Amazon EC2 A1 ARM Graviton Linux Benchmarks Gigabyte R281-Z92 with 2x AMD Epyc 7371 CPUs ARMv8 a1.large: Processor: ARMv8 Cortex-A72 (2 Cores), Motherboard: Amazon EC2 a1.large (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 4096MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1028-aws (aarch64), Compiler: GCC 7.3.0, File-System: ext4 ARMv8 a1.xlarge: Processor: ARMv8 Cortex-A72 (4 Cores), Motherboard: Amazon EC2 a1.xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 8192MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1028-aws (aarch64), Compiler: GCC 7.3.0, File-System: ext4 ARMv8 a1.2xlarge: Processor: ARMv8 Cortex-A72 (8 Cores), Motherboard: Amazon EC2 a1.2xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 16384MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1028-aws (aarch64), Compiler: GCC 7.3.0, File-System: ext4 ARMv8 a1.4xlarge: Processor: ARMv8 Cortex-A72 (16 Cores), Motherboard: Amazon EC2 a1.4xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 31744MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1028-aws (aarch64), Compiler: GCC 7.3.0, File-System: ext4 AMD m5a.large: Processor: AMD EPYC 7571 (1 Core / 2 Threads), Motherboard: Amazon EC2 m5a.large (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 8192MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM AMD m5a.xlarge: Processor: AMD EPYC 7571 (2 Cores / 4 Threads), Motherboard: Amazon EC2 m5a.xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 16384MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM AMD m5a.2xlarge: Processor: AMD EPYC 7571 (4 Cores / 8 Threads), Motherboard: Amazon EC2 m5a.2xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 31744MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM AMD m5a.4xlarge: Processor: AMD EPYC 7571 (8 Cores / 16 Threads), Motherboard: Amazon EC2 m5a.4xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 63488MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM Intel m5.large: Processor: Intel Xeon Platinum 8175M (1 Core / 2 Threads), Motherboard: Amazon EC2 m5.large (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 8192MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM Intel m4.xlarge: Processor: Intel Xeon Platinum 8175M (2 Cores / 4 Threads), Motherboard: Amazon EC2 m5.xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 16384MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM Intel m5.2xlarge: Processor: Intel Xeon Platinum 8175M (4 Cores / 8 Threads), Motherboard: Amazon EC2 m5.2xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 31744MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM Intel m5.4xlarge: Processor: Intel Xeon Platinum 8175M (8 Cores / 16 Threads), Motherboard: Amazon EC2 m5.4xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 63488MB, Disk: 52GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 18.04, Kernel: 4.15.0-1021-aws (x86_64), Compiler: GCC 7.3.0, File-System: ext4, System Layer: KVM Cavium Thunder X2 2x28C: Processor: Unknown (56 Cores / 224 Threads), Motherboard: GIGABYTE MT91-FS1-00 v01000100 (F10 BIOS), Chipset: Cavium CN99xx, Memory: 65536MB, Disk: 320GB TOSHIBA MK3261GS + 8 x 1920GB DOPE1920S5xnNMRI, Graphics: ASPEED Family, Network: Intel 82571EB/82571GB OS: Red Hat Enterprise Linux 7.6, Kernel: 4.14.0-49.13.1.el7a.aarch64 (aarch64) 20180827, Compiler: GCC 4.8.5 20150623, File-System: ext4, Screen Resolution: 1024x768 AMD Epyc 7371 x2: Processor: 2 x AMD EPYC 7371 16-Core @ 3.10GHz (32 Cores / 64 Threads), Motherboard: GIGABYTE MZ91-FS0-00 v01000100 (F11 BIOS), Chipset: AMD Family 17h, Memory: 16 x 8192 MB DDR4-2667MT/s, Disk: 480GB INTEL SSDPED1D480GA + 280GB INTEL SSDPE21D280GA + 16GB, Graphics: ASPEED Family 32GB, Monitor: DELL E1709W, Network: 2 x Intel I350 OS: Fedora 29, Kernel: 4.20.14-200.fc29.x86_64 (x86_64), Display Server: X Server, Compiler: GCC 8.3.1 20190223, File-System: ext4, Screen Resolution: 1440x900 AMD Epyc 7371 x2 second run: Processor: 2 x AMD EPYC 7371 16-Core @ 3.10GHz (32 Cores / 64 Threads), Motherboard: GIGABYTE MZ91-FS0-00 v01000100 (F11 BIOS), Chipset: AMD Family 17h, Memory: 16 x 8192 MB DDR4-2667MT/s, Disk: 480GB INTEL SSDPED1D480GA + 280GB INTEL SSDPE21D280GA + 16GB, Graphics: ASPEED Family 32GB, Monitor: DELL E1709W, Network: 2 x Intel I350 OS: Fedora 29, Kernel: 4.20.14-200.fc29.x86_64 (x86_64), Display Server: X Server, Compiler: GCC 8.3.1 20190223, File-System: ext4, Screen Resolution: 1440x900 Rodinia 2.4 Test: OpenMP LavaMD Seconds < Lower Is Better ARMv8 a1.large .............. 