KVM testing on Ubuntu 18.04 via the Phoronix Test Suite.
c5ad.4xlarge Processor: AMD EPYC 7R32 (8 Cores / 16 Threads), Motherboard: Amazon EC2 c5ad.4xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 32GB, Disk: 2 x 300GB Amazon EC2 NVMe Instance Storage + 9GB Amazon Elastic Block Store, Network: Amazon Elastic
OS: Ubuntu 18.04, Kernel: 5.3.0-1032-aws (x86_64), Compiler: GCC 7.5.0, File-System: ext4, System Layer: KVM
Environment Notes: COST_PERF_PER_HOUR=0.688Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / relatime,rwProcessor Notes: CPU Microcode: 0x8301025Python Notes: Python 2.7.17 + Python 3.6.9Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
c5a.4xlarge Changed Motherboard to Amazon EC2 c5a.4xlarge (1.0 BIOS) .
Changed Disk to 137GB Amazon Elastic Block Store .
Environment Change: COST_PERF_PER_HOUR=0.616Disk Change: NONE / discard,relatime,rw
EC2 c5ad Benchmarks Processor Motherboard Chipset Memory Disk Network OS Kernel Compiler File-System System Layer c5ad.4xlarge c5a.4xlarge AMD EPYC 7R32 (8 Cores / 16 Threads) Amazon EC2 c5ad.4xlarge (1.0 BIOS) Intel 440FX 82441FX PMC 32GB 2 x 300GB Amazon EC2 NVMe Instance Storage + 9GB Amazon Elastic Block Store Amazon Elastic Ubuntu 18.04 5.3.0-1032-aws (x86_64) GCC 7.5.0 ext4 KVM Amazon EC2 c5a.4xlarge (1.0 BIOS) 137GB Amazon Elastic Block Store OpenBenchmarking.org Environment Details - c5ad.4xlarge: COST_PERF_PER_HOUR=0.688 - c5a.4xlarge: COST_PERF_PER_HOUR=0.616 Compiler Details - --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib --with-tune=generic --without-cuda-driver -v Disk Details - c5ad.4xlarge: NONE / relatime,rw - c5a.4xlarge: NONE / discard,relatime,rw Processor Details - CPU Microcode: 0x8301025 Python Details - Python 2.7.17 + Python 3.6.9 Security Details - itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
c5ad.4xlarge vs. c5a.4xlarge Comparison Phoronix Test Suite Baseline +197.1% +197.1% +394.2% +394.2% +591.3% +591.3% 83.4% 9.4% 5.4% 2.4% Write Test 788.2% Read Test Rand Fill Sync HWB Color Space Read While Writing 4% Rand Read 2.7% Rotate P.P.A.S.T 2.1% IOR IOR Facebook RocksDB GraphicsMagick Facebook RocksDB Facebook RocksDB GraphicsMagick Hugin c5ad.4xlarge c5a.4xlarge
EC2 c5ad Benchmarks build-linux-kernel: Time To Compile svt-av1: Enc Mode 8 - 1080p svt-av1: Enc Mode 4 - 1080p x265: H.265 1080p Video Encoding avifenc: 0 avifenc: 2 avifenc: 8 avifenc: 10 graphics-magick: HWB Color Space graphics-magick: Noise-Gaussian graphics-magick: Enhanced graphics-magick: Resizing graphics-magick: Rotate graphics-magick: Sharpen graphics-magick: Swirl rocksdb: Seq Fill rocksdb: Rand Fill rocksdb: Rand Fill Sync rocksdb: Rand Read rocksdb: Read While Writing build2: Time To Compile appleseed: Disney Material appleseed: Material Tester oidn: Memorial openvkl: vklBenchmark openvkl: vklBenchmarkVdbVolume embree: Pathtracer - Asian Dragon embree: Pathtracer - Crown embree: Pathtracer ISPC - Asian Dragon embree: Pathtracer ISPC - Crown ospray: NASA Streamlines - SciVis ospray: NASA Streamlines - Path Tracer ospray: San Miguel - SciVis gromacs: Water Benchmark namd: ATPase Simulation - 327,506 Atoms ior: Write Test ior: Read Test hpcg: rawtherapee: Total Benchmark Time tesseract-ocr: Time To OCR 7 Images ocrmypdf: Processing 60 Page PDF Document hugin: Panorama Photo Assistant + Stitching Time montage: Mosaic of M17, K band, 1.5 deg x 1.5 deg c5ad.4xlarge c5a.4xlarge 83.009 25.580 3.104 43.61 106.817 63.841 6.622 6.271 833 197 165 787 506 100 392 629334 518769 5111 29887633 1526922 107.770 282.360588 270.090637 6.65 98.14 9509058.0530303 9.9683 8.9553 9.9750 8.6148 15.15 2.98 10.56 0.890 2.64664 1200.71 219.84 5.09672 63.730 33.390 39.793 68.029 127.623 83.463 25.436 3.117 44.36 106.666 63.863 6.635 6.293 878 199 165 799 518 100 398 627297 521252 5592 29112755 1467545 109.701 282.401483 269.142691 6.71 98.33 9580048.4696970 9.9442 8.9583 9.9007 8.5938 15.08 2.98 10.53 0.884 2.63938 135.19 403.08 5.07705 63.775 33.701 39.971 69.449 127.497 OpenBenchmarking.org
OSPray Intel OSPray is a portable ray-tracing engine for high-performance, high-fidenlity scientific visualizations. OSPray builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
NAMD NAMD is a parallel molecular dynamics code designed for high-performance simulation of large biomolecular systems. NAMD was developed by the Theoretical and Computational Biophysics Group in the Beckman Institute for Advanced Science and Technology at the University of Illinois at Urbana-Champaign. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org days/ns, Fewer Is Better NAMD 2.13 ATPase Simulation - 327,506 Atoms c5ad.4xlarge c5a.4xlarge 0.5955 1.191 1.7865 2.382 2.9775 SE +/- 0.00251, N = 3 SE +/- 0.00828, N = 3 2.64664 2.63938 Performance / Cost
OpenBenchmarking.org days/ns x Dollar, Fewer Is Better NAMD 2.13 Performance / Cost - ATPase Simulation - 327,506 Atoms c5ad.4xlarge c5a.4xlarge 0.0473 0.0946 0.1419 0.1892 0.2365 0.21 0.18 1. c5ad.4xlarge: $0.688 reported cost per hour, test consumed 6 Minutes, 56 Seconds: cost approximately 0.08 dollar. 2. c5a.4xlarge: $0.616 reported cost per hour, test consumed 6 Minutes, 56 Seconds: cost approximately 0.07 dollar. Result Confidence
OpenBenchmarking.org days/ns, Fewer Is Better NAMD 2.13 ATPase Simulation - 327,506 Atoms c5ad.4xlarge c5a.4xlarge 2 4 6 8 10 Min: 2.64 / Avg: 2.65 / Max: 2.65 Min: 2.62 / Avg: 2.64 / Max: 2.65
High Performance Conjugate Gradient HPCG is the High Performance Conjugate Gradient and is a new scientific benchmark from Sandia National Lans focused for super-computer testing with modern real-world workloads compared to HPCC. Learn more via the OpenBenchmarking.org test page.
RawTherapee RawTherapee is a cross-platform, open-source multi-threaded RAW image processing program. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org Seconds, Fewer Is Better RawTherapee Total Benchmark Time c5ad.4xlarge c5a.4xlarge 14 28 42 56 70 SE +/- 0.06, N = 3 SE +/- 0.04, N = 3 63.73 63.78 1. RawTherapee, version 5.3, command line
An advanced, cross-platform program for developing raw photos.
Website: http://www.rawtherapee.com/
Documentation: http://rawpedia.rawtherapee.com/
Forum: https://discuss.pixls.us/c/software/rawtherapee
Code and bug reports: https://github.com/Beep6581/RawTherapee
Symbols:
<Chevrons> indicate parameters you can change.
