Ryzen 9 5950X AOCC 3.0 Compiler Benchmarking Benchmarks for a future article. GCC 10.2: Processor: AMD Ryzen 9 5950X 16-Core @ 3.40GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR VIII HERO (WI-FI) (3204 BIOS), Chipset: AMD Starship/Matisse, Memory: 32GB, Disk: 2000GB Corsair Force MP600 + 2000GB, Graphics: AMD NAVY_FLOUNDER 12GB (2855/1000MHz), Audio: AMD Device ab28, Monitor: ASUS MG28U, Network: Realtek RTL8125 2.5GbE + Intel I211 + Intel Wi-Fi 6 AX200 OS: Ubuntu 20.10, Kernel: 5.11.6-051106-generic (x86_64), Desktop: GNOME Shell 3.38.2, Display Server: X Server 1.20.9, OpenGL: 4.6 Mesa 21.1.0-devel (git-684f97d 2021-03-12 groovy-oibaf-ppa) (LLVM 11.0.1), Vulkan: 1.2.168, Compiler: GCC 10.2.0, File-System: ext4, Screen Resolution: 3840x2160 LLVM Clang 12: Processor: AMD Ryzen 9 5950X 16-Core @ 3.40GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR VIII HERO (WI-FI) (3204 BIOS), Chipset: AMD Starship/Matisse, Memory: 32GB, Disk: 2000GB Corsair Force MP600 + 2000GB, Graphics: AMD NAVY_FLOUNDER 12GB (2855/1000MHz), Audio: AMD Device ab28, Monitor: ASUS MG28U, Network: Realtek RTL8125 2.5GbE + Intel I211 + Intel Wi-Fi 6 AX200 OS: Ubuntu 20.10, Kernel: 5.11.6-051106-generic (x86_64), Desktop: GNOME Shell 3.38.2, Display Server: X Server 1.20.9, OpenGL: 4.6 Mesa 21.1.0-devel (git-684f97d 2021-03-12 groovy-oibaf-ppa) (LLVM 11.0.1), Vulkan: 1.2.168, Compiler: Clang 12.0.0-++rc3-1~exp1~oibaf~g, File-System: ext4, Screen Resolution: 3840x2160 AMD AOCC 2.3: Processor: AMD Ryzen 9 5950X 16-Core @ 3.40GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR VIII HERO (WI-FI) (3204 BIOS), Chipset: AMD Starship/Matisse, Memory: 32GB, Disk: 2000GB Corsair Force MP600 + 2000GB, Graphics: AMD NAVY_FLOUNDER 12GB (2855/1000MHz), Audio: AMD Device ab28, Monitor: ASUS MG28U, Network: Realtek RTL8125 2.5GbE + Intel I211 + Intel Wi-Fi 6 AX200 OS: Ubuntu 20.10, Kernel: 5.11.6-051106-generic (x86_64), Desktop: GNOME Shell 3.38.2, Display Server: X Server 1.20.9, OpenGL: 4.6 Mesa 21.1.0-devel (git-684f97d 2021-03-12 groovy-oibaf-ppa) (LLVM 11.0.1), Vulkan: 1.2.168, Compiler: Clang 11.0.0, File-System: ext4, Screen Resolution: 3840x2160 AMD AOCC 3.0: Processor: AMD Ryzen 9 5950X 16-Core @ 3.40GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR VIII HERO (WI-FI) (3204 BIOS), Chipset: AMD Starship/Matisse, Memory: 32GB, Disk: 2000GB Corsair Force MP600 + 2000GB, Graphics: AMD NAVY_FLOUNDER 12GB (2855/1000MHz), Audio: AMD Device ab28, Monitor: ASUS MG28U, Network: Realtek RTL8125 2.5GbE + Intel I211 + Intel Wi-Fi 6 AX200 OS: Ubuntu 20.10, Kernel: 5.11.6-051106-generic (x86_64), Desktop: GNOME Shell 3.38.2, Display Server: X Server 1.20.9, OpenGL: 4.6 Mesa 21.1.0-devel (git-684f97d 2021-03-12 groovy-oibaf-ppa) (LLVM 11.0.1), Vulkan: 1.2.168, Compiler: Clang 12.0.0, File-System: ext4, Screen Resolution: 3840x2160 AOM AV1 2.1-rc Encoder Mode: Speed 0 Two-Pass Frames Per Second > Higher Is Better GCC 10.2 ...... 0.37 |================================================== LLVM Clang 12 . 0.42 |========================================================= AOM AV1 2.1-rc Encoder Mode: Speed 4 Two-Pass Frames Per Second > Higher Is Better GCC 10.2 ...... 9.20 |====================================================== LLVM Clang 12 . 9.74 |========================================================= AOM AV1 2.1-rc Encoder Mode: Speed 6 Realtime Frames Per Second > Higher Is Better GCC 10.2 ...... 35.13 |==================================================== LLVM Clang 12 . 37.50 |======================================================== AOM AV1 2.1-rc Encoder Mode: Speed 6 Two-Pass Frames Per Second > Higher Is Better GCC 10.2 ...... 29.43 |======================================================= LLVM Clang 12 . 30.03 |======================================================== AOM AV1 2.1-rc Encoder Mode: Speed 8 Realtime Frames Per Second > Higher Is Better GCC 10.2 ...... 121.13 |======================================================= LLVM Clang 12 . 118.22 |====================================================== ASTC Encoder 2.4 Preset: Medium Seconds < Lower Is Better GCC 10.2 ...... 4.0524 |======================================================= LLVM Clang 12 . 3.5076 |================================================ AMD AOCC 2.3 .. 3.2899 |============================================= AMD AOCC 3.0 .. 3.4040 |============================================== ASTC Encoder 2.4 Preset: Thorough Seconds < Lower Is Better GCC 10.2 ...... 6.9922 |======================================== LLVM Clang 12 . 9.4996 |======================================================= AMD AOCC 2.3 .. 9.2012 |===================================================== AMD AOCC 3.0 .. 9.3493 |====================================================== ASTC Encoder 2.4 Preset: Exhaustive Seconds < Lower Is Better GCC 10.2 ...... 52.93 |======================================================== LLVM Clang 12 . 51.66 |======================================================= AMD AOCC 2.3 .. 50.81 |====================================================== AMD AOCC 3.0 .. 51.45 |====================================================== Basis Universal 1.13 Settings: ETC1S Seconds < Lower Is Better GCC 10.2 ...... 19.90 |==================================================== LLVM Clang 12 . 21.38 |======================================================== AMD AOCC 2.3 .. 21.22 |======================================================== AMD AOCC 3.0 .. 21.37 |======================================================== Basis Universal 1.13 Settings: UASTC Level 0 Seconds < Lower Is Better GCC 10.2 ...... 5.157 |=================================================== LLVM Clang 12 . 5.522 |======================================================= AMD AOCC 2.3 .. 5.453 |====================================================== AMD AOCC 3.0 .. 5.639 |======================================================== Basis Universal 1.13 Settings: UASTC Level 2 Seconds < Lower Is Better GCC 10.2 ...... 15.90 |======================================================= LLVM Clang 12 . 15.91 |======================================================= AMD AOCC 2.3 .. 15.98 |======================================================== AMD AOCC 3.0 .. 16.06 |======================================================== Basis Universal 1.13 Settings: UASTC Level 3 Seconds < Lower Is Better GCC 10.2 ...... 28.13 |======================================================== LLVM Clang 12 . 28.22 |======================================================== AMD AOCC 2.3 .. 28.15 |======================================================== AMD AOCC 3.0 .. 28.17 |======================================================== C-Ray 1.1 Total Time - 4K, 16 Rays Per Pixel Seconds < Lower Is Better GCC 10.2 ...... 25.09 |=============================== LLVM Clang 12 . 44.89 |======================================================== AMD AOCC 2.3 .. 44.53 |======================================================== AMD AOCC 3.0 .. 