Core i9 7980XE GCC 9 Compiler Benchmarks Intel Core i9-7980XE testing with GCC 9 compiler by Michael Larabel. -O3 -march=skylake: Processor: Intel Core i9-7980XE @ 4.20GHz (18 Cores / 36 Threads), Motherboard: ASUS PRIME X299-A (1704 BIOS), Chipset: Intel Sky Lake-E DMI3 Registers, Memory: 16384MB, Disk: Samsung SSD 970 EVO 500GB, Graphics: NVIDIA NV120 12GB, Audio: Realtek ALC1220, Monitor: ASUS PB278, Network: Intel I219-V OS: Clear Linux OS 29350, Kernel: 5.0.14-753.native (x86_64), Desktop: GNOME Shell 3.32.1, Display Server: X Server 1.20.4, Display Driver: nouveau 1.0.16, OpenGL: 4.3 Mesa 19.1.0-devel, Compiler: GCC 9.1.1 20190509 gcc-9-branch@271025 + clang (Clear Linux OS for Intel Architecture) 9.1.1 20190509 gcc-9-branch@271025 + LLVM 8.0.0, File-System: ext4, Screen Resolution: 2560x1440 -O3 -march=skylake-avx512 -mprefer-vector-width=128: Processor: Intel Core i9-7980XE @ 4.20GHz (18 Cores / 36 Threads), Motherboard: ASUS PRIME X299-A (1704 BIOS), Chipset: Intel Sky Lake-E DMI3 Registers, Memory: 16384MB, Disk: Samsung SSD 970 EVO 500GB, Graphics: NVIDIA NV120 12GB, Audio: Realtek ALC1220, Monitor: ASUS PB278, Network: Intel I219-V OS: Clear Linux OS 29350, Kernel: 5.0.14-753.native (x86_64), Desktop: GNOME Shell 3.32.1, Display Server: X Server 1.20.4, Display Driver: nouveau 1.0.16, OpenGL: 4.3 Mesa 19.1.0-devel, Compiler: GCC 9.1.1 20190509 gcc-9-branch@271025 + clang (Clear Linux OS for Intel Architecture) 9.1.1 20190509 gcc-9-branch@271025 + LLVM 8.0.0, File-System: ext4, Screen Resolution: 2560x1440 -O3 -march=skylake-avx512 -mprefer-vector-width=256: Processor: Intel Core i9-7980XE @ 4.20GHz (18 Cores / 36 Threads), Motherboard: ASUS PRIME X299-A (1704 BIOS), Chipset: Intel Sky Lake-E DMI3 Registers, Memory: 16384MB, Disk: Samsung SSD 970 EVO 500GB, Graphics: NVIDIA NV120 12GB, Audio: Realtek ALC1220, Monitor: ASUS PB278, Network: Intel I219-V OS: Clear Linux OS 29350, Kernel: 5.0.14-753.native (x86_64), Desktop: GNOME Shell 3.32.1, Display Server: X Server 1.20.4, Display Driver: nouveau 1.0.16, OpenGL: 4.3 Mesa 19.1.0-devel, Compiler: GCC 9.1.1 20190509 gcc-9-branch@271025 + clang (Clear Linux OS for Intel Architecture) 9.1.1 20190509 gcc-9-branch@271025 + LLVM 8.0.0, File-System: ext4, Screen Resolution: 2560x1440 -O3 -march=skylake-avx512 -mprefer-vector-width=512: Processor: Intel Core i9-7980XE @ 4.20GHz (18 Cores / 36 Threads), Motherboard: ASUS PRIME X299-A (1704 BIOS), Chipset: Intel Sky Lake-E DMI3 Registers, Memory: 16384MB, Disk: Samsung SSD 970 EVO 500GB, Graphics: NVIDIA NV120 12GB, Audio: Realtek ALC1220, Monitor: ASUS PB278, Network: Intel I219-V OS: Clear Linux OS 29350, Kernel: 5.0.14-753.native (x86_64), Desktop: GNOME Shell 3.32.1, Display Server: X Server 1.20.4, Display Driver: nouveau 1.0.16, OpenGL: 4.3 Mesa 19.1.0-devel, Compiler: GCC 9.1.1 20190509 gcc-9-branch@271025 + clang (Clear Linux OS for Intel Architecture) 9.1.1 20190509 gcc-9-branch@271025 + LLVM 8.0.0, File-System: ext4, Screen Resolution: 2560x1440 CppPerformanceBenchmarks 9 Test: Random Numbers Seconds < Lower Is Better -O3 -march=skylake .................................. 846 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 821 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 857 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 846 |==================== MKL-DNN 2019-04-16 Harness: Deconvolution Batch deconv_all - Data Type: f32 ms < Lower Is Better -O3 -march=skylake .................................. 10446 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 10397 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 10405 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 10424 |================== CppPerformanceBenchmarks 9 Test: Math Library Seconds < Lower Is Better -O3 -march=skylake .................................. 260 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 262 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 264 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 263 |==================== FFTW 3.3.6 Build: Float + SSE - Size: 2D FFT Size 4096 Mflops > Higher Is Better -O3 -march=skylake .................................. 17183 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 17522 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 17326 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 17334 |================== FFTW 3.3.6 Build: Stock - Size: 2D FFT Size 4096 Mflops > Higher Is Better -O3 -march=skylake .................................. 6327 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 6863 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 6615 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 5673 |================ MKL-DNN 2019-04-16 Harness: Convolution Batch conv_googlenet_v3 - Data Type: f32 ms < Lower Is Better -O3 -march=skylake .................................. 752 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 749 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 751 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 751 |==================== Parboil 2.5 Test: OpenMP MRI Gridding Seconds < Lower Is Better -O3 -march=skylake .................................. 132 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 135 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 134 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 132 |==================== NGINX Benchmark 1.9.9 Static Web Page Serving Requests Per Second > Higher Is Better -O3 -march=skylake .................................. 47570 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 46938 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 49221 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 44072 |================ CppPerformanceBenchmarks 9 Test: Stepanov Vector Seconds < Lower Is Better -O3 -march=skylake .................................. 82.98 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 82.88 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 83.16 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 98.92 |================== PostgreSQL pgbench 10.3 Scaling: Buffer Test - Test: Normal Load - Mode: Read Only TPS > Higher Is Better -O3 -march=skylake .................................. 520100 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 519930 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 521901 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 498539 |================ PostgreSQL pgbench 10.3 Scaling: Buffer Test - Test: Normal Load - Mode: Read Write TPS > Higher Is Better -O3 -march=skylake .................................. 