1074.00 |======================================== ARMv8 a1.xlarge ............. 541.00 |==================== ARMv8 a1.2xlarge ............ 272.00 |========== ARMv8 a1.4xlarge ............ 138.00 |===== AMD m5a.large ............... 886.00 |================================= AMD m5a.xlarge .............. 443.00 |================ AMD m5a.2xlarge ............. 263.00 |========== AMD m5a.4xlarge ............. 131.00 |===== Intel m5.large .............. 803.00 |============================== Intel m4.xlarge ............. 406.00 |=============== Intel m5.2xlarge ............ 202.00 |======== Intel m5.4xlarge ............ 102.00 |==== Cavium Thunder X2 2x28C ..... 36.17 |= AMD Epyc 7371 x2 ............ 11.58 | AMD Epyc 7371 x2 second run . 11.62 | Rust Prime Benchmark Performance / Cost - Prime Number Test To 200,000,000 Seconds x Dollar < Lower Is Better ARMv8 a1.large ... 1.28 |===== ARMv8 a1.xlarge .. 0.64 |== ARMv8 a1.2xlarge . 0.32 |= ARMv8 a1.4xlarge . 0.16 |= AMD m5a.large .... 13.89 |===================================================== AMD m5a.xlarge ... 6.95 |=========================== AMD m5a.2xlarge .. 5.60 |===================== AMD m5a.4xlarge .. 2.77 |=========== Intel m5.large ... 1.24 |===== Intel m4.xlarge .. 0.62 |== Intel m5.2xlarge . 0.31 |= Intel m5.4xlarge . 0.15 |= PyBench 2018-02-16 Performance / Cost - Total For Average Test Times Milliseconds x Dollar < Lower Is Better ARMv8 a1.xlarge .. 42.44 |================= ARMv8 a1.2xlarge . 85.78 |=================================== ARMv8 a1.4xlarge . 128.16 |==================================================== AMD m5a.xlarge ... 16.64 |======= AMD m5a.2xlarge .. 16.62 |======= AMD m5a.4xlarge .. 33.10 |============= Intel m4.xlarge .. 13.49 |===== Intel m5.2xlarge . 13.55 |===== Intel m5.4xlarge . 27.04 |=========== Zstd Compression 1.3.4 Performance / Cost - Compressing ubuntu-16.04.3-server-i386.img, Compression Level 19 Seconds x Dollar < Lower Is Better ARMv8 a1.large ... 1.71 |====================================================== ARMv8 a1.xlarge .. 1.08 |================================== ARMv8 a1.2xlarge . 0.62 |==================== ARMv8 a1.4xlarge . 0.42 |============= AMD m5a.large .... 1.61 |=================================================== AMD m5a.xlarge ... 0.86 |=========================== AMD m5a.2xlarge .. 1.10 |=================================== AMD m5a.4xlarge .. 0.67 |===================== Intel m5.large ... 1.15 |==================================== Intel m4.xlarge .. 0.70 |====================== Intel m5.2xlarge . 0.36 |=========== Intel m5.4xlarge . 0.20 |====== Rodinia 2.4 Performance / Cost - Test: OpenMP LavaMD Seconds x Dollar < Lower Is Better ARMv8 a1.large ... 53.72 |===================================================== ARMv8 a1.xlarge .. 27.06 |=========================== ARMv8 a1.2xlarge . 13.60 |============= ARMv8 a1.4xlarge . 6.89 |======= AMD m5a.large .... 53.14 |==================================================== AMD m5a.xlarge ... 26.61 |========================== AMD m5a.2xlarge .. 21.00 |===================== AMD m5a.4xlarge .. 10.47 |========== Intel m5.large ... 48.18 |================================================ Intel m4.xlarge .. 28.39 |============================ Intel m5.2xlarge . 14.15 |============== Intel m5.4xlarge . 