44.33 |======================================================= Crafty 25.2 Elapsed Time Nodes Per Second > Higher Is Better GCC 10.2 . 11731249 |========================================================== Crypto++ 8.2 Test: Unkeyed Algorithms MiB/second > Higher Is Better GCC 10.2 ...... 545.91 |====================================================== LLVM Clang 12 . 552.38 |======================================================= AMD AOCC 2.3 .. 550.46 |======================================================= AMD AOCC 3.0 .. 538.88 |====================================================== dav1d 0.8.2 Video Input: Summer Nature 4K FPS > Higher Is Better GCC 10.2 ...... 243.69 |======================================================= LLVM Clang 12 . 244.37 |======================================================= AMD AOCC 2.3 .. 244.15 |======================================================= AMD AOCC 3.0 .. 229.03 |==================================================== dav1d 0.8.2 Video Input: Summer Nature 1080p FPS > Higher Is Better GCC 10.2 ...... 971.79 |======================================================= LLVM Clang 12 . 979.34 |======================================================= AMD AOCC 2.3 .. 976.93 |======================================================= AMD AOCC 3.0 .. 959.29 |====================================================== Etcpak 0.7 Configuration: DXT1 Mpx/s > Higher Is Better GCC 10.2 ...... 1546.30 |======================= LLVM Clang 12 . 3669.19 |====================================================== AMD AOCC 2.3 .. 2986.71 |============================================ AMD AOCC 3.0 .. 3583.01 |===================================================== Etcpak 0.7 Configuration: ETC1 Mpx/s > Higher Is Better GCC 10.2 ...... 386.56 |======================================================= LLVM Clang 12 . 383.47 |======================================================= AMD AOCC 2.3 .. 285.29 |========================================= AMD AOCC 3.0 .. 286.93 |========================================= Etcpak 0.7 Configuration: ETC2 Mpx/s > Higher Is Better GCC 10.2 ...... 245.04 |================================================= LLVM Clang 12 . 272.99 |======================================================= AMD AOCC 2.3 .. 236.47 |================================================ AMD AOCC 3.0 .. 242.03 |================================================= Gcrypt Library 1.9 Seconds < Lower Is Better GCC 10.2 ...... 171.19 |====================================================== LLVM Clang 12 . 172.90 |====================================================== AMD AOCC 2.3 .. 173.31 |====================================================== AMD AOCC 3.0 .. 175.69 |======================================================= GNU Radio Test: Five Back to Back FIR Filters MiB/s > Higher Is Better GCC 10.2 ...... 920.8 |======================================================= LLVM Clang 12 . 911.2 |======================================================= AMD AOCC 2.3 .. 931.6 |======================================================== AMD AOCC 3.0 .. 929.1 |======================================================== GNU Radio Test: Signal Source (Cosine) MiB/s > Higher Is Better GCC 10.2 ...... 4715.4 |====================================================== LLVM Clang 12 . 4769.8 |======================================================= AMD AOCC 2.3 .. 4661.6 |====================================================== AMD AOCC 3.0 .. 4704.7 |====================================================== GNU Radio Test: FIR Filter MiB/s > Higher Is Better GCC 10.2 ...... 1063.5 |====================================================== LLVM Clang 12 . 1060.1 |====================================================== AMD AOCC 2.3 .. 1080.3 |======================================================= AMD AOCC 3.0 .. 1065.4 |====================================================== GNU Radio Test: IIR Filter MiB/s > Higher Is Better GCC 10.2 ...... 843.1 |======================================================= LLVM Clang 12 . 835.4 |======================================================= AMD AOCC 2.3 .. 853.3 |======================================================== AMD AOCC 3.0 .. 838.3 |======================================================= GNU Radio Test: FM Deemphasis Filter MiB/s > Higher Is Better GCC 10.2 ...... 1055.0 |======================================================= LLVM Clang 12 . 1054.9 |======================================================= AMD AOCC 2.3 .. 1061.0 |======================================================= AMD AOCC 3.0 .. 1055.8 |======================================================= GNU Radio Test: Hilbert Transform MiB/s > Higher Is Better GCC 10.2 ...... 515.8 |====================================================== LLVM Clang 12 . 522.8 |======================================================= AMD AOCC 2.3 .. 534.8 |======================================================== AMD AOCC 3.0 .. 523.6 |======================================================= Google SynthMark 20201109 Test: VoiceMark_100 Voices > Higher Is Better GCC 10.2 ...... 966.30 |======================================================= LLVM Clang 12 . 795.81 |============================================= AMD AOCC 2.3 .. 807.37 |============================================== AMD AOCC 3.0 .. 789.22 |============================================= GraphicsMagick 1.3.33 Operation: Swirl Iterations Per Minute > Higher Is Better GCC 10.2 ...... 1166 |========================================================= LLVM Clang 12 . 1108 |====================================================== AMD AOCC 2.3 .. 1131 |======================================================= AMD AOCC 3.0 .. 1083 |===================================================== GraphicsMagick 1.3.33 Operation: Rotate Iterations Per Minute > Higher Is Better GCC 10.2 ...... 1056 |========================================================= LLVM Clang 12 . 1016 |======================================================= AMD AOCC 2.3 .. 928 |================================================== AMD AOCC 3.0 .. 867 |=============================================== GraphicsMagick 1.3.33 Operation: Sharpen Iterations Per Minute > Higher Is Better GCC 10.2 ...... 375 |========================================================== LLVM Clang 12 . 237 |===================================== AMD AOCC 2.3 .. 241 |===================================== AMD AOCC 3.0 .. 240 |===================================== GraphicsMagick 1.3.33 Operation: Enhanced Iterations Per Minute > Higher Is Better GCC 10.2 ...... 439 |======================================================= LLVM Clang 12 . 457 |========================================================= AMD AOCC 2.3 .. 461 |========================================================== AMD AOCC 3.0 .. 452 |========================================================= GraphicsMagick 1.3.33 Operation: Resizing Iterations Per Minute > Higher Is Better GCC 10.2 ...... 