13311 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 13495 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 13576 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 13424 |================== Timed Linux Kernel Compilation 4.18 Time To Compile Seconds < Lower Is Better -O3 -march=skylake .................................. 36.34 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 36.33 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 36.93 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 37.36 |================== Parboil 2.5 Test: OpenMP LBM Seconds < Lower Is Better -O3 -march=skylake .................................. 68.54 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 68.35 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 68.26 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 68.18 |================== Stockfish 9 Total Time Nodes Per Second > Higher Is Better -O3 -march=skylake .................................. 47112136 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 46810779 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 47046623 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 46997822 |=============== Timed MrBayes Analysis 3.1.2 Primate Phylogeny Analysis Seconds < Lower Is Better -O3 -march=skylake .................................. 65.67 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 63.06 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 63.20 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 65.66 |================== NAS Parallel Benchmarks 3.3.1 Test / Class: LU.C Total Mop/s > Higher Is Better -O3 -march=skylake .................................. 33669 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 34962 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 35007 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 33676 |================= GraphicsMagick 1.3.30 Operation: Sharpen Iterations Per Minute > Higher Is Better -O3 -march=skylake .................................. 204 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 207 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 206 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 186 |================== GraphicsMagick 1.3.30 Operation: Noise-Gaussian Iterations Per Minute > Higher Is Better -O3 -march=skylake .................................. 198 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 197 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 198 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 195 |==================== GraphicsMagick 1.3.30 Operation: Swirl Iterations Per Minute > Higher Is Better -O3 -march=skylake .................................. 252 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 251 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 251 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 228 |================== GraphicsMagick 1.3.30 Operation: Enhanced Iterations Per Minute > Higher Is Better -O3 -march=skylake .................................. 227 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 225 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 224 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 196 |================= GraphicsMagick 1.3.30 Operation: Rotate Iterations Per Minute > Higher Is Better -O3 -march=skylake .................................. 291 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 290 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 290 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 250 |================= GraphicsMagick 1.3.30 Operation: HWB Color Space Iterations Per Minute > Higher Is Better -O3 -march=skylake .................................. 288 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 289 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 288 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 253 |================== GraphicsMagick 1.3.30 Operation: Resizing Iterations Per Minute > Higher Is Better -O3 -march=skylake .................................. 277 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 277 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 275 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 243 |================== Memcached mcperf 1.5.10 Method: Get Operations Per Second > Higher Is Better -O3 -march=skylake .................................. 117212 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 107497 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 117621 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 112844 |================ Himeno Benchmark 3.0 Poisson Pressure Solver MFLOPS > Higher Is Better -O3 -march=skylake .................................. 3302 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 3310 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 3292 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2814 |================ CppPerformanceBenchmarks 9 Test: Atol Seconds < Lower Is Better -O3 -march=skylake .................................. 57.08 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 57.07 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 57.00 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 57.06 |================== Rodinia 2.4 Test: OpenMP Streamcluster Seconds < Lower Is Better -O3 -march=skylake .................................. 13.63 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 13.72 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 13.88 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 14.12 |================== Redis 4.0.8 Test: LPUSH Requests Per Second > Higher Is Better -O3 -march=skylake .................................. 2059237 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 1962269 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2013442 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2026656 |================ Redis 4.0.8 Test: GET Requests Per Second > Higher Is Better -O3 -march=skylake .................................. 2958563 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2992962 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2944968 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 3002750 |================ Timed PHP Compilation 7.1.9 Time To Compile Seconds < Lower Is Better -O3 -march=skylake .................................. 51.80 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 51.84 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 51.90 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 52.19 |================== Redis 4.0.8 Test: LPOP Requests Per Second > Higher Is Better -O3 -march=skylake .................................. 