Amazon EC2 Graviton3 Benchmark Comparison Amazon AWS Graviton3 benchmarks by Michael Larabel. c7g.4xlarge Graviton3: Processor: ARMv8 Neoverse-V1 (16 Cores), Motherboard: Amazon EC2 c7g.4xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 32GB, Disk: 193GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 22.04, Kernel: 5.15.0-1004-aws (aarch64), Compiler: GCC 11.2.0, File-System: ext4, System Layer: amazon a1.4xlarge Graviton: Processor: ARMv8 Cortex-A72 (16 Cores), Motherboard: Amazon EC2 a1.4xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 32GB, Disk: 193GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 22.04, Kernel: 5.15.0-1004-aws (aarch64), Compiler: GCC 11.2.0, File-System: ext4, System Layer: amazon c6g.4xlarge Graviton2: Processor: ARMv8 Neoverse-N1 (16 Cores), Motherboard: Amazon EC2 c6g.4xlarge (1.0 BIOS), Chipset: Amazon Device 0200, Memory: 32GB, Disk: 193GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 22.04, Kernel: 5.15.0-1004-aws (aarch64), Compiler: GCC 11.2.0, File-System: ext4, System Layer: amazon c6a.4xlarge EPYC: Processor: AMD EPYC 7R13 (8 Cores / 16 Threads), Motherboard: Amazon EC2 c6a.4xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 32GB, Disk: 193GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 22.04, Kernel: 5.15.0-1004-aws (x86_64), Compiler: GCC 11.2.0, File-System: ext4, System Layer: amazon c6i.4xlarge Xeon: Processor: Intel Xeon Platinum 8375C (8 Cores / 16 Threads), Motherboard: Amazon EC2 c6i.4xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 32GB, Disk: 193GB Amazon Elastic Block Store, Network: Amazon Elastic OS: Ubuntu 22.04, Kernel: 5.15.0-1004-aws (x86_64), Compiler: GCC 11.2.0, File-System: ext4, System Layer: amazon High Performance Conjugate Gradient 3.1 GFLOP/s > Higher Is Better c7g.4xlarge Graviton3 . 26.30580 |============================================= a1.4xlarge Graviton ... 3.77834 |====== c6g.4xlarge Graviton2 . 19.72180 |================================== c6a.4xlarge EPYC ...... 5.06042 |========= c6i.4xlarge Xeon ...... 8.66031 |=============== NAS Parallel Benchmarks 3.4 Test / Class: BT.C Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 10339.53 |================================== a1.4xlarge Graviton ... 3148.18 |========== c6g.4xlarge Graviton2 . 6449.11 |===================== c6a.4xlarge EPYC ...... 13134.46 |=========================================== c6i.4xlarge Xeon ...... 13888.40 |============================================= NAS Parallel Benchmarks 3.4 Test / Class: CG.C Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 6571.95 |================================ a1.4xlarge Graviton ... 1213.15 |====== c6g.4xlarge Graviton2 . 3520.86 |================= c6a.4xlarge EPYC ...... 6169.22 |============================== c6i.4xlarge Xeon ...... 9522.82 |============================================== NAS Parallel Benchmarks 3.4 Test / Class: EP.D Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 934.72 |======================================= a1.4xlarge Graviton ... 339.20 |============== c6g.4xlarge Graviton2 . 558.88 |======================= c6a.4xlarge EPYC ...... 466.21 |=================== c6i.4xlarge Xeon ...... 1103.22 |============================================== NAS Parallel Benchmarks 3.4 Test / Class: FT.C Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 11791.77 |========================== a1.4xlarge Graviton ... 2927.16 |====== c6g.4xlarge Graviton2 . 6244.48 |============== c6a.4xlarge EPYC ...... 18299.96 |======================================== c6i.4xlarge Xeon ...... 20423.57 |============================================= NAS Parallel Benchmarks 3.4 Test / Class: IS.D Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 1041.90 |============================================== a1.4xlarge Graviton ... 197.57 |========= c6g.4xlarge Graviton2 . 372.76 |================ c6a.4xlarge EPYC ...... 541.35 |======================== c6i.4xlarge Xeon ...... 861.57 |====================================== NAS Parallel Benchmarks 3.4 Test / Class: LU.C Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 7730.41 |========= a1.4xlarge Graviton ... 2558.12 |=== c6g.4xlarge Graviton2 . 5133.89 |====== c6a.4xlarge EPYC ...... 25140.55 |============================== c6i.4xlarge Xeon ...... 38136.77 |============================================= NAS Parallel Benchmarks 3.4 Test / Class: MG.C Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 13481.61 |======================= a1.4xlarge Graviton ... 3266.36 |====== c6g.4xlarge Graviton2 . 6720.68 |=========== c6a.4xlarge EPYC ...... 16826.43 |============================= c6i.4xlarge Xeon ...... 