2165 |========================================================= LLVM Clang 12 . 1789 |=============================================== AMD AOCC 2.3 .. 1824 |================================================ AMD AOCC 3.0 .. 1720 |============================================= GraphicsMagick 1.3.33 Operation: Noise-Gaussian Iterations Per Minute > Higher Is Better GCC 10.2 ...... 454 |========================================================== LLVM Clang 12 . 398 |=================================================== AMD AOCC 2.3 .. 402 |=================================================== AMD AOCC 3.0 .. 392 |================================================== GraphicsMagick 1.3.33 Operation: HWB Color Space Iterations Per Minute > Higher Is Better GCC 10.2 ...... 1115 |========================================================= LLVM Clang 12 . 844 |=========================================== AMD AOCC 2.3 .. 848 |=========================================== AMD AOCC 3.0 .. 805 |========================================= JPEG XL 0.3.3 Input: PNG - Encode Speed: 5 MP/s > Higher Is Better GCC 10.2 ...... 74.12 |======================================================== LLVM Clang 12 . 74.77 |======================================================== AMD AOCC 2.3 .. 74.61 |======================================================== AMD AOCC 3.0 .. 73.64 |======================================================= JPEG XL 0.3.3 Input: PNG - Encode Speed: 7 MP/s > Higher Is Better GCC 10.2 ...... 11.20 |======================================================= LLVM Clang 12 . 11.41 |======================================================== AMD AOCC 2.3 .. 11.31 |======================================================== AMD AOCC 3.0 .. 11.17 |======================================================= JPEG XL 0.3.3 Input: PNG - Encode Speed: 8 MP/s > Higher Is Better GCC 10.2 ...... 1.14 |========================================================= LLVM Clang 12 . 1.06 |===================================================== AMD AOCC 2.3 .. 1.04 |==================================================== AMD AOCC 3.0 .. 1.04 |==================================================== JPEG XL 0.3.3 Input: JPEG - Encode Speed: 5 MP/s > Higher Is Better GCC 10.2 ...... 87.35 |======================================================= LLVM Clang 12 . 85.67 |====================================================== AMD AOCC 2.3 .. 89.51 |======================================================== AMD AOCC 3.0 .. 83.54 |==================================================== JPEG XL 0.3.3 Input: JPEG - Encode Speed: 7 MP/s > Higher Is Better GCC 10.2 ...... 87.07 |======================================================= LLVM Clang 12 . 85.81 |====================================================== AMD AOCC 2.3 .. 89.32 |======================================================== AMD AOCC 3.0 .. 83.63 |==================================================== JPEG XL 0.3.3 Input: JPEG - Encode Speed: 8 MP/s > Higher Is Better GCC 10.2 ...... 38.13 |======================================================== LLVM Clang 12 . 36.44 |====================================================== AMD AOCC 2.3 .. 35.87 |===================================================== AMD AOCC 3.0 .. 34.34 |================================================== JPEG XL Decoding 0.3.3 CPU Threads: 1 MP/s > Higher Is Better GCC 10.2 ...... 56.53 |================================================= LLVM Clang 12 . 62.27 |====================================================== AMD AOCC 2.3 .. 64.34 |======================================================== AMD AOCC 3.0 .. 59.92 |==================================================== JPEG XL Decoding 0.3.3 CPU Threads: All MP/s > Higher Is Better GCC 10.2 ...... 210.99 |====================================================== LLVM Clang 12 . 196.34 |=================================================== AMD AOCC 2.3 .. 213.67 |======================================================= AMD AOCC 3.0 .. 191.91 |================================================= libavif avifenc 0.9.0 Encoder Speed: 0 Seconds < Lower Is Better GCC 10.2 ...... 43.62 |======================================================== LLVM Clang 12 . 41.08 |===================================================== AMD AOCC 2.3 .. 40.75 |==================================================== AMD AOCC 3.0 .. 41.03 |===================================================== libavif avifenc 0.9.0 Encoder Speed: 2 Seconds < Lower Is Better GCC 10.2 ...... 23.54 |======================================================== LLVM Clang 12 . 22.07 |===================================================== AMD AOCC 2.3 .. 21.83 |==================================================== AMD AOCC 3.0 .. 22.01 |==================================================== libavif avifenc 0.9.0 Encoder Speed: 6 Seconds < Lower Is Better GCC 10.2 ...... 8.927 |======================================================== LLVM Clang 12 . 8.342 |==================================================== AMD AOCC 2.3 .. 8.384 |===================================================== AMD AOCC 3.0 .. 8.309 |==================================================== libavif avifenc 0.9.0 Encoder Speed: 10 Seconds < Lower Is Better GCC 10.2 ...... 2.934 |======================================================== LLVM Clang 12 . 2.952 |======================================================== AMD AOCC 2.3 .. 2.933 |======================================================== AMD AOCC 3.0 .. 2.941 |======================================================== libavif avifenc 0.9.0 Encoder Speed: 6, Lossless Seconds < Lower Is Better GCC 10.2 ...... 30.98 |======================================================== LLVM Clang 12 . 27.85 |================================================== AMD AOCC 2.3 .. 27.62 |================================================== AMD AOCC 3.0 .. 27.91 |================================================== libavif avifenc 0.9.0 Encoder Speed: 10, Lossless Seconds < Lower Is Better GCC 10.2 ...... 4.875 |======================================================== LLVM Clang 12 . 4.807 |======================================================= AMD AOCC 2.3 .. 4.832 |======================================================== AMD AOCC 3.0 .. 4.837 |======================================================== LibRaw 0.20 Post-Processing Benchmark Mpix/sec > Higher Is Better GCC 10.2 ...... 78.66 |======================================================== LLVM Clang 12 . 54.14 |======================================= AMD AOCC 2.3 .. 50.37 |==================================== AMD AOCC 3.0 .. 52.68 |====================================== Liquid-DSP 2021.01.31 Threads: 1 - Buffer Length: 256 - Filter Length: 57 samples/s > Higher Is Better GCC 10.2 ...... 81844000 |===================================================== LLVM Clang 12 . 77794333 |================================================== AMD AOCC 2.3 .. 75031667 |================================================= AMD AOCC 3.0 .. 