3140283 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 3095145 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 3139485 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 3059844 |================ m-queens 1.2 Time To Solve Seconds < Lower Is Better -O3 -march=skylake .................................. 48.42 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 48.42 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 48.40 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 48.41 |================== Memcached mcperf 1.5.10 Method: Delete Operations Per Second > Higher Is Better -O3 -march=skylake .................................. 120897 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 122131 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 114714 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 115719 |================ 7-Zip Compression 16.02 Compress Speed Test MIPS > Higher Is Better -O3 -march=skylake .................................. 100157 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 99546 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 99960 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 100416 |================= Redis 4.0.8 Test: SET Requests Per Second > Higher Is Better -O3 -march=skylake .................................. 2163293 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2189741 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 1971745 |============== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2135179 |================ Redis 4.0.8 Test: SADD Requests Per Second > Higher Is Better -O3 -march=skylake .................................. 2339849 |============== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2624995 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2449355 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2383273 |=============== Cpuminer-Opt 3.8.8.1 Algorithm: m7m kH/s - Hash Speed > Higher Is Better -O3 -march=skylake .................................. 499 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 503 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 506 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 479 |=================== Cpuminer-Opt 3.8.8.1 Algorithm: sha256t kH/s - Hash Speed > Higher Is Better -O3 -march=skylake .................................. 125653 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 124672 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 124030 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 124567 |================= lzbench 2017-08-08 Test: XZ 0 - Process: Decompression MB/s > Higher Is Better -O3 -march=skylake .................................. 108 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 107 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 107 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 108 |==================== lzbench 2017-08-08 Test: XZ 0 - Process: Compression MB/s > Higher Is Better -O3 -march=skylake .................................. 39 |===================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 39 |===================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 39 |===================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 39 |===================== NAS Parallel Benchmarks 3.3.1 Test / Class: BT.A Total Mop/s > Higher Is Better -O3 -march=skylake .................................. 4391 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 4773 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 4784 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 4391 |================= Cpuminer-Opt 3.8.8.1 Algorithm: deep kH/s - Hash Speed > Higher Is Better -O3 -march=skylake .................................. 15323 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 15497 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 15490 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 14490 |================= lzbench 2017-08-08 Test: Brotli 0 - Process: Decompression MB/s > Higher Is Better -O3 -march=skylake .................................. 620 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 620 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 620 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 621 |==================== lzbench 2017-08-08 Test: Brotli 0 - Process: Compression MB/s > Higher Is Better -O3 -march=skylake .................................. 430 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 429 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 429 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 428 |==================== lzbench 2017-08-08 Test: Zstd 1 - Process: Decompression MB/s > Higher Is Better -O3 -march=skylake .................................. 1137 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 1138 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 1137 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 1138 |=================== lzbench 2017-08-08 Test: Zstd 1 - Process: Compression MB/s > Higher Is Better -O3 -march=skylake .................................. 451 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 451 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 452 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 452 |==================== lzbench 2017-08-08 Test: Libdeflate 1 - Process: Decompression MB/s > Higher Is Better -O3 -march=skylake .................................. 1189 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 1190 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 1191 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 1191 |=================== lzbench 2017-08-08 Test: Libdeflate 1 - Process: Compression MB/s > Higher Is Better -O3 -march=skylake .................................. 248 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 249 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 249 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 249 |==================== C-Ray 1.1 Total Time - 4K, 16 Rays Per Pixel Seconds < Lower Is Better -O3 -march=skylake .................................. 30.87 |=============== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 33.34 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 33.47 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 35.86 |================== CppPerformanceBenchmarks 9 Test: Stepanov Abstraction Seconds < Lower Is Better -O3 -march=skylake .................................. 