26298.81 |============================================= NAS Parallel Benchmarks 3.4 Test / Class: SP.C Total Mop/s > Higher Is Better c7g.4xlarge Graviton3 . 4467.19 |===================== a1.4xlarge Graviton ... 1293.80 |====== c6g.4xlarge Graviton2 . 2356.16 |=========== c6a.4xlarge EPYC ...... 8094.79 |======================================= c6i.4xlarge Xeon ...... 9563.22 |============================================== LeelaChessZero 0.28 Backend: BLAS Nodes Per Second > Higher Is Better c7g.4xlarge Graviton3 . 1103 |======================================= a1.4xlarge Graviton ... 135 |===== c6g.4xlarge Graviton2 . 864 |============================== c6a.4xlarge EPYC ...... 1091 |====================================== c6i.4xlarge Xeon ...... 1397 |================================================= LeelaChessZero 0.28 Backend: Eigen Nodes Per Second > Higher Is Better c7g.4xlarge Graviton3 . 1189 |======================================== a1.4xlarge Graviton ... 128 |==== c6g.4xlarge Graviton2 . 834 |============================ c6a.4xlarge EPYC ...... 1001 |================================= c6i.4xlarge Xeon ...... 1466 |================================================= Rodinia 3.1 Test: OpenMP LavaMD Seconds < Lower Is Better c7g.4xlarge Graviton3 . 143.33 |=================== a1.4xlarge Graviton ... 360.30 |=============================================== c6g.4xlarge Graviton2 . 215.67 |============================ c6a.4xlarge EPYC ...... 224.33 |============================= c6i.4xlarge Xeon ...... 281.39 |===================================== Rodinia 3.1 Test: OpenMP CFD Solver Seconds < Lower Is Better c7g.4xlarge Graviton3 . 10.48 |============ a1.4xlarge Graviton ... 41.45 |================================================ c6g.4xlarge Graviton2 . 17.04 |==================== c6a.4xlarge EPYC ...... 21.79 |========================= c6i.4xlarge Xeon ...... 20.45 |======================== Rodinia 3.1 Test: OpenMP Streamcluster Seconds < Lower Is Better c7g.4xlarge Graviton3 . 13.30 |============= a1.4xlarge Graviton ... 47.43 |================================================ c6g.4xlarge Graviton2 . 15.48 |================ c6a.4xlarge EPYC ...... 18.38 |=================== c6i.4xlarge Xeon ...... 23.51 |======================== Algebraic Multi-Grid Benchmark 1.2 Figure Of Merit > Higher Is Better c7g.4xlarge Graviton3 . 1258807333 |=========================================== a1.4xlarge Graviton ... 186716933 |====== c6g.4xlarge Graviton2 . 932652900 |================================ c6a.4xlarge EPYC ...... 267670700 |========= c6i.4xlarge Xeon ...... 661364767 |======================= Timed MrBayes Analysis 3.2.7 Primate Phylogeny Analysis Seconds < Lower Is Better c7g.4xlarge Graviton3 . 251.40 |================== a1.4xlarge Graviton ... 644.79 |=============================================== c6g.4xlarge Graviton2 . 384.75 |============================ c6a.4xlarge EPYC ...... 120.64 |========= c6i.4xlarge Xeon ...... 134.92 |========== Xcompact3d Incompact3d 2021-03-11 Input: input.i3d 129 Cells Per Direction Seconds < Lower Is Better c7g.4xlarge Graviton3 . 8.01671425 |====== a1.4xlarge Graviton ... 53.77062740 |========================================== c6g.4xlarge Graviton2 . 11.57335470 |========= c6a.4xlarge EPYC ...... 28.27976610 |====================== c6i.4xlarge Xeon ...... 17.86827720 |============== Xcompact3d Incompact3d 2021-03-11 Input: input.i3d 193 Cells Per Direction Seconds < Lower Is Better c7g.4xlarge Graviton3 . 29.13 |======= a1.4xlarge Graviton ... 182.58 |=============================================== c6g.4xlarge Graviton2 . 41.02 |=========== c6a.4xlarge EPYC ...... 110.77 |============================= c6i.4xlarge Xeon ...... 69.22 |================== LAMMPS Molecular Dynamics Simulator 29Oct2020 Model: Rhodopsin Protein ns/day > Higher Is Better c7g.4xlarge Graviton3 . 11.291 |=============================================== a1.4xlarge Graviton ... 3.245 |============== c6g.4xlarge Graviton2 . 7.935 |================================= c6a.4xlarge EPYC ...... 5.067 |===================== c6i.4xlarge Xeon ...... 6.220 |========================== LULESH 2.0.3 z/s > Higher Is Better c7g.4xlarge Graviton3 . 10940.94 |============================================= a1.4xlarge Graviton ... 2328.27 |========== c6g.4xlarge Graviton2 . 6016.16 |========================= c6a.4xlarge EPYC ...... 5452.11 |====================== c6i.4xlarge Xeon ...... 8112.37 |================================= WebP Image Encode 1.1 Encode Settings: Quality 100, Lossless Encode Time - Seconds < Lower Is Better c7g.4xlarge Graviton3 . 22.77 |================== a1.4xlarge Graviton ... 61.80 |================================================ c6g.4xlarge Graviton2 . 31.08 |======================== c6a.4xlarge EPYC ...... 