OpenCL Radeon Open-Source vs. Catalyst Radeon R600g and RadeonSI Gallium3D Clover OpenCL benchmarks with clpeak using Linux 4.5 + Mesa 11.2-devel. Compared to Catalyst. Benchmark results by Michael Larabel of Phoronix.com for a future article.
HTML result view exported from: https://openbenchmarking.org/result/1602026-GA-OPENCLAMD14&grs&sro&rro .
Processor Motherboard Chipset Memory Disk Graphics Audio Monitor Network OS Kernel Desktop Display Server Display Driver OpenGL Compiler File-System Screen Resolution Open-Source Catalyst HD 6870 HD 7950 R9 285 R9 290 R7 370 HD 6870 HD 7950 R9 285 R9 290 R7 370 Intel Core i7-5960X @ 3.50GHz (16 Cores) Gigabyte X99-UD4-CF Intel Xeon E7 v3/Xeon 16384MB 2 x 120GB SanDisk SDSSDA12 + 120GB Samsung SSD 850 Sapphire AMD Radeon HD 6870 1024MB Realtek ALC1150 DELL P2415Q Intel Connection Ubuntu 15.10 4.5.0-rc2amdgpu-pp+ (x86_64) 20160201 Unity X Server 1.17.2 radeon 7.6.99 3.3 Mesa 11.2.0-devel (padoka PPA) Gallium 0.4 GCC 5.2.1 20151010 ext4 3840x2160 XFX AMD Radeon HD 7950/8950 OEM / R9 280 3072MB 4.1 Mesa 11.2.0-devel (padoka PPA) Gallium 0.4 XFX AMD Radeon R9 285/380 2048MB amdgpu 1.0.99 XFX AMD Radeon R9 290 4096MB radeon 7.6.99 MSI AMD Radeon R7 370 / R9 270/370 OEM 4096MB Sapphire AMD Radeon HD 6800 1024MB 4.2.0-27-generic (x86_64) fglrx 15.20.3 4.5.13399 XFX AMD Radeon HD 7900 3072MB Supported device 6939 2048MB Supported device 67B1 4096MB MSI AMD Radeon R7 370 4096MB OpenBenchmarking.org Kernel Details - Open-Source: HD 6870, Open-Source: HD 7950, Open-Source: R9 285, Open-Source: R9 290, Open-Source: R7 370: amdgpu.powerplay=1 Compiler Details - --build=x86_64-linux-gnu --disable-browser-plugin --disable-vtable-verify --disable-werror --enable-checking=release --enable-clocale=gnu --enable-gnu-unique-object --enable-gtk-cairo --enable-java-awt=gtk --enable-java-home --enable-languages=c,ada,c++,java,go,d,fortran,objc,obj-c++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-arch-directory=amd64 --with-default-libstdcxx-abi=new --with-multilib-list=m32,m64,mx32 --with-tune=generic -v Processor Details - Scaling Governor: intel_pstate performance Graphics Details - Open-Source: HD 6870: EXA - Open-Source: HD 7950: GLAMOR - Open-Source: R9 285: GLAMOR - Open-Source: R9 290: GLAMOR - Open-Source: R7 370: GLAMOR Environment Details - Catalyst: HD 6870, Catalyst: HD 7950, Catalyst: R9 285, Catalyst: R9 290, Catalyst: R7 370: LIBGL_DRIVERS_PATH=/usr/lib/fglrx/dri:/usr/lib/x86_64-linux-gnu/dri:/usr/lib/dri:/usr/lib32/fglrx/dri:/usr/lib/i386-linux-gnu/dri
clpeak: Integer Compute INT clpeak: Double-Precision Double clpeak: Transfer Bandwidth enqueueWriteBuffer clpeak: Single-Precision Float clpeak: Transfer Bandwidth enqueueReadBuffer clpeak: Kernel Latency Open-Source Catalyst HD 6870 HD 7950 R9 285 R9 290 R7 370 HD 6870 HD 7950 R9 285 R9 290 R7 370 182.44 5.26 360.92 3.58 145.85 491.62 547.03 15.72 2345.12 10.40 105.54 650.80 204.75 14.82 3102.60 10.09 143.39 1578.01 596.55 12.32 4608.55 9.82 118.71 418.74 131.49 15.77 2031.91 10.41 110.24 199.38 9.49 398.29 8.80 24.90 516.74 536.84 13.88 2292.54 10.47 34.03 647.08 202.71 13.21 3177.62 9.00 26.78 1591.74 598.81 14.81 4721.18 10.64 37.77 415.67 128.73 14.97 2095.91 10.67 35.42 OpenBenchmarking.org