78734000 |=================================================== Liquid-DSP 2021.01.31 Threads: 16 - Buffer Length: 256 - Filter Length: 57 samples/s > Higher Is Better GCC 10.2 ...... 1111200000 |=================================================== LLVM Clang 12 . 1067666667 |================================================= AMD AOCC 2.3 .. 1067266667 |================================================= AMD AOCC 3.0 .. 1086033333 |================================================== Liquid-DSP 2021.01.31 Threads: 32 - Buffer Length: 256 - Filter Length: 57 samples/s > Higher Is Better GCC 10.2 ...... 1164966667 |============================================ LLVM Clang 12 . 1335233333 |=================================================== AMD AOCC 2.3 .. 1332333333 |=================================================== AMD AOCC 3.0 .. 1334900000 |=================================================== LZ4 Compression 1.9.3 Compression Level: 1 - Compression Speed MB/s > Higher Is Better GCC 10.2 ...... 12330.56 |==================================================== LLVM Clang 12 . 12227.47 |==================================================== AMD AOCC 2.3 .. 12456.17 |===================================================== AMD AOCC 3.0 .. 12124.81 |==================================================== LZ4 Compression 1.9.3 Compression Level: 1 - Decompression Speed MB/s > Higher Is Better GCC 10.2 ...... 13771.1 |====================================================== LLVM Clang 12 . 13305.3 |==================================================== AMD AOCC 2.3 .. 13595.3 |===================================================== AMD AOCC 3.0 .. 13144.8 |==================================================== LZ4 Compression 1.9.3 Compression Level: 3 - Compression Speed MB/s > Higher Is Better GCC 10.2 ...... 72.36 |======================================================== LLVM Clang 12 . 68.30 |===================================================== AMD AOCC 2.3 .. 72.21 |======================================================== AMD AOCC 3.0 .. 72.30 |======================================================== LZ4 Compression 1.9.3 Compression Level: 3 - Decompression Speed MB/s > Higher Is Better GCC 10.2 ...... 13400.1 |====================================================== LLVM Clang 12 . 13082.4 |===================================================== AMD AOCC 2.3 .. 13212.6 |===================================================== AMD AOCC 3.0 .. 13010.7 |==================================================== LZ4 Compression 1.9.3 Compression Level: 9 - Compression Speed MB/s > Higher Is Better GCC 10.2 ...... 71.13 |======================================================== LLVM Clang 12 . 64.43 |=================================================== AMD AOCC 2.3 .. 68.80 |====================================================== AMD AOCC 3.0 .. 68.40 |====================================================== LZ4 Compression 1.9.3 Compression Level: 9 - Decompression Speed MB/s > Higher Is Better GCC 10.2 ...... 13397.7 |====================================================== LLVM Clang 12 . 13129.4 |===================================================== AMD AOCC 2.3 .. 13188.4 |===================================================== AMD AOCC 3.0 .. 12981.8 |==================================================== Mobile Neural Network 1.1.3 Model: SqueezeNetV1.0 ms < Lower Is Better GCC 10.2 . 5.081 |============================================================= Mobile Neural Network 1.1.3 Model: resnet-v2-50 ms < Lower Is Better GCC 10.2 . 25.07 |============================================================= Mobile Neural Network 1.1.3 Model: MobileNetV2_224 ms < Lower Is Better GCC 10.2 . 3.240 |============================================================= Mobile Neural Network 1.1.3 Model: mobilenet-v1-1.0 ms < Lower Is Better GCC 10.2 . 2.351 |============================================================= Mobile Neural Network 1.1.3 Model: inception-v3 ms < Lower Is Better GCC 10.2 . 32.34 |============================================================= NCNN 20201218 Target: CPU - Model: mobilenet ms < Lower Is Better GCC 10.2 ...... 12.42 |======================================================== LLVM Clang 12 . 11.53 |==================================================== AMD AOCC 2.3 .. 10.96 |================================================= AMD AOCC 3.0 .. 11.28 |=================================================== NCNN 20201218 Target: CPU-v2-v2 - Model: mobilenet-v2 ms < Lower Is Better GCC 10.2 ...... 4.43 |========================================================= LLVM Clang 12 . 3.79 |================================================= AMD AOCC 2.3 .. 3.52 |============================================= AMD AOCC 3.0 .. 3.53 |============================================= NCNN 20201218 Target: CPU-v3-v3 - Model: mobilenet-v3 ms < Lower Is Better GCC 10.2 ...... 3.85 |========================================================= LLVM Clang 12 . 3.33 |================================================= AMD AOCC 2.3 .. 3.06 |============================================= AMD AOCC 3.0 .. 3.07 |============================================= NCNN 20201218 Target: CPU - Model: shufflenet-v2 ms < Lower Is Better GCC 10.2 ...... 4.23 |========================================================= LLVM Clang 12 . 4.04 |====================================================== AMD AOCC 2.3 .. 3.79 |=================================================== AMD AOCC 3.0 .. 3.87 |==================================================== NCNN 20201218 Target: CPU - Model: mnasnet ms < Lower Is Better GCC 10.2 ...... 3.93 |========================================================= LLVM Clang 12 . 3.45 |================================================== AMD AOCC 2.3 .. 3.16 |============================================== AMD AOCC 3.0 .. 3.18 |============================================== NCNN 20201218 Target: CPU - Model: efficientnet-b0 ms < Lower Is Better GCC 10.2 ...... 5.32 |========================================================= LLVM Clang 12 . 4.80 |=================================================== AMD AOCC 2.3 .. 4.53 |================================================= AMD AOCC 3.0 .. 4.50 |================================================ NCNN 20201218 Target: CPU - Model: blazeface ms < Lower Is Better GCC 10.2 ...... 1.83 |========================================================= LLVM Clang 12 . 1.73 |====================================================== AMD AOCC 2.3 .. 1.57 |================================================= AMD AOCC 3.0 .. 1.56 |================================================= NCNN 20201218 Target: CPU - Model: googlenet ms < Lower Is Better GCC 10.2 ...... 12.76 |======================================================== LLVM Clang 12 . 12.53 |======================================================= AMD AOCC 2.3 .. 11.94 |==================================================== AMD AOCC 3.0 .. 