26.71 |===================== c6i.4xlarge Xeon ...... 21.12 |================ WebP Image Encode 1.1 Encode Settings: Quality 100, Lossless, Highest Compression Encode Time - Seconds < Lower Is Better c7g.4xlarge Graviton3 . 48.21 |================== a1.4xlarge Graviton ... 124.71 |=============================================== c6g.4xlarge Graviton2 . 66.15 |========================= c6a.4xlarge EPYC ...... 48.68 |================== c6i.4xlarge Xeon ...... 41.81 |================ simdjson 1.0 Throughput Test: Kostya GB/s > Higher Is Better c7g.4xlarge Graviton3 . 1.94 |================================== a1.4xlarge Graviton ... 0.63 |=========== c6g.4xlarge Graviton2 . 1.19 |===================== c6a.4xlarge EPYC ...... 2.80 |================================================= c6i.4xlarge Xeon ...... 2.46 |=========================================== simdjson 1.0 Throughput Test: LargeRandom GB/s > Higher Is Better c7g.4xlarge Graviton3 . 0.70 |==================================== a1.4xlarge Graviton ... 0.30 |=============== c6g.4xlarge Graviton2 . 0.49 |========================= c6a.4xlarge EPYC ...... 0.95 |================================================= c6i.4xlarge Xeon ...... 0.86 |============================================ simdjson 1.0 Throughput Test: PartialTweets GB/s > Higher Is Better c7g.4xlarge Graviton3 . 2.62 |=================================== a1.4xlarge Graviton ... 0.78 |========== c6g.4xlarge Graviton2 . 1.51 |==================== c6a.4xlarge EPYC ...... 3.64 |================================================ c6i.4xlarge Xeon ...... 3.71 |================================================= simdjson 1.0 Throughput Test: DistinctUserID GB/s > Higher Is Better c7g.4xlarge Graviton3 . 2.69 |=============================== a1.4xlarge Graviton ... 0.80 |========= c6g.4xlarge Graviton2 . 1.53 |================= c6a.4xlarge EPYC ...... 4.30 |================================================= c6i.4xlarge Xeon ...... 4.30 |================================================= DaCapo Benchmark 9.12-MR1 Java Test: H2 msec < Lower Is Better c7g.4xlarge Graviton3 . 2951 |===================== a1.4xlarge Graviton ... 6740 |================================================= c6g.4xlarge Graviton2 . 3964 |============================= c6a.4xlarge EPYC ...... 3019 |====================== c6i.4xlarge Xeon ...... 2921 |===================== DaCapo Benchmark 9.12-MR1 Java Test: Jython msec < Lower Is Better c7g.4xlarge Graviton3 . 3940 |=============== a1.4xlarge Graviton ... 12997 |================================================ c6g.4xlarge Graviton2 . 5626 |===================== c6a.4xlarge EPYC ...... 4616 |================= c6i.4xlarge Xeon ...... 4013 |=============== DaCapo Benchmark 9.12-MR1 Java Test: Tradesoap msec < Lower Is Better c7g.4xlarge Graviton3 . 3524 |=============== a1.4xlarge Graviton ... 11182 |================================================ c6g.4xlarge Graviton2 . 4506 |=================== c6a.4xlarge EPYC ...... 4052 |================= c6i.4xlarge Xeon ...... 3815 |================ DaCapo Benchmark 9.12-MR1 Java Test: Tradebeans msec < Lower Is Better c7g.4xlarge Graviton3 . 3203 |================= a1.4xlarge Graviton ... 9045 |================================================= c6g.4xlarge Graviton2 . 4344 |======================== c6a.4xlarge EPYC ...... 3167 |================= c6i.4xlarge Xeon ...... 2928 |================ Zstd Compression 1.5.0 Compression Level: 3 - Compression Speed MB/s > Higher Is Better c6g.4xlarge Graviton2 . 2888.3 |============================= c6a.4xlarge EPYC ...... 2784.0 |============================ c7g.4xlarge Graviton3 . 4639.1 |=============================================== a1.4xlarge Graviton ... 633.9 |====== c6g.4xlarge Graviton2 . 2878.8 |============================= c6a.4xlarge EPYC ...... 2768.7 |============================ c6i.4xlarge Xeon ...... 3440.6 |=================================== Zstd Compression 1.5.0 Compression Level: 19 - Compression Speed MB/s > Higher Is Better c7g.4xlarge Graviton3 . 41.2 |================================================= a1.4xlarge Graviton ... 16.9 |==================== c6g.4xlarge Graviton2 . 34.6 |========================================= c6a.4xlarge EPYC ...... 30.0 |==================================== c6i.4xlarge Xeon ...... 38.1 |============================================= Zstd Compression 1.5.0 Compression Level: 19 - Decompression Speed MB/s > Higher Is Better c7g.4xlarge Graviton3 . 3050.3 |=============================================== a1.4xlarge Graviton ... 1121.7 |================= c6g.4xlarge Graviton2 . 2051.6 |================================ c6a.4xlarge EPYC ...... 