clpeak OpenCL Test: Integer Compute INT Open-Source Catalyst OpenBenchmarking.org GIOPS, More Is Better clpeak OpenCL Test: Integer Compute INT R9 290 R9 285 R7 370 HD 7950 HD 6870 300 600 900 1200 1500 SE +/- 0.49, N = 3 SE +/- 0.05, N = 3 SE +/- 0.01, N = 3 SE +/- 7.09, N = 6 SE +/- 0.16, N = 3 SE +/- 0.05, N = 3 SE +/- 1.77, N = 3 SE +/- 3.65, N = 3 SE +/- 21.03, N = 6 SE +/- 0.02, N = 3 1578.01 650.80 418.74 491.62 182.44 1591.74 647.08 415.67 516.74 199.38
clpeak OpenCL Test: Double-Precision Double Open-Source Catalyst OpenBenchmarking.org GFLOPS, More Is Better clpeak OpenCL Test: Double-Precision Double R9 290 R9 285 R7 370 HD 7950 130 260 390 520 650 SE +/- 0.10, N = 3 SE +/- 0.03, N = 3 SE +/- 0.00, N = 3 SE +/- 16.28, N = 6 SE +/- 0.01, N = 3 SE +/- 0.22, N = 3 SE +/- 0.36, N = 3 SE +/- 11.37, N = 6 596.55 204.75 131.49 547.03 598.81 202.71 128.73 536.84
clpeak OpenCL Test: Transfer Bandwidth enqueueWriteBuffer Open-Source Catalyst OpenBenchmarking.org GBPS, More Is Better clpeak OpenCL Test: Transfer Bandwidth enqueueWriteBuffer R9 290 R9 285 R7 370 HD 7950 HD 6870 4 8 12 16 20 SE +/- 0.06, N = 3 SE +/- 0.59, N = 6 SE +/- 0.00, N = 3 SE +/- 0.05, N = 3 SE +/- 0.00, N = 3 SE +/- 0.03, N = 3 SE +/- 0.19, N = 3 SE +/- 0.00, N = 3 SE +/- 0.04, N = 3 SE +/- 0.12, N = 3 12.32 14.82 15.77 15.72 5.26 14.81 13.21 14.97 13.88 9.49
clpeak OpenCL Test: Single-Precision Float Open-Source Catalyst OpenBenchmarking.org GFLOPS, More Is Better clpeak OpenCL Test: Single-Precision Float R9 290 R9 285 R7 370 HD 7950 HD 6870 1000 2000 3000 4000 5000 SE +/- 2.13, N = 3 SE +/- 1.52, N = 3 SE +/- 0.22, N = 3 SE +/- 47.92, N = 6 SE +/- 0.08, N = 3 SE +/- 0.95, N = 3 SE +/- 0.24, N = 3 SE +/- 2.25, N = 3 SE +/- 44.66, N = 6 SE +/- 0.00, N = 3 4608.55 3102.60 2031.91 2345.12 360.92 4721.18 3177.62 2095.91 2292.54 398.29
clpeak OpenCL Test: Transfer Bandwidth enqueueReadBuffer Open-Source Catalyst OpenBenchmarking.org GBPS, More Is Better clpeak OpenCL Test: Transfer Bandwidth enqueueReadBuffer R9 290 R9 285 R7 370 HD 7950 HD 6870 3 6 9 12 15 SE +/- 0.25, N = 6 SE +/- 0.30, N = 6 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.10, N = 6 SE +/- 0.13, N = 3 SE +/- 0.07, N = 3 SE +/- 0.04, N = 3 SE +/- 0.11, N = 3 SE +/- 0.08, N = 3 9.82 10.09 10.41 10.40 3.58 10.64 9.00 10.67 10.47 8.80
clpeak OpenCL Test: Kernel Latency Open-Source Catalyst OpenBenchmarking.org us, Fewer Is Better clpeak OpenCL Test: Kernel Latency R9 290 R9 285 R7 370 HD 7950 HD 6870 30 60 90 120 150 SE +/- 2.08, N = 6 SE +/- 14.32, N = 6 SE +/- 3.85, N = 6 SE +/- 1.78, N = 6 SE +/- 8.80, N = 6 SE +/- 2.12, N = 6 SE +/- 0.38, N = 3 SE +/- 0.52, N = 5 SE +/- 0.68, N = 3 SE +/- 0.28, N = 3 118.71 143.39 110.24 105.54 145.85 37.77 26.78 35.42 34.03 24.90
Phoronix Test Suite v10.8.5