[Square brackets] mean the parameter is optional.
The pipe symbol | indicates a choice of one or the other.
The dash symbol - denotes a range of possible values from one to the other.
Usage:
rawtherapee-cli -c <dir>|<files> Convert files in batch with default parameters.
rawtherapee-cli <other options> -c <dir>|<files> Convert files in batch with your own settings.
Options:
rawtherapee-cli[-o <output>|-O <output>] [-q] [-a] [-s|-S] [-p <one.pp3> [-p <two.pp3> ...] ] [-d] [ -j[1-100] [-js<1-3>] | [-b<8|16>] [-t[z] | [-n]] ] [-Y] [-f] -c <input>
-c <files> Specify one or more input files or directory.
When specifying directories, Rawtherapee will look for images files that comply with the
selected extensions (see also '-a').
-c must be the last option.
-o <file>|<dir> Set output file or folder.
Saves output file alongside input file if -o is not specified.
-O <file>|<dir> Set output file or folder and copy pp3 file into it.
Saves output file alongside input file if -O is not specified.
-q Quick-start mode. Does not load cached files to speedup start time.
-a Process all supported image file types when specifying a folder, even those
not currently selected in Preferences > File Browser > Parsed Extensions.
-s Use the existing sidecar file to build the processing parameters,
e.g. for photo.raw there should be a photo.raw.pp3 file in the same folder.
If the sidecar file does not exist, neutral values will be used.
-S Like -s but skip if the sidecar file does not exist.
-p <file.pp3> Specify processing profile to be used for all conversions.
You can specify as many sets of "-p <file.pp3>" options as you like,
each will be built on top of the previous one, as explained below.
-d Use the default raw or non-raw processing profile as set in
Preferences > Image Processing > Default Processing Profile
-j[1-100] Specify output to be JPEG (default, if -t and -n are not set).
Optionally, specify compression 1-100 (default value: 92).
-js<1-3> Specify the JPEG chroma subsampling parameter, where:
1 = Best compression: 2x2, 1x1, 1x1 (4:2:0)
Chroma halved vertically and horizontally.
2 = Balanced (default): 2x1, 1x1, 1x1 (4:2:2)
Chroma halved horizontally.
3 = Best quality: 1x1, 1x1, 1x1 (4:4:4)
No chroma subsampling.
-b<8|16> Specify bit depth per channel (default value: 16 for TIFF, 8 for PNG).
Only applies to TIFF and PNG output, JPEG is always 8.
-t[z] Specify output to be TIFF.
Uncompressed by default, or deflate compression with 'z'.
-n Specify output to be compressed PNG.
Compression is hard-coded to 6.
-Y Overwrite output if present.
-f Use the custom fast-export processing pipeline.
Your pp3 files can be incomplete, RawTherapee will build the final values as follows:
1- A new processing profile is created using neutral values,
2- If the "-d" option is set, the values are overridden by those found in
the default raw or non-raw processing profile.
3- If one or more "-p" options are set, the values are overridden by those
found in these processing profiles.
4- If the "-s" or "-S" options are set, the values are finally overridden by those
found in the sidecar files.
The processing profiles are processed in the order specified on the command line. Performance / Cost
OpenBenchmarking.org Seconds x Dollar, Fewer Is Better RawTherapee Performance / Cost - Total Benchmark Time c5ad.4xlarge c5a.4xlarge 0.5738 1.1476 1.7214 2.2952 2.869 2.55 2.55 1. c5ad.4xlarge: $0.688 reported cost per hour, test consumed 3 Minutes, 27 Seconds: cost approximately 0.04 dollar. 2. c5a.4xlarge: $0.616 reported cost per hour, test consumed 3 Minutes, 27 Seconds: cost approximately 0.04 dollar. Result Confidence
OpenBenchmarking.org Seconds, Fewer Is Better RawTherapee Total Benchmark Time c5ad.4xlarge c5a.4xlarge 12 24 36 48 60 Min: 63.67 / Avg: 63.73 / Max: 63.84 Min: 63.73 / Avg: 63.77 / Max: 63.86 1. RawTherapee, version 5.3, command line
An advanced, cross-platform program for developing raw photos.