2907.5 |============================================= c6i.4xlarge Xeon ...... 2582.0 |======================================== Zstd Compression 1.5.0 Compression Level: 19, Long Mode - Compression Speed MB/s > Higher Is Better c7g.4xlarge Graviton3 . 39.5 |================================================= a1.4xlarge Graviton ... 16.0 |==================== c6g.4xlarge Graviton2 . 31.0 |====================================== c6a.4xlarge EPYC ...... 25.9 |================================ c6i.4xlarge Xeon ...... 33.8 |========================================== Zstd Compression 1.5.0 Compression Level: 19, Long Mode - Decompression Speed MB/s > Higher Is Better c7g.4xlarge Graviton3 . 3240.6 |=============================================== a1.4xlarge Graviton ... 1213.9 |================== c6g.4xlarge Graviton2 . 2196.3 |================================ c6a.4xlarge EPYC ...... 2826.0 |========================================= c6i.4xlarge Xeon ...... 2666.1 |======================================= TSCP 1.81 AI Chess Performance Nodes Per Second > Higher Is Better c7g.4xlarge Graviton3 . 1370094 |============================================ a1.4xlarge Graviton ... 538500 |================= c6g.4xlarge Graviton2 . 872313 |============================ c6a.4xlarge EPYC ...... 1442631 |============================================== c6i.4xlarge Xeon ...... 1272596 |========================================= ACES DGEMM 1.0 Sustained Floating-Point Rate GFLOP/s > Higher Is Better c7g.4xlarge Graviton3 . 5.853864 |============================================= a1.4xlarge Graviton ... 0.891391 |======= c6g.4xlarge Graviton2 . 4.785123 |===================================== c6a.4xlarge EPYC ...... 2.432432 |=================== c6i.4xlarge Xeon ...... 2.230545 |================= Coremark 1.0 CoreMark Size 666 - Iterations Per Second Iterations/Sec > Higher Is Better c7g.4xlarge Graviton3 . 405413.86 |============================================ a1.4xlarge Graviton ... 203869.40 |====================== c6g.4xlarge Graviton2 . 315464.34 |================================== c6a.4xlarge EPYC ...... 345133.44 |===================================== c6i.4xlarge Xeon ...... 285378.84 |=============================== 7-Zip Compression 21.06 Test: Compression Rating MIPS > Higher Is Better c7g.4xlarge Graviton3 . 97824 |================================================ a1.4xlarge Graviton ... 32498 |================ c6g.4xlarge Graviton2 . 71285 |=================================== c6a.4xlarge EPYC ...... 62562 |=============================== c6i.4xlarge Xeon ...... 66631 |================================= 7-Zip Compression 21.06 Test: Decompression Rating MIPS > Higher Is Better c7g.4xlarge Graviton3 . 73054 |================================================ a1.4xlarge Graviton ... 40891 |=========================== c6g.4xlarge Graviton2 . 59445 |======================================= c6a.4xlarge EPYC ...... 57318 |====================================== c6i.4xlarge Xeon ...... 45653 |============================== Stockfish 13 Total Time Nodes Per Second > Higher Is Better c7g.4xlarge Graviton3 . 27608891 |============================================= a1.4xlarge Graviton ... 10980430 |================== c6g.4xlarge Graviton2 . 21679245 |=================================== c6a.4xlarge EPYC ...... 23857623 |======================================= c6i.4xlarge Xeon ...... 22081961 |==================================== asmFish 2018-07-23 1024 Hash Memory, 26 Depth Nodes/second > Higher Is Better c7g.4xlarge Graviton3 . 32134123 |============================================= a1.4xlarge Graviton ... 15331550 |===================== c6g.4xlarge Graviton2 . 26540482 |===================================== c6a.4xlarge EPYC ...... 26187688 |===================================== c6i.4xlarge Xeon ...... 23746200 |================================= libavif avifenc 0.10 Encoder Speed: 0 Seconds < Lower Is Better c7g.4xlarge Graviton3 . 256.84 |================ a1.4xlarge Graviton ... 768.30 |=============================================== c6g.4xlarge Graviton2 . 406.94 |========================= c6a.4xlarge EPYC ...... 195.53 |============ c6i.4xlarge Xeon ...... 204.99 |============= libavif avifenc 0.10 Encoder Speed: 2 Seconds < Lower Is Better c7g.4xlarge Graviton3 . 141.70 |=============== a1.4xlarge Graviton ... 449.02 |=============================================== c6g.4xlarge Graviton2 . 238.21 |========================= c6a.4xlarge EPYC ...... 93.95 |========== c6i.4xlarge Xeon ...... 97.74 |========== libavif avifenc 0.10 Encoder Speed: 6, Lossless Seconds < Lower Is Better c7g.4xlarge Graviton3 . 11.91 |================= a1.4xlarge Graviton ... 33.99 |================================================ c6g.4xlarge Graviton2 . 