32.49 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 32.44 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 32.44 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 37.27 |================== Cpuminer-Opt 3.8.8.1 Algorithm: myr-gr kH/s - Hash Speed > Higher Is Better -O3 -march=skylake .................................. 16360 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 15560 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 15695 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 15310 |================= Cpuminer-Opt 3.8.8.1 Algorithm: lbry kH/s - Hash Speed > Higher Is Better -O3 -march=skylake .................................. 48547 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 52793 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 53473 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 53620 |================== TTSIOD 3D Renderer 2.3b Phong Rendering With Soft-Shadow Mapping FPS > Higher Is Better -O3 -march=skylake .................................. 710 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 738 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 729 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 669 |================== LuaJIT 2.1-git Test: Composite Mflops > Higher Is Better -O3 -march=skylake .................................. 1606 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 1605 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 1607 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 1608 |=================== AOBench Size: 2048 x 2048 - Total Time Seconds < Lower Is Better -O3 -march=skylake .................................. 31.17 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 31.15 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 31.11 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 31.29 |================== Cpuminer-Opt 3.8.8.1 Algorithm: skein kH/s - Hash Speed > Higher Is Better -O3 -march=skylake .................................. 56227 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 62813 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 62560 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 62027 |================== CppPerformanceBenchmarks 9 Test: Ctype Seconds < Lower Is Better -O3 -march=skylake .................................. 29.69 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 30.57 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 30.89 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 30.56 |================== Apache Benchmark 2.4.29 Static Web Page Serving Requests Per Second > Higher Is Better -O3 -march=skylake .................................. 36040 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 35982 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 35946 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 33240 |================= Crafty 25.2 Elapsed Time Nodes Per Second > Higher Is Better -O3 -march=skylake .................................. 8852846 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 8840395 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 8845689 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 8840945 |================ Botan 2.8.0 Test: AES-256 - Decrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 4212 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 3840 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 3855 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 3839 |================= Botan 2.8.0 Test: AES-256 - Encrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 4220 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 3847 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 3833 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 3835 |================= MKL-DNN 2019-04-16 Harness: Deconvolution Batch deconv_1d - Data Type: f32 ms < Lower Is Better -O3 -march=skylake .................................. 23.03 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 23.01 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 22.93 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 23.07 |================== MKL-DNN 2019-04-16 Harness: Convolution Batch conv_alexnet - Data Type: f32 ms < Lower Is Better -O3 -march=skylake .................................. 1822 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 1820 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 1819 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 1819 |=================== NAS Parallel Benchmarks 3.3.1 Test / Class: SP.A Total Mop/s > Higher Is Better -O3 -march=skylake .................................. 4062 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 4073 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 4185 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 4080 |=================== Botan 2.8.0 Test: Twofish - Decrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 393 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 395 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 395 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 395 |==================== Botan 2.8.0 Test: Twofish - Encrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 400 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 401 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 401 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 401 |==================== Botan 2.8.0 Test: Blowfish - Decrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 293 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 294 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 293 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 294 |==================== Botan 2.8.0 Test: Blowfish - Encrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 295 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 295 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 295 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 295 |==================== Botan 2.8.0 Test: CAST-256 - Decrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 145 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 146 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 146 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 146 |==================== Botan 2.8.0 Test: CAST-256 - Encrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 145 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 145 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 145 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 145 |==================== Botan 2.8.0 Test: KASUMI - Decrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 95.01 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 95.26 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 95.17 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 95.30 |================== Botan 2.8.0 Test: KASUMI - Encrypt MiB/s > Higher Is Better -O3 -march=skylake .................................. 