7.12 |======= x264 2018-09-25 Performance / Cost - H.264 Video Encoding Frames Per Second Per Dollar > Higher Is Better ARMv8 a1.xlarge .. 876.00 |======== ARMv8 a1.2xlarge . 1662.00 |============== ARMv8 a1.4xlarge . 3243.00 |============================ AMD m5a.xlarge ... 1620.00 |============== AMD m5a.2xlarge .. 2654.00 |======================= AMD m5a.4xlarge .. 5090.00 |============================================ Intel m5.large ... 921.00 |======== Intel m4.xlarge .. 1596.00 |============== Intel m5.2xlarge . 3062.00 |========================== Intel m5.4xlarge . 5955.00 |=================================================== LAME MP3 Encoding 3.100 Performance / Cost - WAV To MP3 Seconds x Dollar < Lower Is Better ARMv8 a1.xlarge .. 0.83 |================== ARMv8 a1.2xlarge . 0.83 |================== ARMv8 a1.4xlarge . 2.49 |====================================================== AMD m5a.xlarge ... 0.47 |========== AMD m5a.2xlarge .. 0.48 |========== AMD m5a.4xlarge .. 1.41 |=============================== Intel m4.xlarge .. 0.38 |======== Intel m5.2xlarge . 0.38 |======== Intel m5.4xlarge . 1.15 |========================= PHPBench 0.8.1 Performance / Cost - PHP Benchmark Suite Score Per Dollar > Higher Is Better ARMv8 a1.xlarge .. 16846700.00 |================ ARMv8 a1.2xlarge . 16746500.00 |=============== ARMv8 a1.4xlarge . 8367200.00 |======== AMD m5a.2xlarge .. 34255300.00 |================================ AMD m5a.4xlarge .. 18320550.00 |================= Intel m5.2xlarge . 50818900.00 |=============================================== Intel m5.4xlarge . 50880500.00 |=============================================== Rust Prime Benchmark Prime Number Test To 200,000,000 Seconds < Lower Is Better ARMv8 a1.large .............. 128.30 |=========== ARMv8 a1.xlarge ............. 64.22 |====== ARMv8 a1.2xlarge ............ 32.08 |=== ARMv8 a1.4xlarge ............ 16.06 |= AMD m5a.large ............... 463.02 |========================================= AMD m5a.xlarge .............. 231.78 |===================== AMD m5a.2xlarge ............. 139.88 |============ AMD m5a.4xlarge ............. 69.33 |====== Intel m5.large .............. 123.86 |=========== Intel m4.xlarge ............. 62.14 |====== Intel m5.2xlarge ............ 31.01 |=== Intel m5.4xlarge ............ 15.49 |= Cavium Thunder X2 2x28C ..... 8.35 |= AMD Epyc 7371 x2 ............ 11.72 |= AMD Epyc 7371 x2 second run . 11.74 |= Rust Mandelbrot Time To Complete Serial/Parallel Mandelbrot Seconds < Lower Is Better ARMv8 a1.large .............. 243.00 |========================================= ARMv8 a1.xlarge ............. 152.00 |========================== ARMv8 a1.2xlarge ............ 106.00 |================== ARMv8 a1.4xlarge ............ 83.98 |============== AMD m5a.large ............... 228.74 |======================================= AMD m5a.xlarge .............. 138.01 |======================= AMD m5a.2xlarge ............. 100.95 |================= AMD m5a.4xlarge ............. 74.44 |============= Intel m5.large .............. 216.10 |==================================== Intel m4.xlarge ............. 133.38 |======================= Intel m5.2xlarge ............ 92.20 |================ Intel m5.4xlarge ............ 71.25 |============ Cavium Thunder X2 2x28C ..... 92.27 |================ AMD Epyc 7371 x2 ............ 40.57 |======= AMD Epyc 7371 x2 second run . 40.55 |======= PHPBench 0.8.1 PHP Benchmark Suite Score > Higher Is Better ARMv8 a1.large .............. 168855 |============== ARMv8 a1.xlarge ............. 168467 |============== ARMv8 a1.2xlarge ............ 167465 |============= ARMv8 a1.4xlarge ............ 167344 |============= AMD m5a.large ............... 365665 |============================= AMD m5a.xlarge .............. 369362 |============================== AMD m5a.2xlarge ............. 342553 |=========================== AMD m5a.4xlarge ............. 366411 |============================= Intel m5.large .............. 