16.52 |======================= c6a.4xlarge EPYC ...... 16.39 |======================= c6i.4xlarge Xeon ...... 17.53 |========================= Timed Apache Compilation 2.4.41 Time To Compile Seconds < Lower Is Better c7g.4xlarge Graviton3 . 26.94 |================= a1.4xlarge Graviton ... 74.74 |================================================ c6g.4xlarge Graviton2 . 34.20 |====================== c6a.4xlarge EPYC ...... 23.53 |=============== c6i.4xlarge Xeon ...... 22.53 |============== Timed Gem5 Compilation 21.2 Time To Compile Seconds < Lower Is Better c7g.4xlarge Graviton3 . 391.17 |================ a1.4xlarge Graviton ... 1155.62 |============================================== c6g.4xlarge Graviton2 . 488.81 |=================== c6a.4xlarge EPYC ...... 515.20 |===================== c6i.4xlarge Xeon ...... 469.94 |=================== Timed ImageMagick Compilation 6.9.0 Time To Compile Seconds < Lower Is Better c7g.4xlarge Graviton3 . 27.90 |============== a1.4xlarge Graviton ... 93.63 |================================================ c6g.4xlarge Graviton2 . 40.33 |===================== c6a.4xlarge EPYC ...... 32.63 |================= c6i.4xlarge Xeon ...... 29.74 |=============== Timed LLVM Compilation 13.0 Build System: Ninja Seconds < Lower Is Better c7g.4xlarge Graviton3 . 544.93 |============== a1.4xlarge Graviton ... 1784.60 |============================================== c6g.4xlarge Graviton2 . 682.98 |================== c6a.4xlarge EPYC ...... 760.34 |==================== c6i.4xlarge Xeon ...... 685.70 |================== Timed Node.js Compilation 17.3 Time To Compile Seconds < Lower Is Better c7g.4xlarge Graviton3 . 497.58 |============= a1.4xlarge Graviton ... 1765.91 |============================================== c6g.4xlarge Graviton2 . 628.40 |================ c6a.4xlarge EPYC ...... 664.35 |================= c6i.4xlarge Xeon ...... 604.62 |================ Timed PHP Compilation 7.4.2 Time To Compile Seconds < Lower Is Better c7g.4xlarge Graviton3 . 69.48 |================= a1.4xlarge Graviton ... 196.03 |=============================================== c6g.4xlarge Graviton2 . 88.90 |===================== c6a.4xlarge EPYC ...... 67.08 |================ c6i.4xlarge Xeon ...... 64.34 |=============== Build2 0.13 Time To Compile Seconds < Lower Is Better c7g.4xlarge Graviton3 . 115.02 |=============== a1.4xlarge Graviton ... 353.91 |=============================================== c6g.4xlarge Graviton2 . 142.28 |=================== c6a.4xlarge EPYC ...... 150.99 |==================== c6i.4xlarge Xeon ...... 136.80 |================== C-Ray 1.1 Total Time - 4K, 16 Rays Per Pixel Seconds < Lower Is Better c7g.4xlarge Graviton3 . 38.52 |================= a1.4xlarge Graviton ... 104.76 |=============================================== c6g.4xlarge Graviton2 . 62.32 |============================ c6a.4xlarge EPYC ...... 69.35 |=============================== c6i.4xlarge Xeon ...... 92.55 |========================================== POV-Ray 3.7.0.7 Trace Time Seconds < Lower Is Better c7g.4xlarge Graviton3 . 37.86 |=================== a1.4xlarge Graviton ... 93.80 |================================================ c6g.4xlarge Graviton2 . 51.05 |========================== c6a.4xlarge EPYC ...... 49.44 |========================= c6i.4xlarge Xeon ...... 52.78 |=========================== m-queens 1.2 Time To Solve Seconds < Lower Is Better c7g.4xlarge Graviton3 . 66.82 |============================ a1.4xlarge Graviton ... 110.37 |=============================================== c6g.4xlarge Graviton2 . 75.22 |================================ c6a.4xlarge EPYC ...... 72.33 |=============================== c6i.4xlarge Xeon ...... 91.23 |======================================= N-Queens 1.0 Elapsed Time Seconds < Lower Is Better c7g.4xlarge Graviton3 . 21.54 |================================ a1.4xlarge Graviton ... 32.29 |================================================ c6g.4xlarge Graviton2 . 23.14 |================================== c6a.4xlarge EPYC ...... 16.38 |======================== c6i.4xlarge Xeon ...... 18.84 |============================ Ngspice 34 Circuit: C2670 Seconds < Lower Is Better c7g.4xlarge Graviton3 . 198.22 |==================== a1.4xlarge Graviton ... 473.90 |=============================================== c6g.4xlarge Graviton2 . 263.72 |========================== c6a.4xlarge EPYC ...... 245.89 |======================== c6i.4xlarge Xeon ...... 147.89 |=============== Ngspice 34 Circuit: C7552 Seconds < Lower Is Better c7g.4xlarge Graviton3 . 191.29 |=================== a1.4xlarge Graviton ... 480.79 |=============================================== c6g.4xlarge Graviton2 . 