Benchmarking the effect of unrolling on AARCH64 Tegra X1 vs S905 1.5GHz Android TV boxen NVIDIA Tegra X1: Processor: Cortex A57 rev 1 @ 1.91GHz (4 Cores), Motherboard: foster_e_hdd, Memory: 3072MB, Disk: 500GB Seagate ST500LM000-1EJ16 + 16GB SDW16G + 32GB 00000, Graphics: NVIDIA TEGRA OS: Ubuntu 14.10, Kernel: 3.10.61 (aarch64), Display Driver: fbdev 0.4.4, Compiler: GCC 4.9.1 + CUDA 6.5, File-System: ext4, Screen Resolution: 1920x2400 MXQ PRO+ Debian: Processor: AArch64 rev 4 @ 2.02GHz (4 Cores), Motherboard: Amlogic, Memory: 2048MB, Disk: 60GB A + 16GB AGND3R + 16GB SD16G OS: Debian 8.3, Kernel: 3.14.65-odroidc2 (aarch64), Compiler: GCC 4.9.2, File-System: ext4, Screen Resolution: 1280x1440 Mini MXIII GCC7: Processor: Unknown @ 1.50GHz (4 Cores), Memory: 2048MB, Disk: 16GB NCard + 32GB 00000 OS: Ubuntu 16.04, Kernel: 3.14.65-61 (aarch64), Compiler: GCC 7.0.0 20160904 + LLVM 3.8.0, File-System: ext4, Screen Resolution: 1280x1440 Mini MXIII GCC5: Processor: Unknown @ 1.50GHz (4 Cores), Memory: 2048MB, Disk: 16GB NCard + 32GB 00000 OS: Ubuntu 16.04, Kernel: 3.14.65-61 (aarch64), Compiler: GCC 5.3.1 20160413 + LLVM 3.8.0, File-System: ext4, Screen Resolution: 1280x1440 Mini MXIII GCC5 unrolled: Processor: Unknown @ 1.54GHz (4 Cores), Motherboard: Amlogic, Memory: 2048MB, Disk: 16GB NCard + 32GB 00000 OS: Ubuntu 16.04, Kernel: 3.14.79-vegas95 (aarch64), Compiler: GCC 5.4.0 20160609 + Clang 3.8.0-2ubuntu4 + LLVM 3.8.0, File-System: ext4, Screen Resolution: 1280x1440 Mini MXIII GCC7 unrolled: Processor: Unknown @ 1.54GHz (4 Cores), Motherboard: Amlogic, Memory: 2048MB, Disk: 16GB NCard + 32GB 00000 OS: Ubuntu 16.04, Kernel: 3.14.79-vegas95 (aarch64), Compiler: GCC 7.0.1 20170220 + Clang 3.8.0-2ubuntu4 + LLVM 3.8.0, File-System: ext4, Screen Resolution: 1280x1440 FFTW 3.3.4 Build: Stock - Size: 2D FFT Size 2048 Mflops > Higher Is Better NVIDIA Tegra X1 .......... 215.66 |============================================ MXQ PRO+ Debian .......... 108.88 |====================== Mini MXIII GCC7 .......... 196.93 |======================================== Mini MXIII GCC5 .......... 196.79 |======================================== Mini MXIII GCC5 unrolled . 194.66 |======================================== Mini MXIII GCC7 unrolled . 188.14 |====================================== Fhourstones 3.1 Complex Connect-4 Solving Kpos / sec > Higher Is Better NVIDIA Tegra X1 .......... 4705.57 |=========================================== MXQ PRO+ Debian .......... 2619.87 |======================== Mini MXIII GCC7 .......... 3136.27 |============================= Mini MXIII GCC5 .......... 3044.93 |============================ Mini MXIII GCC5 unrolled . 3228.87 |============================== Mini MXIII GCC7 unrolled . 3446.17 |=============================== VP8 libvpx Encoding 1.3.0 vpxenc Frames Per Second > Higher Is Better NVIDIA Tegra X1 .......... 11.52 |============================================= MXQ PRO+ Debian .......... 5.78 |======================= Mini MXIII GCC7 .......... 6.81 |=========================== Mini MXIII GCC5 .......... 6.69 |========================== Mini MXIII GCC5 unrolled . 6.75 |========================== Mini MXIII GCC7 unrolled . 7.27 |============================ Timed Apache Compilation 2.4.7 Time To Compile Seconds < Lower Is Better NVIDIA Tegra X1 .......... 134.51 |=========== MXQ PRO+ Debian .......... 535.29 |============================================ Mini MXIII GCC7 .......... 245.49 |==================== Mini MXIII GCC5 .......... 250.33 |===================== Mini MXIII GCC5 unrolled . 270.46 |====================== Mini MXIII GCC7 unrolled . 267.69 |====================== C-Ray 1.1 Total Time Seconds < Lower Is Better NVIDIA Tegra X1 .......... 84.74 |=================== MXQ PRO+ Debian .......... 182.11 |========================================== Mini MXIII GCC7 .......... 192.70 |============================================ Mini MXIII GCC5 .......... 186.85 |=========================================== Mini MXIII GCC5 unrolled . 157.42 |==================================== Mini MXIII GCC7 unrolled . 151.77 |=================================== Primesieve 5.4.2 1e12 Prime Number Generation Seconds < Lower Is Better NVIDIA Tegra X1 .......... 340.53 |======================= MXQ PRO+ Debian .......... 656.71 |============================================ Mini MXIII GCC7 .......... 575.31 |======================================= Mini MXIII GCC5 .......... 