ROG CROSSHAIR X670E HERO (1101 BIOS) and AMD Radeon RX 7900 XTX 24GB on Ubuntu 22.04 via the Phoronix Test Suite. a: Processor: AMD Ryzen 9 7950X 16-Core @ 4.50GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR X670E HERO (1101 BIOS), Chipset: AMD Device 14d8, Memory: 32GB, Disk: 2048GB SOLIDIGM SSDPFKKW020X7 + 2000GB, Graphics: AMD Radeon RX 7900 XTX 24GB (2304/1249MHz), Audio: AMD Device ab30, Monitor: ASUS MG28U, Network: Intel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411 OS: Ubuntu 22.04, Kernel: 6.3.0-060300rc7daily20230417-generic (x86_64), Desktop: GNOME Shell 42.5, Display Server: X Server 1.21.1.3 + Wayland, OpenGL: 4.6 Mesa 23.2.0-devel (git-f6fb189 2023-04-18 jammy-oibaf-ppa) (LLVM 15.0.7 DRM 3.52), Vulkan: 1.3.246, Compiler: GCC 11.3.0, File-System: ext4, Screen Resolution: 3840x2160 b: Processor: AMD Ryzen 9 7950X 16-Core @ 4.50GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR X670E HERO (1101 BIOS), Chipset: AMD Device 14d8, Memory: 32GB, Disk: 2048GB SOLIDIGM SSDPFKKW020X7 + 2000GB, Graphics: AMD Radeon RX 7900 XTX 24GB (2304/1249MHz), Audio: AMD Device ab30, Monitor: ASUS MG28U, Network: Intel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411 OS: Ubuntu 22.04, Kernel: 6.3.0-060300rc7daily20230417-generic (x86_64), Desktop: GNOME Shell 42.5, Display Server: X Server 1.21.1.3 + Wayland, OpenGL: 4.6 Mesa 23.2.0-devel (git-f6fb189 2023-04-18 jammy-oibaf-ppa) (LLVM 15.0.7 DRM 3.52), Vulkan: 1.3.246, Compiler: GCC 11.3.0, File-System: ext4, Screen Resolution: 3840x2160 c: Processor: AMD Ryzen 9 7950X 16-Core @ 4.50GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR X670E HERO (1101 BIOS), Chipset: AMD Device 14d8, Memory: 32GB, Disk: 2048GB SOLIDIGM SSDPFKKW020X7 + 2000GB, Graphics: AMD Radeon RX 7900 XTX 24GB (2304/1249MHz), Audio: AMD Device ab30, Monitor: ASUS MG28U, Network: Intel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411 OS: Ubuntu 22.04, Kernel: 6.3.0-060300rc7daily20230417-generic (x86_64), Desktop: GNOME Shell 42.5, Display Server: X Server 1.21.1.3 + Wayland, OpenGL: 4.6 Mesa 23.2.0-devel (git-f6fb189 2023-04-18 jammy-oibaf-ppa) (LLVM 15.0.7 DRM 3.52), Vulkan: 1.3.246, Compiler: GCC 11.3.0, File-System: ext4, Screen Resolution: 3840x2160 DeepRec Model: BST - Data Type: BF16 Throughput > Higher Is Better a . 5442.76 |================================================================== b . 5390.11 |================================================================= c . 5397.29 |================================================================= DeepRec Model: BST - Data Type: BF16 Gstep / sec > Higher Is Better a . 10.63 |==================================================================== b . 10.53 |=================================================================== c . 10.54 |=================================================================== DeepRec Model: BST - Data Type: FP32 Throughput > Higher Is Better a . 10459.37 |================================================================= b . 10469.73 |================================================================= c . 10450.02 |================================================================= DeepRec Model: BST - Data Type: FP32 Gstep / sec > Higher Is Better a . 20.43 |==================================================================== b . 20.45 |==================================================================== c . 20.41 |==================================================================== DeepRec Model: DIN - Data Type: BF16 Throughput > Higher Is Better a . 39620.03 |================================================================= b . 39371.49 |================================================================= c . 39664.62 |================================================================= DeepRec Model: DIN - Data Type: BF16 Gstep / sec > Higher Is Better a . 77.39 |==================================================================== b . 76.90 |=================================================================== c . 77.47 |==================================================================== DeepRec Model: DIN - Data Type: FP32 Throughput > Higher Is Better a . 45933.94 |================================================================= b . 46189.13 |================================================================= c . 46184.24 |================================================================= DeepRec Model: DIN - Data Type: FP32 Gstep / sec > Higher Is Better a . 89.71 |==================================================================== b . 90.21 |==================================================================== c . 90.20 |==================================================================== DeepRec Model: PLE - Data Type: BF16 Throughput > Higher Is Better a . 19539.80 |================================================================= b . 