255.21 |========================= c6a.4xlarge EPYC ...... 180.36 |================== c6i.4xlarge Xeon ...... 161.08 |================ Google SynthMark 20201109 Test: VoiceMark_100 Voices > Higher Is Better c7g.4xlarge Graviton3 . 675.64 |=============================================== a1.4xlarge Graviton ... 331.07 |======================= c6g.4xlarge Graviton2 . 470.39 |================================= c6a.4xlarge EPYC ...... 663.07 |============================================== c6i.4xlarge Xeon ...... 565.69 |======================================= SecureMark 1.0.4 Benchmark: SecureMark-TLS marks > Higher Is Better c7g.4xlarge Graviton3 . 183708 |===================================== a1.4xlarge Graviton ... 74356 |=============== c6g.4xlarge Graviton2 . 120301 |========================= c6a.4xlarge EPYC ...... 213288 |=========================================== c6i.4xlarge Xeon ...... 230549 |=============================================== OpenSSL 3.0 Algorithm: SHA256 byte/s > Higher Is Better c7g.4xlarge Graviton3 . 13722045973 |========================================== a1.4xlarge Graviton ... 6785689517 |===================== c6g.4xlarge Graviton2 . 10723184083 |================================= c6a.4xlarge EPYC ...... 11691403353 |==================================== c6i.4xlarge Xeon ...... 7096993937 |====================== OpenSSL 3.0 Algorithm: RSA4096 sign/s > Higher Is Better c7g.4xlarge Graviton3 . 2546.4 |=============================================== a1.4xlarge Graviton ... 588.3 |=========== c6g.4xlarge Graviton2 . 660.6 |============ c6a.4xlarge EPYC ...... 2088.9 |======================================= c6i.4xlarge Xeon ...... 2161.3 |======================================== OpenSSL 3.0 Algorithm: RSA4096 verify/s > Higher Is Better c7g.4xlarge Graviton3 . 178460.4 |============================================= a1.4xlarge Graviton ... 45328.6 |=========== c6g.4xlarge Graviton2 . 53951.5 |============== c6a.4xlarge EPYC ...... 136784.2 |================================== c6i.4xlarge Xeon ...... 140964.4 |==================================== Liquid-DSP 2021.01.31 Threads: 16 - Buffer Length: 256 - Filter Length: 57 samples/s > Higher Is Better c7g.4xlarge Graviton3 . 383606667 |================================= a1.4xlarge Graviton ... 165513333 |============== c6g.4xlarge Graviton2 . 262890000 |======================= c6a.4xlarge EPYC ...... 509746667 |============================================ c6i.4xlarge Xeon ...... 373100000 |================================ GROMACS 2022.1 Implementation: MPI CPU - Input: water_GMX50_bare Ns Per Day > Higher Is Better c7g.4xlarge Graviton3 . 1.128 |===================================== a1.4xlarge Graviton ... 0.316 |========== c6g.4xlarge Graviton2 . 0.781 |========================== c6a.4xlarge EPYC ...... 1.004 |================================= c6i.4xlarge Xeon ...... 1.452 |================================================ TensorFlow Lite 2022-05-18 Model: SqueezeNet Microseconds < Lower Is Better c7g.4xlarge Graviton3 . 3257.94 |============ a1.4xlarge Graviton ... 12014.70 |============================================= c6g.4xlarge Graviton2 . 3969.35 |=============== c6a.4xlarge EPYC ...... 3103.12 |============ c6i.4xlarge Xeon ...... 2983.93 |=========== TensorFlow Lite 2022-05-18 Model: Inception V4 Microseconds < Lower Is Better c7g.4xlarge Graviton3 . 41855.1 |========== a1.4xlarge Graviton ... 188910.0 |============================================= c6g.4xlarge Graviton2 . 46793.9 |=========== c6a.4xlarge EPYC ...... 44920.6 |=========== c6i.4xlarge Xeon ...... 41185.7 |========== TensorFlow Lite 2022-05-18 Model: NASNet Mobile Microseconds < Lower Is Better c7g.4xlarge Graviton3 . 11591.90 |================= a1.4xlarge Graviton ... 30986.70 |============================================= c6g.4xlarge Graviton2 . 14985.40 |====================== c6a.4xlarge EPYC ...... 9266.86 |============= c6i.4xlarge Xeon ...... 10900.60 |================ TensorFlow Lite 2022-05-18 Model: Mobilenet Float Microseconds < Lower Is Better c7g.4xlarge Graviton3 . 2156.60 |========== a1.4xlarge Graviton ... 9990.15 |============================================== c6g.4xlarge Graviton2 . 2500.87 |============ c6a.4xlarge EPYC ...... 2159.72 |========== c6i.4xlarge Xeon ...... 1965.07 |========= TensorFlow Lite 2022-05-18 Model: Mobilenet Quant Microseconds < Lower Is Better c7g.4xlarge Graviton3 . 1502.95 |============ a1.4xlarge Graviton ... 5724.66 |============================================== c6g.4xlarge Graviton2 . 1980.24 |================ c6a.4xlarge EPYC ...... 3847.96 |=============================== c6i.4xlarge Xeon ...... 