11.96 |==================================================== NCNN 20201218 Target: CPU - Model: vgg16 ms < Lower Is Better GCC 10.2 ...... 57.89 |======================================================= LLVM Clang 12 . 57.51 |======================================================= AMD AOCC 2.3 .. 58.11 |======================================================= AMD AOCC 3.0 .. 58.96 |======================================================== NCNN 20201218 Target: CPU - Model: resnet18 ms < Lower Is Better GCC 10.2 ...... 14.11 |======================================================== LLVM Clang 12 . 13.91 |======================================================= AMD AOCC 2.3 .. 14.08 |======================================================== AMD AOCC 3.0 .. 13.86 |======================================================= NCNN 20201218 Target: CPU - Model: alexnet ms < Lower Is Better GCC 10.2 ...... 10.82 |====================================================== LLVM Clang 12 . 11.14 |======================================================== AMD AOCC 2.3 .. 11.01 |======================================================= AMD AOCC 3.0 .. 11.11 |======================================================== NCNN 20201218 Target: CPU - Model: resnet50 ms < Lower Is Better GCC 10.2 ...... 25.67 |======================================================== LLVM Clang 12 . 23.54 |=================================================== AMD AOCC 2.3 .. 23.31 |=================================================== AMD AOCC 3.0 .. 23.29 |=================================================== NCNN 20201218 Target: CPU - Model: yolov4-tiny ms < Lower Is Better GCC 10.2 ...... 20.77 |===================================================== LLVM Clang 12 . 21.70 |======================================================= AMD AOCC 2.3 .. 21.66 |======================================================= AMD AOCC 3.0 .. 21.93 |======================================================== NCNN 20201218 Target: CPU - Model: squeezenet_ssd ms < Lower Is Better GCC 10.2 ...... 13.77 |======================================================== LLVM Clang 12 . 12.50 |=================================================== AMD AOCC 2.3 .. 12.74 |==================================================== AMD AOCC 3.0 .. 12.40 |================================================== NCNN 20201218 Target: CPU - Model: regnety_400m ms < Lower Is Better GCC 10.2 ...... 17.61 |======================================================== LLVM Clang 12 . 17.06 |====================================================== AMD AOCC 2.3 .. 12.19 |======================================= AMD AOCC 3.0 .. 12.30 |======================================= Ngspice 34 Circuit: C2670 Seconds < Lower Is Better GCC 10.2 ...... 71.60 |======================================================= LLVM Clang 12 . 72.46 |======================================================= AMD AOCC 2.3 .. 72.78 |======================================================== AMD AOCC 3.0 .. 73.28 |======================================================== Ngspice 34 Circuit: C7552 Seconds < Lower Is Better GCC 10.2 ...... 62.82 |====================================================== LLVM Clang 12 . 64.52 |======================================================== AMD AOCC 2.3 .. 64.89 |======================================================== AMD AOCC 3.0 .. 64.91 |======================================================== Ogg Audio Encoding 1.3.4 WAV To Ogg Seconds < Lower Is Better GCC 10.2 ...... 13.58 |============================================== LLVM Clang 12 . 13.37 |============================================= AMD AOCC 2.3 .. 16.56 |======================================================== AMD AOCC 3.0 .. 16.54 |======================================================== oneDNN 2.1.2 Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 3.95979 |==================================================== LLVM Clang 12 . 4.11930 |====================================================== AMD AOCC 3.0 .. 4.09663 |====================================================== oneDNN 2.1.2 Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 9.25967 |==================================================== LLVM Clang 12 . 9.59442 |====================================================== AMD AOCC 3.0 .. 9.57194 |====================================================== oneDNN 2.1.2 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 17.29 |======================================================== LLVM Clang 12 . 17.34 |======================================================== AMD AOCC 3.0 .. 17.36 |======================================================== oneDNN 2.1.2 Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 4.46777 |====================================================== LLVM Clang 12 . 2.46561 |============================== AMD AOCC 3.0 .. 2.46850 |============================== oneDNN 2.1.2 Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 3.55467 |===================================================== LLVM Clang 12 . 3.64485 |====================================================== AMD AOCC 3.0 .. 3.58364 |===================================================== oneDNN 2.1.2 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 2757.52 |====================================================== LLVM Clang 12 . 2757.75 |====================================================== AMD AOCC 3.0 .. 2760.94 |====================================================== oneDNN 2.1.2 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 1773.67 |===================================================== LLVM Clang 12 . 1792.27 |====================================================== AMD AOCC 3.0 .. 1760.57 |===================================================== oneDNN 2.1.2 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU ms < Lower Is Better GCC 10.2 ...... 0.638664 |===================================================== LLVM Clang 12 . 0.641231 |===================================================== AMD AOCC 3.0 .. 0.643093 |===================================================== ONNX Runtime 1.6 Model: yolov4 - Device: OpenMP CPU Inferences Per Minute > Higher Is Better GCC 10.2 ...... 433 |====================================================== LLVM Clang 12 . 426 |===================================================== AMD AOCC 2.3 .. 456 |========================================================= AMD AOCC 3.0 .. 465 |========================================================== ONNX Runtime 1.6 Model: bertsquad-10 - Device: OpenMP CPU Inferences Per Minute > Higher Is Better GCC 10.2 ...... 614 |======================================================= LLVM Clang 12 . 634 |========================================================= AMD AOCC 2.3 .. 646 |========================================================== AMD AOCC 3.0 .. 649 |========================================================== ONNX Runtime 1.6 Model: fcn-resnet101-11 - Device: OpenMP CPU Inferences Per Minute > Higher Is Better GCC 10.2 ...... 