510419 |========================================= Intel m4.xlarge ............. 510819 |========================================= Intel m5.2xlarge ............ 508189 |========================================= Intel m5.4xlarge ............ 508805 |========================================= Cavium Thunder X2 2x28C ..... 93883 |======== AMD Epyc 7371 x2 ............ 469456 |====================================== AMD Epyc 7371 x2 second run . 471135 |====================================== PHP Micro Benchmarks Test: Zend micro_bench Seconds < Lower Is Better ARMv8 a1.large .............. 8.94 |============================ ARMv8 a1.xlarge ............. 8.88 |============================ ARMv8 a1.2xlarge ............ 9.04 |============================ ARMv8 a1.4xlarge ............ 9.09 |============================ AMD m5a.large ............... 3.14 |========== AMD m5a.xlarge .............. 3.19 |========== AMD m5a.2xlarge ............. 3.26 |========== AMD m5a.4xlarge ............. 3.11 |========== Intel m5.large .............. 2.65 |======== Intel m4.xlarge ............. 2.64 |======== Intel m5.2xlarge ............ 2.68 |======== Intel m5.4xlarge ............ 2.65 |======== Cavium Thunder X2 2x28C ..... 13.53 |========================================== AMD Epyc 7371 x2 ............ 2.54 |======== AMD Epyc 7371 x2 second run . 2.56 |======== PHP Micro Benchmarks Test: Zend bench Seconds < Lower Is Better ARMv8 a1.large .............. 1.81 |=========================== ARMv8 a1.xlarge ............. 1.81 |=========================== ARMv8 a1.2xlarge ............ 1.82 |=========================== ARMv8 a1.4xlarge ............ 1.79 |=========================== AMD m5a.large ............... 0.74 |=========== AMD m5a.xlarge .............. 0.73 |=========== AMD m5a.2xlarge ............. 0.84 |============= AMD m5a.4xlarge ............. 0.74 |=========== Intel m5.large .............. 0.55 |======== Intel m4.xlarge ............. 0.54 |======== Intel m5.2xlarge ............ 0.54 |======== Intel m5.4xlarge ............ 0.54 |======== Cavium Thunder X2 2x28C ..... 2.87 |=========================================== AMD Epyc 7371 x2 ............ 0.56 |======== AMD Epyc 7371 x2 second run . 0.56 |======== m-queens 1.2 Time To Solve Seconds < Lower Is Better ARMv8 a1.large .............. 882.00 |====================================== ARMv8 a1.xlarge ............. 441.00 |=================== ARMv8 a1.2xlarge ............ 221.00 |========= ARMv8 a1.4xlarge ............ 110.00 |===== AMD m5a.large ............... 786.00 |================================== AMD m5a.xlarge .............. 392.00 |================= AMD m5a.2xlarge ............. 233.00 |========== AMD m5a.4xlarge ............. 114.00 |===== Intel m5.large .............. 956.00 |========================================= Intel m4.xlarge ............. 480.00 |===================== Intel m5.2xlarge ............ 239.00 |========== Intel m5.4xlarge ............ 120.00 |===== AMD Epyc 7371 x2 ............ 20.96 |= AMD Epyc 7371 x2 second run . 20.95 |= Rodinia 2.4 Performance / Cost - Test: OpenMP CFD Solver Seconds x Dollar < Lower Is Better ARMv8 a1.large ... 1.81 |================================================ ARMv8 a1.xlarge .. 1.11 |============================= ARMv8 a1.2xlarge . 0.62 |================ ARMv8 a1.4xlarge . 0.41 |=========== AMD m5a.large .... 1.69 |============================================= AMD m5a.xlarge ... 0.90 |======================== AMD m5a.2xlarge .. 1.17 |=============================== AMD m5a.4xlarge .. 0.69 |================== Intel m5.large ... 1.83 |================================================ Intel m4.xlarge .. 2.04 |====================================================== Intel m5.2xlarge . 0.96 |========================= Intel m5.4xlarge . 0.50 |============= Apache Benchmark 2.4.29 Static Web Page Serving Requests Per Second > Higher Is Better ARMv8 a1.large .............. 5542 |========= ARMv8 a1.xlarge ............. 