3967.39 |================================ TensorFlow Lite 2022-05-18 Model: Inception ResNet V2 Microseconds < Lower Is Better c7g.4xlarge Graviton3 . 40051.3 |=========== a1.4xlarge Graviton ... 171169.0 |============================================= c6g.4xlarge Graviton2 . 45955.7 |============ c6a.4xlarge EPYC ...... 41366.6 |=========== c6i.4xlarge Xeon ...... 41179.7 |=========== ASTC Encoder 3.2 Preset: Thorough Seconds < Lower Is Better c7g.4xlarge Graviton3 . 13.9248 |=================== a1.4xlarge Graviton ... 33.5198 |============================================== c6g.4xlarge Graviton2 . 16.5222 |======================= c6a.4xlarge EPYC ...... 7.9818 |=========== c6i.4xlarge Xeon ...... 7.2625 |========== ASTC Encoder 3.2 Preset: Exhaustive Seconds < Lower Is Better c7g.4xlarge Graviton3 . 139.38 |======================== a1.4xlarge Graviton ... 277.77 |=============================================== c6g.4xlarge Graviton2 . 159.20 |=========================== c6a.4xlarge EPYC ...... 72.39 |============ c6i.4xlarge Xeon ...... 69.64 |============ Stress-NG 0.14 Test: Crypto Bogo Ops/s > Higher Is Better c7g.4xlarge Graviton3 . 23181.81 |============================================= a1.4xlarge Graviton ... 11985.38 |======================= c6g.4xlarge Graviton2 . 17924.18 |=================================== c6a.4xlarge EPYC ...... 13556.06 |========================== c6i.4xlarge Xeon ...... 10210.34 |==================== Stress-NG 0.14 Test: IO_uring Bogo Ops/s > Higher Is Better c7g.4xlarge Graviton3 . 843015.78 |=================================== a1.4xlarge Graviton ... 918172.37 |====================================== c6g.4xlarge Graviton2 . 770521.81 |================================ c6a.4xlarge EPYC ...... 768723.46 |================================ c6i.4xlarge Xeon ...... 1037943.37 |=========================================== Stress-NG 0.14 Test: CPU Stress Bogo Ops/s > Higher Is Better c7g.4xlarge Graviton3 . 5029.71 |================= a1.4xlarge Graviton ... 2366.00 |======== c6g.4xlarge Graviton2 . 3404.94 |============ c6a.4xlarge EPYC ...... 13304.50 |============================================= c6i.4xlarge Xeon ...... 12527.16 |========================================== Stress-NG 0.14 Test: Vector Math Bogo Ops/s > Higher Is Better c7g.4xlarge Graviton3 . 55258.17 |============================================= a1.4xlarge Graviton ... 27341.47 |====================== c6g.4xlarge Graviton2 . 37753.89 |=============================== c6a.4xlarge EPYC ...... 53787.61 |============================================ c6i.4xlarge Xeon ...... 40140.30 |================================= Stress-NG 0.14 Test: Memory Copying Bogo Ops/s > Higher Is Better c7g.4xlarge Graviton3 . 6693.32 |============================================== a1.4xlarge Graviton ... 798.24 |===== c6g.4xlarge Graviton2 . 2903.00 |==================== c6a.4xlarge EPYC ...... 3551.80 |======================== c6i.4xlarge Xeon ...... 3150.49 |====================== GPAW 22.1 Input: Carbon Nanotube Seconds < Lower Is Better c7g.4xlarge Graviton3 . 155.18 |========= a1.4xlarge Graviton ... 769.35 |=============================================== c6g.4xlarge Graviton2 . 215.53 |============= c6a.4xlarge EPYC ...... 302.96 |=================== c6i.4xlarge Xeon ...... 202.11 |============ PyBench 2018-02-16 Total For Average Test Times Milliseconds < Lower Is Better c7g.4xlarge Graviton3 . 1185 |================= a1.4xlarge Graviton ... 3452 |================================================= c6g.4xlarge Graviton2 . 1741 |========================= c6a.4xlarge EPYC ...... 1961 |============================ c6i.4xlarge Xeon ...... 997 |============== nginx 1.21.1 Concurrent Requests: 100 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 345710.87 |======================================= a1.4xlarge Graviton ... 143155.48 |================ c6g.4xlarge Graviton2 . 307349.36 |=================================== c6a.4xlarge EPYC ...... 388010.76 |============================================ c6i.4xlarge Xeon ...... 356302.84 |======================================== nginx 1.21.1 Concurrent Requests: 200 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 352380.98 |======================================== a1.4xlarge Graviton ... 141436.20 |================ c6g.4xlarge Graviton2 . 308938.67 |=================================== c6a.4xlarge EPYC ...... 390932.79 |============================================ c6i.4xlarge Xeon ...... 356829.93 |======================================== nginx 1.21.1 Concurrent Requests: 500 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 346613.34 |======================================= a1.4xlarge Graviton ... 139414.84 |================ c6g.4xlarge Graviton2 . 