99 |======================================================= LLVM Clang 12 . 104 |========================================================== AMD AOCC 2.3 .. 103 |========================================================= AMD AOCC 3.0 .. 102 |========================================================= ONNX Runtime 1.6 Model: shufflenet-v2-10 - Device: OpenMP CPU Inferences Per Minute > Higher Is Better GCC 10.2 ...... 15049 |================================================= LLVM Clang 12 . 14972 |================================================= AMD AOCC 2.3 .. 17105 |======================================================== AMD AOCC 3.0 .. 15474 |=================================================== ONNX Runtime 1.6 Model: super-resolution-10 - Device: OpenMP CPU Inferences Per Minute > Higher Is Better GCC 10.2 ...... 6721 |========================================================= LLVM Clang 12 . 5937 |================================================== AMD AOCC 2.3 .. 6067 |=================================================== AMD AOCC 3.0 .. 5976 |=================================================== OpenFOAM 8 Input: Motorbike 30M Seconds < Lower Is Better GCC 10.2 ...... 97.75 |====================================================== LLVM Clang 12 . 100.16 |======================================================= Opus Codec Encoding 1.3.1 WAV To Opus Encode Seconds < Lower Is Better GCC 10.2 ...... 5.484 |======================================================= LLVM Clang 12 . 5.589 |======================================================== POV-Ray 3.7.0.7 Trace Time Seconds < Lower Is Better GCC 10.2 ...... 24.09 |======================================================== LLVM Clang 12 . 22.12 |=================================================== AMD AOCC 2.3 .. 22.54 |==================================================== AMD AOCC 3.0 .. 22.54 |==================================================== QuantLib 1.21 MFLOPS > Higher Is Better GCC 10.2 ...... 3196.9 |=============================================== LLVM Clang 12 . 3538.5 |==================================================== AMD AOCC 2.3 .. 3710.4 |======================================================= AMD AOCC 3.0 .. 3646.4 |====================================================== Redis 6.0.9 Test: LPOP Requests Per Second > Higher Is Better GCC 10.2 ...... 3549910.50 |================================================ LLVM Clang 12 . 3649832.58 |================================================= AMD AOCC 2.3 .. 3589202.59 |================================================= AMD AOCC 3.0 .. 3766645.92 |=================================================== Redis 6.0.9 Test: SADD Requests Per Second > Higher Is Better GCC 10.2 ...... 3041527.37 |=================================================== LLVM Clang 12 . 2954866.80 |================================================== AMD AOCC 2.3 .. 2961165.67 |================================================== AMD AOCC 3.0 .. 2948093.68 |================================================= Redis 6.0.9 Test: LPUSH Requests Per Second > Higher Is Better GCC 10.2 ...... 2222217.52 |================================================ LLVM Clang 12 . 2212779.00 |================================================ AMD AOCC 2.3 .. 2351340.56 |=================================================== AMD AOCC 3.0 .. 2345671.03 |=================================================== Redis 6.0.9 Test: GET Requests Per Second > Higher Is Better GCC 10.2 ...... 3470419.90 |================================================ LLVM Clang 12 . 3624414.37 |=================================================== AMD AOCC 2.3 .. 3658044.77 |=================================================== AMD AOCC 3.0 .. 3545388.92 |================================================= Redis 6.0.9 Test: SET Requests Per Second > Higher Is Better GCC 10.2 ...... 2640316.17 |================================================= LLVM Clang 12 . 2762047.50 |=================================================== AMD AOCC 2.3 .. 2719539.83 |================================================== AMD AOCC 3.0 .. 2719036.20 |================================================== RNNoise 2020-06-28 Seconds < Lower Is Better GCC 10.2 ...... 14.20 |======================================================= LLVM Clang 12 . 14.33 |======================================================= AMD AOCC 2.3 .. 14.04 |====================================================== AMD AOCC 3.0 .. 14.47 |======================================================== simdjson 0.8.2 Throughput Test: Kostya GB/s > Higher Is Better GCC 10.2 ...... 3.72 |========================================================= LLVM Clang 12 . 3.71 |========================================================= AMD AOCC 2.3 .. 3.53 |====================================================== AMD AOCC 3.0 .. 3.61 |======================================================= simdjson 0.8.2 Throughput Test: LargeRandom GB/s > Higher Is Better GCC 10.2 ...... 1.22 |========================================================= LLVM Clang 12 . 1.14 |===================================================== AMD AOCC 2.3 .. 1.11 |==================================================== AMD AOCC 3.0 .. 1.12 |==================================================== simdjson 0.8.2 Throughput Test: PartialTweets GB/s > Higher Is Better GCC 10.2 ...... 5.64 |==================================================== LLVM Clang 12 . 6.18 |========================================================= AMD AOCC 2.3 .. 5.92 |======================================================= AMD AOCC 3.0 .. 6.04 |======================================================== simdjson 0.8.2 Throughput Test: DistinctUserID GB/s > Higher Is Better GCC 10.2 ...... 5.73 |==================================================== LLVM Clang 12 . 6.26 |========================================================= AMD AOCC 2.3 .. 6.12 |======================================================== AMD AOCC 3.0 .. 6.23 |========================================================= Smallpt 1.0 Global Illumination Renderer; 128 Samples Seconds < Lower Is Better GCC 10.2 . 4.674 |============================================================= SQLite Speedtest 3.30 Timed Time - Size 1,000 Seconds < Lower Is Better GCC 10.2 ...... 42.60 |====================================================== LLVM Clang 12 . 43.26 |======================================================= AMD AOCC 2.3 .. 44.11 |======================================================== SVT-AV1 0.8 Encoder Mode: Enc Mode 4 - Input: 1080p Frames Per Second > Higher Is Better GCC 10.2 ...... 6.137 |================================================== LLVM Clang 12 . 6.823 |======================================================= AMD AOCC 2.3 .. 6.859 |======================================================== AMD AOCC 3.0 .. 6.917 |======================================================== SVT-AV1 0.8 Encoder Mode: Enc Mode 8 - Input: 1080p Frames Per Second > Higher Is Better GCC 10.2 ...... 