AMD A10-6800K Compiler GCC vs. LLVM/Clang Performance AMD A10-6800K Richland APU testing of GCC 4.8.1 and LLVM Clang 3.3 Benchmarking by Michael Larabel for a future article on Phoronix.com. GCC 4.8.1: Processor: AMD A10-6800K APU @ 4.70GHz (4 Cores), Motherboard: MSI FM2-A85XA-G65 (MS-7793) v1.0, Chipset: AMD Family 15h, Memory: 7168MB, Disk: 64GB OCZ AGILITY, Graphics: AMD Radeon HD 8670D 768MB (844/1066MHz), Audio: ATI R6xx HDMI, Monitor: SyncMaster, Network: Realtek RTL8111/8168 OS: Ubuntu 13.10, Kernel: 3.9.0-6-generic (x86_64), Desktop: Xfce 4.10, Display Server: X Server 1.13.3, Display Driver: fglrx 13.10.10, OpenGL: 4.2.12337, Compiler: GCC 4.8, File-System: ext4, Screen Resolution: 2560x1600 LLVM Clang 3.3: Processor: AMD A10-6800K APU @ 4.70GHz (4 Cores), Motherboard: MSI FM2-A85XA-G65 (MS-7793) v1.0, Chipset: AMD Family 15h, Memory: 7168MB, Disk: 64GB OCZ AGILITY, Graphics: AMD Radeon HD 8670D 768MB (844/1066MHz), Audio: ATI R6xx HDMI, Monitor: SyncMaster, Network: Realtek RTL8111/8168 OS: Ubuntu 13.10, Kernel: 3.9.0-6-generic (x86_64), Desktop: Xfce 4.10, Display Server: X Server 1.13.3, Display Driver: fglrx 13.10.10, OpenGL: 4.2.12337, Compiler: Clang 3.3-2, File-System: ext4, Screen Resolution: 2560x1600 1306306-SO-RICHLANDG80-vs-8120: Processor: AMD FX-8120 Eight-Core @ 3.10GHz (8 Cores), Motherboard: ASRock 990FX Extreme3, Chipset: AMD ATI RD890 bridge, Memory: 8192MB, Disk: 20GB INTEL SSDSA2VP02 + 2 x 1000GB Western Digital WD10EADS-00L + 300GB Maxtor 6V300F0, Graphics: NVIDIA GeForce GTX 560 Ti 1024MB (835/2050MHz), Audio: NVIDIA GPU 16 HDMI/DP, Network: Broadcom NetLink BCM57781 Gigabit PCIe OS: Ubuntu 12.04, Kernel: 3.2.0-49-generic (x86_64), Desktop: Openbox 3.5.0, Display Server: X Server 1.11.3, Display Driver: NVIDIA 310.14, OpenGL: 4.3.0 NVIDIA 310.14, Compiler: GCC 4.6, File-System: ext4, Screen Resolution: 1920x1080 Apache Benchmark 2.4.3 Static Web Page Serving Requests Per Second > Higher Is Better GCC 4.8.1 ...................... 21944.40 |=================================== LLVM Clang 3.3 ................. 22273.00 |==================================== 1306306-SO-RICHLANDG80-vs-8120 . 11268.30 |================== BLAKE2 20121223 Phoronix Test Suite v4.8.0m1 Cycles Per Byte < Lower Is Better GCC 4.8.1 ...................... 8.99 |===================================== LLVM Clang 3.3 ................. 9.68 |======================================== 1306306-SO-RICHLANDG80-vs-8120 . 7.43 |=============================== Botan 1.10.3 Test: Tiger Mbytes/s > Higher Is Better GCC 4.8.1 ...................... 342.44 |==================================== LLVM Clang 3.3 ................. 359.42 |====================================== 1306306-SO-RICHLANDG80-vs-8120 . 231.95 |========================= Botan 1.10.3 Test: KASUMI Mbytes/s > Higher Is Better GCC 4.8.1 ...................... 77.91 |======================================= LLVM Clang 3.3 ................. 78.76 |======================================= 1306306-SO-RICHLANDG80-vs-8120 . 50.24 |========================= Botan 1.10.3 Test: AES-256 Mbytes/s > Higher Is Better GCC 4.8.1 ...................... 134.10 |============================ LLVM Clang 3.3 ................. 180.25 |====================================== 1306306-SO-RICHLANDG80-vs-8120 . 92.13 |=================== Botan 1.10.3 Test: Twofish Mbytes/s > Higher Is Better GCC 4.8.1 ...................... 219.46 |====================================== LLVM Clang 3.3 ................. 216.68 |====================================== 1306306-SO-RICHLANDG80-vs-8120 . 148.28 |========================== Botan 1.10.3 Test: CAST-256 Mbytes/s > Higher Is Better GCC 4.8.1 ...................... 88.88 |============================= LLVM Clang 3.3 ................. 115.20 |====================================== 1306306-SO-RICHLANDG80-vs-8120 . 69.08 |======================= Botan 1.10.3 Test: X9.19-MAC Mbytes/s > Higher Is Better GCC 4.8.1 ...................... 82.94 |======================================= LLVM Clang 3.3 ................. 76.16 |==================================== 1306306-SO-RICHLANDG80-vs-8120 . 53.71 |========================= C-Ray 1.1 Total Time Seconds < Lower Is Better GCC 4.8.1 ...................... 33.10 |==================== LLVM Clang 3.3 ................. 58.32 |==================================== 1306306-SO-RICHLANDG80-vs-8120 . 63.68 |======================================= FFmpeg 1.1 H.264 HD To NTSC DV Seconds < Lower Is Better GCC 4.8.1 ...................... 15.57 |===================== LLVM Clang 3.3 ................. 15.47 |===================== 1306306-SO-RICHLANDG80-vs-8120 . 29.38 |======================================= FLAC Audio Encoding 1.3.0 WAV To FLAC Seconds < Lower Is Better GCC 4.8.1 ...................... 6.81 |============================ LLVM Clang 3.3 ................. 7.11 |============================== 1306306-SO-RICHLANDG80-vs-8120 . 