7484 |============= ARMv8 a1.2xlarge ............ 7111 |============ ARMv8 a1.4xlarge ............ 7231 |============ AMD m5a.large ............... 9225 |================ AMD m5a.xlarge .............. 13871 |======================== AMD m5a.2xlarge ............. 10534 |================== AMD m5a.4xlarge ............. 10274 |================== Intel m5.large .............. 14337 |======================== Intel m4.xlarge ............. 21696 |===================================== Intel m5.2xlarge ............ 24547 |========================================== Intel m5.4xlarge ............ 24607 |========================================== Cavium Thunder X2 2x28C ..... 8370 |============== AMD Epyc 7371 x2 ............ 10846 |=================== AMD Epyc 7371 x2 second run . 11171 |=================== PyBench 2018-02-16 Total For Average Test Times Milliseconds < Lower Is Better ARMv8 a1.large .............. 4266 |=========================================== ARMv8 a1.xlarge ............. 4244 |=========================================== ARMv8 a1.2xlarge ............ 4289 |=========================================== ARMv8 a1.4xlarge ............ 4272 |=========================================== AMD m5a.large ............... 1657 |================= AMD m5a.xlarge .............. 1664 |================= AMD m5a.2xlarge ............. 1662 |================= AMD m5a.4xlarge ............. 1655 |================= Intel m5.large .............. 1353 |============== Intel m4.xlarge ............. 1349 |============== Intel m5.2xlarge ............ 1355 |============== Intel m5.4xlarge ............ 1352 |============== AMD Epyc 7371 x2 ............ 1606 |================ AMD Epyc 7371 x2 second run . 1600 |================ Perl Benchmarks Test: Pod2html Seconds < Lower Is Better ARMv8 a1.large .............. 0.42504815 |==================================== ARMv8 a1.xlarge ............. 0.43503233 |===================================== ARMv8 a1.2xlarge ............ 0.43043023 |===================================== ARMv8 a1.4xlarge ............ 0.42981722 |===================================== AMD m5a.large ............... 0.21205971 |================== AMD m5a.xlarge .............. 0.20664451 |================== AMD m5a.2xlarge ............. 0.20577741 |================== AMD m5a.4xlarge ............. 0.20555748 |================= Intel m5.large .............. 0.14030294 |============ Intel m4.xlarge ............. 0.14096660 |============ Intel m5.2xlarge ............ 0.14085884 |============ Intel m5.4xlarge ............ 0.14023221 |============ Cavium Thunder X2 2x28C ..... 0.36356409 |=============================== AMD Epyc 7371 x2 second run . 0.16294567 |============== Rust Mandelbrot Performance / Cost - Time To Complete Serial/Parallel Mandelbrot Seconds x Dollar < Lower Is Better ARMv8 a1.large ... 2.43 |============================= ARMv8 a1.xlarge .. 1.52 |================== ARMv8 a1.2xlarge . 2.13 |========================= ARMv8 a1.4xlarge . 2.52 |============================== AMD m5a.large .... 4.57 |====================================================== AMD m5a.xlarge ... 2.76 |================================= AMD m5a.2xlarge .. 3.03 |==================================== AMD m5a.4xlarge .. 2.98 |=================================== Intel m5.large ... 4.32 |=================================================== Intel m4.xlarge .. 2.67 |================================ Intel m5.2xlarge . 2.77 |================================= Intel m5.4xlarge . 3.56 |========================================== 7-Zip Compression 16.02 Compress Speed Test MIPS > Higher Is Better ARMv8 a1.large .............. 3868 |= ARMv8 a1.xlarge ............. 6866 |== ARMv8 a1.2xlarge ............ 13589 |===== ARMv8 a1.4xlarge ............ 25230 |========= AMD m5a.large ............... 4711 |== AMD m5a.xlarge .............. 9091 |=== AMD m5a.2xlarge ............. 14702 |===== AMD m5a.4xlarge ............. 