96.02 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 96.22 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 96.21 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 96.25 |================== Memcached mcperf 1.5.10 Method: Set Operations Per Second > Higher Is Better -O3 -march=skylake .................................. 72585 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 73045 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 72688 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 72940 |================== Memcached mcperf 1.5.10 Method: Add Operations Per Second > Higher Is Better -O3 -march=skylake .................................. 72123 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 73307 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 72947 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 73346 |================== dav1d 0.3 Video Input: Summer Nature 4K Seconds < Lower Is Better -O3 -march=skylake .................................. 19.15 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 19.47 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 19.34 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 19.67 |================== Memcached mcperf 1.5.10 Method: Append Operations Per Second > Higher Is Better -O3 -march=skylake .................................. 78294 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 78021 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 77703 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 77932 |================== Memcached mcperf 1.5.10 Method: Replace Operations Per Second > Higher Is Better -O3 -march=skylake .................................. 78460 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 78432 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 77373 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 78798 |================== Memcached mcperf 1.5.10 Method: Prepend Operations Per Second > Higher Is Better -O3 -march=skylake .................................. 77484 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 77547 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 79076 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 79234 |================== Rodinia 2.4 Test: OpenMP LavaMD Seconds < Lower Is Better -O3 -march=skylake .................................. 16.96 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 16.99 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 16.95 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 16.98 |================== XZ Compression 5.2.4 Compressing ubuntu-16.04.3-server-i386.img, Compression Level 9 Seconds < Lower Is Better -O3 -march=skylake .................................. 15.88 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 15.84 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 15.80 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 15.99 |================== MKL-DNN 2019-04-16 Harness: IP Batch 1D - Data Type: f32 ms < Lower Is Better -O3 -march=skylake .................................. 20.79 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 20.84 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 20.74 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 20.66 |================== FLAC Audio Encoding 1.3.2 WAV To FLAC Seconds < Lower Is Better -O3 -march=skylake .................................. 8.77 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 8.76 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 8.66 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 9.14 |=================== Xsbench 2017-07-06 Lookups/s > Higher Is Better -O3 -march=skylake .................................. 5164164 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 5171354 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 5140078 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 5166043 |================ CppPerformanceBenchmarks 9 Test: Function Objects Seconds < Lower Is Better -O3 -march=skylake .................................. 13.62 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 13.28 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 13.70 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 13.39 |================== Primesieve 7.4 1e12 Prime Number Generation Seconds < Lower Is Better -O3 -march=skylake .................................. 12.79 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 12.80 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 12.78 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 12.96 |================== Rodinia 2.4 Test: OpenMP CFD Solver Seconds < Lower Is Better -O3 -march=skylake .................................. 11.52 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 11.47 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 11.31 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 11.71 |================== x265 3.0 H.265 1080p Video Encoding Frames Per Second > Higher Is Better -O3 -march=skylake .................................. 53.71 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 53.34 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 53.97 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 50.05 |================= NAS Parallel Benchmarks 3.3.1 Test / Class: FT.B Total Mop/s > Higher Is Better -O3 -march=skylake .................................. 9376 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 9337 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 9331 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 9387 |=================== CLOMP 3.3 Static OMP Speedup Speedup > Higher Is Better -O3 -march=skylake .................................. 2.12 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2.11 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2.12 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2.12 |=================== SVT-AV1 2019-03-07 1080p 8-bit YUV To AV1 Video Encode Frames Per Second > Higher Is Better -O3 -march=skylake .................................. 24.44 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 24.37 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 24.34 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 24.43 |================== Zstd Compression 1.3.4 Compressing ubuntu-16.04.3-server-i386.img, Compression Level 19 Seconds < Lower Is Better -O3 -march=skylake .................................. 