19521.64 |================================================================= c . 19562.81 |================================================================= DeepRec Model: PLE - Data Type: BF16 Gstep / sec > Higher Is Better a . 38.16 |==================================================================== b . 38.13 |==================================================================== c . 38.21 |==================================================================== DeepRec Model: PLE - Data Type: FP32 Throughput > Higher Is Better a . 34192.93 |================================================================= b . 33588.20 |================================================================ c . 33755.33 |================================================================ DeepRec Model: PLE - Data Type: FP32 Gstep / sec > Higher Is Better a . 66.78 |==================================================================== b . 65.60 |=================================================================== c . 65.93 |=================================================================== DeepRec Model: DLRM - Data Type: BF16 Throughput > Higher Is Better a . 40034.28 |================================================================= b . 39866.20 |================================================================= c . 39957.45 |================================================================= DeepRec Model: DLRM - Data Type: BF16 Gstep / sec > Higher Is Better a . 78.19 |==================================================================== b . 77.86 |==================================================================== c . 78.04 |==================================================================== DeepRec Model: DLRM - Data Type: FP32 Throughput > Higher Is Better a . 153853.71 |================================================================ b . 154485.76 |================================================================ c . 153879.88 |================================================================ DeepRec Model: DLRM - Data Type: FP32 Gstep / sec > Higher Is Better a . 300.50 |=================================================================== b . 301.73 |=================================================================== c . 300.55 |=================================================================== DeepRec Model: MMOE - Data Type: BF16 Throughput > Higher Is Better a . 24768.41 |================================================================= b . 24744.99 |================================================================= c . 24737.78 |================================================================= DeepRec Model: MMOE - Data Type: BF16 Gstep / sec > Higher Is Better a . 48.38 |==================================================================== b . 48.33 |==================================================================== c . 48.32 |==================================================================== DeepRec Model: MMOE - Data Type: FP32 Throughput > Higher Is Better a . 98473.37 |================================================================ b . 99439.16 |================================================================= c . 99015.53 |================================================================= DeepRec Model: MMOE - Data Type: FP32 Gstep / sec > Higher Is Better a . 192.33 |================================================================== b . 194.22 |=================================================================== c . 193.39 |=================================================================== DeepRec Model: DCNv2 - Data Type: BF16 Throughput > Higher Is Better a . 10981.04 |================================================================= b . 11010.90 |================================================================= c . 10997.10 |================================================================= DeepRec Model: DCNv2 - Data Type: BF16 Gstep / sec > Higher Is Better a . 21.45 |==================================================================== b . 21.51 |==================================================================== c . 21.48 |==================================================================== DeepRec Model: DCNv2 - Data Type: FP32 Throughput > Higher Is Better a . 15386.73 |================================================================= b . 15394.56 |================================================================= c . 15407.15 |================================================================= DeepRec Model: DCNv2 - Data Type: FP32 Gstep / sec > Higher Is Better a . 30.05 |==================================================================== b . 30.07 |==================================================================== c . 30.09 |====================================================================