28557 |========== Intel m5.large .............. 5658 |== Intel m4.xlarge ............. 10769 |==== Intel m5.2xlarge ............ 20803 |======= Intel m5.4xlarge ............ 39564 |============== Cavium Thunder X2 2x28C ..... 110396 |======================================== AMD Epyc 7371 x2 ............ 114240 |========================================= AMD Epyc 7371 x2 second run . 113331 |========================================= Go Benchmarks Test: json Nanoseconds/Operation < Lower Is Better ARMv8 a1.large .............. 78842192 |======================================= ARMv8 a1.xlarge ............. 39753346 |==================== ARMv8 a1.2xlarge ............ 20320318 |========== ARMv8 a1.4xlarge ............ 10540841 |===== AMD m5a.large ............... 66986350 |================================= AMD m5a.xlarge .............. 33246787 |================ AMD m5a.2xlarge ............. 20044569 |========== AMD m5a.4xlarge ............. 10340655 |===== Intel m5.large .............. 58808172 |============================= Intel m4.xlarge ............. 29757990 |=============== Intel m5.2xlarge ............ 15057128 |======= Intel m5.4xlarge ............ 7700381 |==== Cavium Thunder X2 2x28C ..... 3286870 |== AMD Epyc 7371 x2 ............ 3155112 |== AMD Epyc 7371 x2 second run . 2998824 |= Cost To Run Tests Cost / Price Per Hour Cost / Price Per Hour < Lower Is Better ARMv8 a1.large ... 0.25 |======================== ARMv8 a1.xlarge .. 0.19 |================== ARMv8 a1.2xlarge . 0.26 |========================= ARMv8 a1.4xlarge . 0.40 |======================================= AMD m5a.large .... 0.23 |====================== AMD m5a.xlarge ... 0.27 |========================== AMD m5a.2xlarge .. 0.40 |======================================= AMD m5a.4xlarge .. 0.56 |====================================================== Intel m5.large ... 0.23 |====================== Intel m4.xlarge .. 0.27 |========================== Intel m5.2xlarge . 0.35 |================================== Intel m5.4xlarge . 0.49 |=============================================== 7-Zip Compression 16.02 Performance / Cost - Compress Speed Test MIPS Per Dollar > Higher Is Better ARMv8 a1.2xlarge . 1358900.00 |========================== ARMv8 a1.4xlarge . 2523000.00 |================================================ AMD m5a.xlarge ... 909100.00 |================= AMD m5a.2xlarge .. 1470200.00 |============================ AMD m5a.4xlarge .. 951900.00 |================== Intel m4.xlarge .. 1076900.00 |==================== Intel m5.2xlarge . 2080300.00 |======================================== Intel m5.4xlarge . 1318800.00 |========================= Go Benchmarks Test: build Nanoseconds/Operation < Lower Is Better ARMv8 a1.large .............. 39591611983 |================================== ARMv8 a1.xlarge ............. 39673825966 |=================================== ARMv8 a1.2xlarge ............ 40031137671 |=================================== ARMv8 a1.4xlarge ............ 41340905115 |==================================== AMD m5a.large ............... 21534657315 |=================== AMD m5a.xlarge .............. 20758758226 |================== AMD m5a.2xlarge ............. 20802086649 |================== AMD m5a.4xlarge ............. 21097798242 |================== Intel m5.large .............. 16632861257 |============== Intel m4.xlarge ............. 17865141058 |================ Intel m5.2xlarge ............ 17520780038 |=============== Intel m5.4xlarge ............ 17197591367 |=============== Cavium Thunder X2 2x28C ..... 40032868756 |=================================== AMD Epyc 7371 x2 ............ 23286023963 |==================== AMD Epyc 7371 x2 second run . 23367749753 |==================== LAME MP3 Encoding 3.100 WAV To MP3 Seconds < Lower Is Better ARMv8 a1.large .............. 83.11 |======================================== ARMv8 a1.xlarge ............. 83.12 |======================================== ARMv8 a1.2xlarge ............ 