51.77 |============================================ LLVM Clang 12 . 65.31 |======================================================== AMD AOCC 2.3 .. 65.75 |======================================================== AMD AOCC 3.0 .. 64.28 |======================================================= SVT-VP9 0.1 Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p Frames Per Second > Higher Is Better GCC 10.2 ...... 235.04 |====================================================== LLVM Clang 12 . 223.50 |==================================================== AMD AOCC 2.3 .. 238.11 |======================================================= AMD AOCC 3.0 .. 225.17 |==================================================== SVT-VP9 0.1 Tuning: Visual Quality Optimized - Input: Bosphorus 1080p Frames Per Second > Higher Is Better GCC 10.2 ...... 228.96 |======================================================= LLVM Clang 12 . 219.12 |==================================================== AMD AOCC 2.3 .. 230.19 |======================================================= AMD AOCC 3.0 .. 221.59 |===================================================== Sysbench 1.0.20 Test: CPU Events Per Second > Higher Is Better GCC 10.2 ...... 91743.72 | LLVM Clang 12 . 2445437.63 |= AMD AOCC 2.3 .. 210804861.92 |================================================= AMD AOCC 3.0 .. 210533984.51 |================================================= Tachyon 0.99b6 Total Time Seconds < Lower Is Better GCC 10.2 ...... 44.39 |====================================================== LLVM Clang 12 . 45.07 |======================================================= AMD AOCC 2.3 .. 46.13 |======================================================== AMD AOCC 3.0 .. 45.00 |======================================================= Timed Godot Game Engine Compilation 3.2.3 Time To Compile Seconds < Lower Is Better GCC 10.2 ...... 79.52 |======================================================== LLVM Clang 12 . 80.21 |======================================================== AMD AOCC 2.3 .. 79.94 |======================================================== AMD AOCC 3.0 .. 79.87 |======================================================== Timed LLVM Compilation 10.0 Time To Compile Seconds < Lower Is Better GCC 10.2 ...... 370.57 |================================= LLVM Clang 12 . 302.70 |=========================== AMD AOCC 2.3 .. 576.19 |==================================================== AMD AOCC 3.0 .. 610.96 |======================================================= Timed MrBayes Analysis 3.2.7 Primate Phylogeny Analysis Seconds < Lower Is Better GCC 10.2 ...... 59.87 |======================================================== LLVM Clang 12 . 59.23 |======================================================= AMD AOCC 2.3 .. 59.07 |======================================================= AMD AOCC 3.0 .. 57.99 |====================================================== TNN 0.2.3 Target: CPU - Model: MobileNet v2 ms < Lower Is Better GCC 10.2 ...... 216.28 |============================================ LLVM Clang 12 . 270.79 |======================================================= AMD AOCC 2.3 .. 252.45 |=================================================== AMD AOCC 3.0 .. 260.66 |===================================================== TNN 0.2.3 Target: CPU - Model: SqueezeNet v1.1 ms < Lower Is Better GCC 10.2 ...... 211.57 |======================================================= LLVM Clang 12 . 206.36 |====================================================== AMD AOCC 2.3 .. 203.64 |===================================================== AMD AOCC 3.0 .. 204.60 |===================================================== TSCP 1.81 AI Chess Performance Nodes Per Second > Higher Is Better GCC 10.2 ...... 1965773 |============================================== LLVM Clang 12 . 2148154 |================================================== AMD AOCC 2.3 .. 2314225 |====================================================== AMD AOCC 3.0 .. 2283512 |===================================================== WavPack Audio Encoding 5.3 WAV To WavPack Seconds < Lower Is Better GCC 10.2 ...... 10.15 |======================================================= LLVM Clang 12 . 10.31 |======================================================== AMD AOCC 2.3 .. 10.29 |======================================================== AMD AOCC 3.0 .. 10.34 |======================================================== WebP Image Encode 1.1 Encode Settings: Default Encode Time - Seconds < Lower Is Better GCC 10.2 ...... 1.042 |======================================================== LLVM Clang 12 . 0.977 |===================================================== AMD AOCC 2.3 .. 0.979 |===================================================== AMD AOCC 3.0 .. 1.007 |====================================================== WebP Image Encode 1.1 Encode Settings: Quality 100 Encode Time - Seconds < Lower Is Better GCC 10.2 ...... 1.652 |======================================================= LLVM Clang 12 . 1.638 |======================================================= AMD AOCC 2.3 .. 1.673 |======================================================== AMD AOCC 3.0 .. 1.669 |======================================================== WebP Image Encode 1.1 Encode Settings: Quality 100, Lossless Encode Time - Seconds < Lower Is Better GCC 10.2 ...... 13.99 |======================================================== LLVM Clang 12 . 13.91 |======================================================== AMD AOCC 2.3 .. 13.64 |======================================================= AMD AOCC 3.0 .. 13.84 |======================================================= WebP Image Encode 1.1 Encode Settings: Quality 100, Highest Compression Encode Time - Seconds < Lower Is Better GCC 10.2 ...... 5.242 |======================================================== LLVM Clang 12 . 4.674 |================================================== AMD AOCC 2.3 .. 4.609 |================================================= AMD AOCC 3.0 .. 4.937 |===================================================== WebP Image Encode 1.1 Encode Settings: Quality 100, Lossless, Highest Compression Encode Time - Seconds < Lower Is Better GCC 10.2 ...... 28.81 |======================================================== LLVM Clang 12 . 28.67 |======================================================== AMD AOCC 2.3 .. 28.43 |======================================================= AMD AOCC 3.0 .. 28.19 |======================================================= WebP2 Image Encode 20210126 Encode Settings: Default Seconds < Lower Is Better GCC 10.2 ...... 2.274 |======================================================== LLVM Clang 12 . 2.134 |===================================================== AMD AOCC 2.3 .. 2.144 |===================================================== AMD AOCC 3.0 .. 2.165 |===================================================== WebP2 Image Encode 20210126 Encode Settings: Quality 75, Compression Effort 7 Seconds < Lower Is Better GCC 10.2 ...... 