561.50 |====================================== Mini MXIII GCC5 unrolled . 570.56 |====================================== Mini MXIII GCC7 unrolled . 524.88 |=================================== Smallpt 1.0 Global Illumination Renderer; 100 Samples Seconds < Lower Is Better NVIDIA Tegra X1 .......... 1211 |============================================== MXQ PRO+ Debian .......... 219 |======== Mini MXIII GCC7 .......... 169 |====== Mini MXIII GCC5 .......... 179 |======= Mini MXIII GCC5 unrolled . 169 |====== Mini MXIII GCC7 unrolled . 166 |====== Stockfish 2014-11-26 Total Time ms < Lower Is Better NVIDIA Tegra X1 .......... 11431 |======================= MXQ PRO+ Debian .......... 22793 |============================================= Mini MXIII GCC7 .......... 20504 |======================================== Mini MXIII GCC5 .......... 22375 |============================================ Mini MXIII GCC5 unrolled . 22197 |============================================ Mini MXIII GCC7 unrolled . 20446 |======================================== FLAC Audio Encoding 1.3.1 WAV To FLAC Seconds < Lower Is Better NVIDIA Tegra X1 .......... 40.00 |========= MXQ PRO+ Debian .......... 187.93 |============================================ Mini MXIII GCC7 .......... 153.48 |==================================== Mini MXIII GCC5 .......... 158.16 |===================================== Mini MXIII GCC5 unrolled . 162.45 |====================================== Mini MXIII GCC7 unrolled . 163.56 |====================================== FFmpeg 2.6.2 H.264 HD To NTSC DV Seconds < Lower Is Better NVIDIA Tegra X1 .......... 74.91 |=============== MXQ PRO+ Debian .......... 222.78 |============================================ Mini MXIII GCC7 .......... 191.79 |====================================== Mini MXIII GCC5 .......... 195.16 |======================================= Mini MXIII GCC5 unrolled . 151.28 |============================== Mini MXIII GCC7 unrolled . 150.39 |============================== N-Queens 1.0 Elapsed Time Seconds < Lower Is Better NVIDIA Tegra X1 .......... 109.64 |============================= MXQ PRO+ Debian .......... 164.23 |============================================ Mini MXIII GCC7 .......... 156.09 |========================================== Mini MXIII GCC5 .......... 152.03 |========================================= Mini MXIII GCC5 unrolled . 140.45 |====================================== Mini MXIII GCC7 unrolled . 154.18 |========================================= PostgreSQL pgbench 9.4.3 Scaling: Buffer Test - Test: Normal Load - Mode: Read Write TPS > Higher Is Better NVIDIA Tegra X1 .......... 510.88 |================================== MXQ PRO+ Debian .......... 652.15 |============================================ Mini MXIII GCC5 .......... 626.65 |========================================== Mini MXIII GCC7 .......... 598.25 |======================================== Mini MXIII GCC7 unrolled . 619.53 |========================================== Mini MXIII GCC5 unrolled . 543.32 |===================================== PostgreSQL pgbench 9.4.3 Scaling: Buffer Test - Test: Single Thread - Mode: Read Write TPS > Higher Is Better NVIDIA Tegra X1 .......... 223.82 |=================================== MXQ PRO+ Debian .......... 281.37 |============================================ Mini MXIII GCC5 .......... 210.77 |================================= Mini MXIII GCC7 .......... 144.95 |======================= Mini MXIII GCC7 unrolled . 194.57 |============================== Mini MXIII GCC5 unrolled . 202.28 |================================ Redis 3.0.1 Test: GET Requests Per Second > Higher Is Better NVIDIA Tegra X1 .......... 615979.06 |========================================= MXQ PRO+ Debian .......... 247721.10 |================ Mini MXIII GCC7 .......... 264703.30 |================== Mini MXIII GCC5 .......... 270455.57 |================== Mini MXIII GCC7 unrolled . 272703.28 |================== Mini MXIII GCC5 unrolled . 256786.90 |================= Redis 3.0.1 Test: SET Requests Per Second > Higher Is Better NVIDIA Tegra X1 .......... 432160.48 |========================================= MXQ PRO+ Debian .......... 186038.34 |================== Mini MXIII GCC7 .......... 194408.30 |================== Mini MXIII GCC5 .......... 197144.29 |=================== Mini MXIII GCC7 unrolled . 200337.42 |=================== Mini MXIII GCC5 unrolled . 197794.05 |===================