83.26 |======================================== ARMv8 a1.4xlarge ............ 83.03 |======================================== AMD m5a.large ............... 47.92 |======================= AMD m5a.xlarge .............. 47.02 |======================= AMD m5a.2xlarge ............. 48.09 |======================= AMD m5a.4xlarge ............. 47.05 |======================= Intel m5.large .............. 38.32 |================== Intel m4.xlarge ............. 38.32 |================== Intel m5.2xlarge ............ 38.32 |================== Intel m5.4xlarge ............ 38.32 |================== Cavium Thunder X2 2x28C ..... 87.44 |========================================== AMD Epyc 7371 x2 ............ 35.39 |================= AMD Epyc 7371 x2 second run . 35.40 |================= Perl Benchmarks Test: Interpreter Seconds < Lower Is Better ARMv8 a1.large .............. 0.00243051 |===================================== ARMv8 a1.xlarge ............. 0.00243618 |===================================== ARMv8 a1.2xlarge ............ 0.00238317 |==================================== ARMv8 a1.4xlarge ............ 0.00240482 |===================================== AMD m5a.large ............... 0.00105898 |================ AMD m5a.xlarge .............. 0.00105854 |================ AMD m5a.2xlarge ............. 0.00115248 |================== AMD m5a.4xlarge ............. 0.00131657 |==================== Intel m5.large .............. 0.00099317 |=============== Intel m4.xlarge ............. 0.00106437 |================ Intel m5.2xlarge ............ 0.00109555 |================= Intel m5.4xlarge ............ 0.00104766 |================ Cavium Thunder X2 2x28C ..... 0.00200625 |============================== AMD Epyc 7371 x2 second run . 0.00206108 |=============================== Go Benchmarks Performance / Cost - Test: build Nanoseconds/Operation x Dollar < Lower Is Better ARMv8 a1.2xlarge . 400311376.71 |============================================= ARMv8 a1.4xlarge . 413409051.15 |============================================== AMD m5a.2xlarge .. 208020866.49 |======================= AMD m5a.4xlarge .. 210977982.42 |======================= Intel m5.2xlarge . 175207800.38 |=================== Intel m5.4xlarge . 171975913.67 |=================== Go Benchmarks Performance / Cost - Test: garbage Nanoseconds/Operation x Dollar < Lower Is Better ARMv8 a1.4xlarge . 25993.15 |====================================== AMD m5a.4xlarge .. 21646.82 |================================ Intel m5.2xlarge . 33876.64 |================================================== Intel m5.4xlarge . 16922.23 |========================= Perl Benchmarks Performance / Cost - Test: Pod2html Seconds x Dollar < Lower Is Better ARMv8 a1.2xlarge . 0.01 |=========================== ARMv8 a1.4xlarge . 0.02 |====================================================== AMD m5a.4xlarge .. 0.01 |=========================== Intel m5.4xlarge . 0.01 |=========================== m-queens 1.2 Performance / Cost - Time To Solve Seconds x Dollar < Lower Is Better ARMv8 a1.large ... 35.29 |======================== ARMv8 a1.xlarge .. 17.65 |============ ARMv8 a1.2xlarge . 8.83 |====== ARMv8 a1.4xlarge . 4.42 |=== AMD m5a.large .... 47.19 |================================= AMD m5a.xlarge ... 23.54 |================ AMD m5a.2xlarge .. 16.29 |=========== AMD m5a.4xlarge .. 8.00 |====== Intel m5.large ... 76.46 |===================================================== Intel m4.xlarge .. 38.36 |=========================== Intel m5.2xlarge . 19.14 |============= Intel m5.4xlarge . 9.58 |======= Rodinia 2.4 Test: OpenMP CFD Solver Seconds < Lower Is Better ARMv8 a1.large .............. 181.00 |========================================= ARMv8 a1.xlarge ............. 111.00 |========================= ARMv8 a1.2xlarge ............ 61.89 |============== ARMv8 a1.4xlarge ............ 40.92 |========= AMD m5a.large ............... 