111.80 |======================================================= LLVM Clang 12 . 103.01 |=================================================== AMD AOCC 2.3 .. 106.38 |==================================================== AMD AOCC 3.0 .. 105.72 |==================================================== WebP2 Image Encode 20210126 Encode Settings: Quality 95, Compression Effort 7 Seconds < Lower Is Better GCC 10.2 ...... 203.81 |======================================================= LLVM Clang 12 . 188.31 |=================================================== AMD AOCC 2.3 .. 193.60 |==================================================== AMD AOCC 3.0 .. 193.43 |==================================================== WebP2 Image Encode 20210126 Encode Settings: Quality 100, Compression Effort 5 Seconds < Lower Is Better GCC 10.2 ...... 6.414 |===================================================== LLVM Clang 12 . 6.789 |======================================================== AMD AOCC 2.3 .. 6.288 |==================================================== AMD AOCC 3.0 .. 6.364 |==================================================== WebP2 Image Encode 20210126 Encode Settings: Quality 100, Lossless Compression Seconds < Lower Is Better GCC 10.2 ...... 367.37 |======================================================= LLVM Clang 12 . 349.02 |==================================================== AMD AOCC 2.3 .. 357.90 |====================================================== AMD AOCC 3.0 .. 356.82 |===================================================== x264 2019-12-17 H.264 Video Encoding Frames Per Second > Higher Is Better GCC 10.2 ...... 208.93 |====================================================== LLVM Clang 12 . 213.57 |======================================================= AMD AOCC 2.3 .. 210.72 |====================================================== AMD AOCC 3.0 .. 210.35 |====================================================== x265 3.4 Video Input: Bosphorus 4K Frames Per Second > Higher Is Better GCC 10.2 ...... 27.83 |======================================================== LLVM Clang 12 . 28.02 |======================================================== AMD AOCC 2.3 .. 27.49 |======================================================= AMD AOCC 3.0 .. 26.96 |====================================================== x265 3.4 Video Input: Bosphorus 1080p Frames Per Second > Higher Is Better GCC 10.2 ...... 89.80 |======================================================= LLVM Clang 12 . 92.16 |======================================================== AMD AOCC 2.3 .. 89.74 |======================================================= AMD AOCC 3.0 .. 88.70 |====================================================== Zstd Compression 1.4.9 Compression Level: 8 - Compression Speed MB/s > Higher Is Better LLVM Clang 12 . 1078.1 |===================================================== GCC 10.2 ...... 1057.4 |==================================================== LLVM Clang 12 . 1043.2 |=================================================== AMD AOCC 2.3 .. 1117.6 |======================================================= AMD AOCC 3.0 .. 1096.3 |====================================================== Zstd Compression 1.4.9 Compression Level: 8 - Decompression Speed MB/s > Higher Is Better GCC 10.2 ...... 4617.1 |======================================================= LLVM Clang 12 . 4352.4 |==================================================== AMD AOCC 2.3 .. 4468.2 |===================================================== AMD AOCC 3.0 .. 4463.1 |===================================================== Zstd Compression 1.4.9 Compression Level: 19 - Compression Speed MB/s > Higher Is Better GCC 10.2 ...... 51.6 |========================================================= LLVM Clang 12 . 51.3 |========================================================= AMD AOCC 2.3 .. 50.7 |======================================================== AMD AOCC 3.0 .. 50.5 |======================================================== Zstd Compression 1.4.9 Compression Level: 19 - Decompression Speed MB/s > Higher Is Better GCC 10.2 ...... 4251.7 |======================================================= LLVM Clang 12 . 4000.9 |==================================================== AMD AOCC 2.3 .. 4097.3 |===================================================== AMD AOCC 3.0 .. 3608.8 |=============================================== Zstd Compression 1.4.9 Compression Level: 3, Long Mode - Compression Speed MB/s > Higher Is Better GCC 10.2 ...... 1425.9 |======================================================= LLVM Clang 12 . 1191.6 |============================================== AMD AOCC 2.3 .. 1166.4 |============================================= AMD AOCC 3.0 .. 1186.0 |============================================== Zstd Compression 1.4.9 Compression Level: 3, Long Mode - Decompression Speed MB/s > Higher Is Better GCC 10.2 ...... 4737.1 |======================================================= LLVM Clang 12 . 4456.6 |==================================================== AMD AOCC 2.3 .. 4586.2 |===================================================== AMD AOCC 3.0 .. 4543.8 |===================================================== Zstd Compression 1.4.9 Compression Level: 8, Long Mode - Compression Speed MB/s > Higher Is Better AMD AOCC 2.3 .. 1025.4 |================================================== LLVM Clang 12 . 1034.5 |=================================================== AMD AOCC 3.0 .. 1024.5 |================================================== GCC 10.2 ...... 1122.6 |======================================================= LLVM Clang 12 . 1039.1 |=================================================== AMD AOCC 2.3 .. 1026.1 |================================================== AMD AOCC 3.0 .. 1023.5 |================================================== Zstd Compression 1.4.9 Compression Level: 8, Long Mode - Decompression Speed MB/s > Higher Is Better GCC 10.2 ..... 4886.2 |======================================================== AMD AOCC 2.3 . 4805.6 |======================================================= Zstd Compression 1.4.9 Compression Level: 19, Long Mode - Compression Speed MB/s > Higher Is Better GCC 10.2 ...... 36.6 |========================================================= LLVM Clang 12 . 36.8 |========================================================= AMD AOCC 2.3 .. 36.7 |========================================================= AMD AOCC 3.0 .. 36.4 |======================================================== Zstd Compression 1.4.9 Compression Level: 19, Long Mode - Decompression Speed MB/s > Higher Is Better GCC 10.2 ...... 4350.9 |======================================================= LLVM Clang 12 . 3957.9 |================================================== AMD AOCC 2.3 .. 4024.7 |=================================================== AMD AOCC 3.0 .. 3978.2 |==================================================