9.56 |======================================== GraphicsMagick 1.3.16 Operation: Blur Iterations Per Minute > Higher Is Better GCC 4.8.1 ...................... 116 |========================================= LLVM Clang 3.3 ................. 73 |========================== 1306306-SO-RICHLANDG80-vs-8120 . 77 |=========================== GraphicsMagick 1.3.16 Operation: Sharpen Iterations Per Minute > Higher Is Better GCC 4.8.1 ...................... 92 |========================================== LLVM Clang 3.3 ................. 26 |============ 1306306-SO-RICHLANDG80-vs-8120 . 47 |===================== GraphicsMagick 1.3.16 Operation: Resizing Iterations Per Minute > Higher Is Better GCC 4.8.1 ...................... 152 |========================================= LLVM Clang 3.3 ................. 86 |======================= 1306306-SO-RICHLANDG80-vs-8120 . 98 |========================== GraphicsMagick 1.3.16 Operation: HWB Color Space Iterations Per Minute > Higher Is Better GCC 4.8.1 ...................... 159 |========================================= LLVM Clang 3.3 ................. 121 |=============================== 1306306-SO-RICHLANDG80-vs-8120 . 123 |================================ GraphicsMagick 1.3.16 Operation: Local Adaptive Thresholding Iterations Per Minute > Higher Is Better GCC 4.8.1 ...................... 83 |========================================== LLVM Clang 3.3 ................. 42 |===================== 1306306-SO-RICHLANDG80-vs-8120 . 70 |=================================== Himeno Benchmark 3.0 Poisson Pressure Solver MFLOPS > Higher Is Better GCC 4.8.1 ...................... 835.46 |====================================== LLVM Clang 3.3 ................. 779.11 |=================================== 1306306-SO-RICHLANDG80-vs-8120 . 519.34 |======================== LAME MP3 Encoding 3.99.3 WAV To MP3 Seconds < Lower Is Better GCC 4.8.1 ...................... 16.81 |=========================== LLVM Clang 3.3 ................. 16.46 |========================== 1306306-SO-RICHLANDG80-vs-8120 . 24.56 |======================================= Primesieve 4.2 1e12 Prime Number Generation Seconds < Lower Is Better GCC 4.8.1 ...................... 298.53 |================ LLVM Clang 3.3 ................. 696.12 |====================================== 1306306-SO-RICHLANDG80-vs-8120 . 245.09 |============= SciMark 2.0 Computational Test: Monte Carlo Mflops > Higher Is Better GCC 4.8.1 ...................... 505.89 |====================================== LLVM Clang 3.3 ................. 490.89 |===================================== 1306306-SO-RICHLANDG80-vs-8120 . 316.98 |======================== SciMark 2.0 Computational Test: Fast Fourier Transform Mflops > Higher Is Better GCC 4.8.1 ...................... 67.22 |================================== LLVM Clang 3.3 ................. 77.30 |======================================= 1306306-SO-RICHLANDG80-vs-8120 . 65.70 |================================= SciMark 2.0 Computational Test: Sparse Matrix Multiply Mflops > Higher Is Better GCC 4.8.1 ...................... 1008.88 |==================================== LLVM Clang 3.3 ................. 1049.75 |===================================== 1306306-SO-RICHLANDG80-vs-8120 . 820.03 |============================= SciMark 2.0 Computational Test: Dense LU Matrix Factorization Mflops > Higher Is Better GCC 4.8.1 ...................... 1221.85 |=============================== LLVM Clang 3.3 ................. 1236.14 |================================ 1306306-SO-RICHLANDG80-vs-8120 . 1450.28 |===================================== SciMark 2.0 Computational Test: Jacobi Successive Over-Relaxation Mflops > Higher Is Better GCC 4.8.1 ...................... 859.39 |======================= LLVM Clang 3.3 ................. 1358.12 |===================================== 1306306-SO-RICHLANDG80-vs-8120 . 618.79 |================= Smallpt 1.0 Global Illumination Renderer; 100 Samples Seconds < Lower Is Better GCC 4.8.1 ...................... 57 |=========== LLVM Clang 3.3 ................. 212 |========================================= 1306306-SO-RICHLANDG80-vs-8120 . 137 |========================== Timed HMMer Search 2.3.2 Pfam Database Search Seconds < Lower Is Better GCC 4.8.1 ...................... 16.10 |===================================== LLVM Clang 3.3 ................. 16.35 |====================================== 1306306-SO-RICHLANDG80-vs-8120 . 16.77 |======================================= Timed MAFFT Alignment 6.864 Multiple Sequence Alignment Seconds < Lower Is Better GCC 4.8.1 ...................... 9.23 |================================= LLVM Clang 3.3 ................. 10.91 |======================================= 1306306-SO-RICHLANDG80-vs-8120 . 9.11 |================================= x264 2013-06-08 H.264 Video Encoding Frames Per Second > Higher Is Better GCC 4.8.1 ...... 97.88 |======================================================= LLVM Clang 3.3 . 94.48 |=====================================================