169.13 |====================================== AMD m5a.xlarge .............. 90.43 |==================== AMD m5a.2xlarge ............. 58.60 |============= AMD m5a.4xlarge ............. 34.60 |======== Intel m5.large .............. 182.64 |========================================= Intel m4.xlarge ............. 102.08 |======================= Intel m5.2xlarge ............ 48.18 |=========== Intel m5.4xlarge ............ 25.23 |====== Cavium Thunder X2 2x28C ..... 18.53 |==== AMD Epyc 7371 x2 ............ 11.11 |== AMD Epyc 7371 x2 second run . 11.14 |=== Go Benchmarks Test: garbage Nanoseconds/Operation < Lower Is Better ARMv8 a1.large .............. 16850985 |======================================= ARMv8 a1.xlarge ............. 8715063 |==================== ARMv8 a1.2xlarge ............ 4727623 |=========== ARMv8 a1.4xlarge ............ 2599315 |====== AMD m5a.large ............... 13241728 |=============================== AMD m5a.xlarge .............. 7023812 |================ AMD m5a.2xlarge ............. 4112557 |========== AMD m5a.4xlarge ............. 2164682 |===== Intel m5.large .............. 12482078 |============================= Intel m4.xlarge ............. 6786423 |================ Intel m5.2xlarge ............ 3387664 |======== Intel m5.4xlarge ............ 1692223 |==== Cavium Thunder X2 2x28C ..... 1092542 |=== AMD Epyc 7371 x2 ............ 1298243 |=== AMD Epyc 7371 x2 second run . 1299055 |=== Apache Benchmark 2.4.29 Performance / Cost - Static Web Page Serving Requests Per Second Per Dollar > Higher Is Better ARMv8 a1.large ... 554229.00 |============ ARMv8 a1.xlarge .. 748416.00 |================= ARMv8 a1.2xlarge . 237034.00 |===== ARMv8 a1.4xlarge . 144625.40 |=== AMD m5a.large .... 922537.00 |==================== AMD m5a.xlarge ... 1387149.00 |=============================== AMD m5a.2xlarge .. 351143.67 |======== AMD m5a.4xlarge .. 171233.33 |==== Intel m5.large ... 1433747.00 |================================ Intel m4.xlarge .. 2169568.00 |================================================ Intel m5.2xlarge . 1227359.50 |=========================== Intel m5.4xlarge . 820236.33 |================== Zstd Compression 1.3.4 Compressing ubuntu-16.04.3-server-i386.img, Compression Level 19 Seconds < Lower Is Better ARMv8 a1.large .............. 171.00 |========================================= ARMv8 a1.xlarge ............. 108.00 |========================== ARMv8 a1.2xlarge ............ 62.04 |=============== ARMv8 a1.4xlarge ............ 41.85 |========== AMD m5a.large ............... 161.43 |======================================= AMD m5a.xlarge .............. 86.05 |===================== AMD m5a.2xlarge ............. 54.93 |============= AMD m5a.4xlarge ............. 33.27 |======== Intel m5.large .............. 114.58 |=========================== Intel m4.xlarge ............. 69.50 |================= Intel m5.2xlarge ............ 35.56 |========= Intel m5.4xlarge ............ 19.73 |===== Cavium Thunder X2 2x28C ..... 16.52 |==== AMD Epyc 7371 x2 ............ 14.74 |==== AMD Epyc 7371 x2 second run . 14.66 |==== x264 2018-09-25 H.264 Video Encoding Frames Per Second > Higher Is Better ARMv8 a1.large ... 5.38 |===== ARMv8 a1.xlarge .. 8.76 |======== ARMv8 a1.2xlarge . 16.62 |=============== ARMv8 a1.4xlarge . 32.43 |============================= AMD m5a.large .... 9.51 |======== AMD m5a.xlarge ... 16.20 |============== AMD m5a.2xlarge .. 26.54 |======================== AMD m5a.4xlarge .. 50.90 |============================================= Intel m5.large ... 9.21 |======== Intel m4.xlarge .. 15.96 |============== Intel m5.2xlarge . 30.62 |=========================== Intel m5.4xlarge . 59.55 |===================================================== Go Benchmarks Performance / Cost - Test: json Nanoseconds/Operation x Dollar < Lower Is Better AMD m5a.4xlarge .. 103406.55 |================================================= Intel m5.4xlarge . 77003.81 |====================================