9.71 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 9.59 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 9.69 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 9.76 |=================== LAME MP3 Encoding 3.100 WAV To MP3 Seconds < Lower Is Better -O3 -march=skylake .................................. 9.32 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 9.05 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 9.30 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 10.19 |================== Parboil 2.5 Test: OpenMP Stencil Seconds < Lower Is Better -O3 -march=skylake .................................. 7.74 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 7.98 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 7.77 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 7.80 |=================== NAS Parallel Benchmarks 3.3.1 Test / Class: EP.C Total Mop/s > Higher Is Better -O3 -march=skylake .................................. 2038 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2090 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2109 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2113 |=================== dav1d 0.3 Video Input: Summer Nature 1080p Seconds < Lower Is Better -O3 -march=skylake .................................. 6.93 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 6.97 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 6.95 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 7.05 |=================== libjpeg-turbo tjbench 1.5.3 Test: Decompression Throughput Megapixels/sec > Higher Is Better -O3 -march=skylake .................................. 196 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 196 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 194 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 197 |==================== t-test1 2017-01-13 Threads: 2 Seconds < Lower Is Better -O3 -march=skylake .................................. 5.93 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 5.81 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 5.82 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 5.82 |=================== FFmpeg 4.0.2 H.264 HD To NTSC DV Seconds < Lower Is Better -O3 -march=skylake .................................. 5.45 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 5.50 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 5.47 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 5.45 |=================== x264 2018-09-25 H.264 Video Encoding Frames Per Second > Higher Is Better -O3 -march=skylake .................................. 139 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 140 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 140 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 140 |==================== Timed HMMer Search 2.3.2 Pfam Database Search Seconds < Lower Is Better -O3 -march=skylake .................................. 3.49 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 3.40 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 3.50 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 3.88 |=================== NAS Parallel Benchmarks 3.3.1 Test / Class: LU.A Total Mop/s > Higher Is Better -O3 -march=skylake .................................. 37394 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 39791 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 39343 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 37305 |================= SVT-VP9 2019-02-17 1080p 8-bit YUV To VP9 Video Encode Frames Per Second > Higher Is Better -O3 -march=skylake .................................. 342 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 346 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 347 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 337 |=================== PolyBench-C 4.2 Test: 3 Matrix Multiplications Seconds < Lower Is Better -O3 -march=skylake .................................. 2.64 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2.61 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2.63 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2.63 |=================== Parboil 2.5 Test: OpenMP CUTCP Seconds < Lower Is Better -O3 -march=skylake .................................. 2.38 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2.38 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2.38 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2.41 |=================== CloverLeaf Lagrangian-Eulerian Hydrodynamics Seconds < Lower Is Better -O3 -march=skylake .................................. 2.32 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 2.29 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 2.29 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 2.30 |=================== SVT-HEVC 2019-02-03 1080p 8-bit YUV To HEVC Video Encode Frames Per Second > Higher Is Better -O3 -march=skylake .................................. 304 |================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 339 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 337 |==================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 304 |================== TSCP 1.81 AI Chess Performance Nodes Per Second > Higher Is Better -O3 -march=skylake .................................. 1368057 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 1388054 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 1380406 |================ -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 1373514 |================ NAS Parallel Benchmarks 3.3.1 Test / Class: FT.A Total Mop/s > Higher Is Better -O3 -march=skylake .................................. 8581 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 8581 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 8544 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 8543 |=================== BLAKE2 20170307 Cycles Per Byte < Lower Is Better -O3 -march=skylake .................................. 4.02 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 4.02 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 4.02 |=================== -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 4.02 |=================== Geometric Mean Of All Test Results Result Composite - Core i9 7980XE GCC 9 Compiler Benchmarks Geometric Mean > Higher Is Better -O3 -march=skylake .................................. 414.04 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=128 . 415.90 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=256 . 415.02 |================= -O3 -march=skylake-avx512 -mprefer-vector-width=512 . 405.16 |=================