310596.58 |=================================== c6a.4xlarge EPYC ...... 389030.11 |============================================ c6i.4xlarge Xeon ...... 351672.92 |======================================== nginx 1.21.1 Concurrent Requests: 1000 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 346814.75 |======================================= a1.4xlarge Graviton ... 138205.11 |================ c6g.4xlarge Graviton2 . 308213.13 |=================================== c6a.4xlarge EPYC ...... 388657.76 |============================================ c6i.4xlarge Xeon ...... 347345.49 |======================================= ONNX Runtime 1.11 Model: GPT-2 - Device: CPU - Executor: Standard Inferences Per Minute > Higher Is Better c7g.4xlarge Graviton3 . 7990 |================================================= a1.4xlarge Graviton ... 2312 |============== c6g.4xlarge Graviton2 . 6948 |=========================================== c6a.4xlarge EPYC ...... 5617 |================================== c6i.4xlarge Xeon ...... 7944 |================================================= ONNX Runtime 1.11 Model: bertsquad-12 - Device: CPU - Executor: Standard Inferences Per Minute > Higher Is Better c7g.4xlarge Graviton3 . 407 |========================== a1.4xlarge Graviton ... 115 |======= c6g.4xlarge Graviton2 . 322 |===================== c6a.4xlarge EPYC ...... 488 |================================ c6i.4xlarge Xeon ...... 773 |================================================== ONNX Runtime 1.11 Model: fcn-resnet101-11 - Device: CPU - Executor: Standard Inferences Per Minute > Higher Is Better c7g.4xlarge Graviton3 . 38 |============== a1.4xlarge Graviton ... 10 |==== c6g.4xlarge Graviton2 . 28 |========== c6a.4xlarge EPYC ...... 65 |======================= c6i.4xlarge Xeon ...... 139 |================================================== ONNX Runtime 1.11 Model: ArcFace ResNet-100 - Device: CPU - Executor: Standard Inferences Per Minute > Higher Is Better c7g.4xlarge Graviton3 . 609 |====================== a1.4xlarge Graviton ... 165 |====== c6g.4xlarge Graviton2 . 334 |============ c6a.4xlarge EPYC ...... 1192 |=========================================== c6i.4xlarge Xeon ...... 1374 |================================================= ONNX Runtime 1.11 Model: super-resolution-10 - Device: CPU - Executor: Standard Inferences Per Minute > Higher Is Better c7g.4xlarge Graviton3 . 2817 |===================================== a1.4xlarge Graviton ... 757 |========== c6g.4xlarge Graviton2 . 2072 |=========================== c6a.4xlarge EPYC ...... 3696 |================================================= c6i.4xlarge Xeon ...... 3450 |============================================== Apache HTTP Server 2.4.48 Concurrent Requests: 100 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 67231.88 |=================================== a1.4xlarge Graviton ... 18636.43 |========== c6g.4xlarge Graviton2 . 46995.35 |======================== c6a.4xlarge EPYC ...... 77567.69 |======================================== c6i.4xlarge Xeon ...... 86545.57 |============================================= Apache HTTP Server 2.4.48 Concurrent Requests: 200 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 73676.95 |=================================== a1.4xlarge Graviton ... 20887.58 |========== c6g.4xlarge Graviton2 . 50059.97 |======================== c6a.4xlarge EPYC ...... 83070.00 |======================================== c6i.4xlarge Xeon ...... 94458.22 |============================================= Apache HTTP Server 2.4.48 Concurrent Requests: 500 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 73546.32 |==================================== a1.4xlarge Graviton ... 20133.49 |========== c6g.4xlarge Graviton2 . 50077.81 |========================= c6a.4xlarge EPYC ...... 81995.64 |======================================== c6i.4xlarge Xeon ...... 91746.57 |============================================= Apache HTTP Server 2.4.48 Concurrent Requests: 1000 Requests Per Second > Higher Is Better c7g.4xlarge Graviton3 . 72719.33 |========================================= a1.4xlarge Graviton ... 19278.68 |=========== c6g.4xlarge Graviton2 . 46629.45 |========================== c6a.4xlarge EPYC ...... 71537.11 |======================================== c6i.4xlarge Xeon ...... 79830.96 |============================================= PHPBench 0.8.1 PHP Benchmark Suite Score > Higher Is Better c7g.4xlarge Graviton3 . 666484 |====================================== a1.4xlarge Graviton ... 241259 |============== c6g.4xlarge Graviton2 . 449855 |========================== c6a.4xlarge EPYC ...... 480741 |=========================== c6i.4xlarge Xeon ...... 828186 |===============================================