Website: http://www.rawtherapee.com/
Documentation: http://rawpedia.rawtherapee.com/
Forum: https://discuss.pixls.us/c/software/rawtherapee
Code and bug reports: https://github.com/Beep6581/RawTherapee
Symbols:
<Chevrons> indicate parameters you can change.
[Square brackets] mean the parameter is optional.
The pipe symbol | indicates a choice of one or the other.
The dash symbol - denotes a range of possible values from one to the other.
Usage:
rawtherapee-cli -c <dir>|<files> Convert files in batch with default parameters.
rawtherapee-cli <other options> -c <dir>|<files> Convert files in batch with your own settings.
Options:
rawtherapee-cli[-o <output>|-O <output>] [-q] [-a] [-s|-S] [-p <one.pp3> [-p <two.pp3> ...] ] [-d] [ -j[1-100] [-js<1-3>] | [-b<8|16>] [-t[z] | [-n]] ] [-Y] [-f] -c <input>
-c <files> Specify one or more input files or directory.
When specifying directories, Rawtherapee will look for images files that comply with the
selected extensions (see also '-a').
-c must be the last option.
-o <file>|<dir> Set output file or folder.
Saves output file alongside input file if -o is not specified.
-O <file>|<dir> Set output file or folder and copy pp3 file into it.
Saves output file alongside input file if -O is not specified.
-q Quick-start mode. Does not load cached files to speedup start time.
-a Process all supported image file types when specifying a folder, even those
not currently selected in Preferences > File Browser > Parsed Extensions.
-s Use the existing sidecar file to build the processing parameters,
e.g. for photo.raw there should be a photo.raw.pp3 file in the same folder.
If the sidecar file does not exist, neutral values will be used.
-S Like -s but skip if the sidecar file does not exist.
-p <file.pp3> Specify processing profile to be used for all conversions.
You can specify as many sets of "-p <file.pp3>" options as you like,
each will be built on top of the previous one, as explained below.
-d Use the default raw or non-raw processing profile as set in
Preferences > Image Processing > Default Processing Profile
-j[1-100] Specify output to be JPEG (default, if -t and -n are not set).
Optionally, specify compression 1-100 (default value: 92).
-js<1-3> Specify the JPEG chroma subsampling parameter, where:
1 = Best compression: 2x2, 1x1, 1x1 (4:2:0)
Chroma halved vertically and horizontally.
2 = Balanced (default): 2x1, 1x1, 1x1 (4:2:2)
Chroma halved horizontally.
3 = Best quality: 1x1, 1x1, 1x1 (4:4:4)
No chroma subsampling.
-b<8|16> Specify bit depth per channel (default value: 16 for TIFF, 8 for PNG).
Only applies to TIFF and PNG output, JPEG is always 8.
-t[z] Specify output to be TIFF.
Uncompressed by default, or deflate compression with 'z'.
-n Specify output to be compressed PNG.
Compression is hard-coded to 6.
-Y Overwrite output if present.
-f Use the custom fast-export processing pipeline.
Your pp3 files can be incomplete, RawTherapee will build the final values as follows:
1- A new processing profile is created using neutral values,
2- If the "-d" option is set, the values are overridden by those found in
the default raw or non-raw processing profile.
3- If one or more "-p" options are set, the values are overridden by those
found in these processing profiles.
4- If the "-s" or "-S" options are set, the values are finally overridden by those
found in the sidecar files.
The processing profiles are processed in the order specified on the command line.
Tesseract OCR Tesseract-OCR is the open-source optical character recognition (OCR) engine for the conversion of text within images to raw text output. This test profile relies upon a system-supplied Tesseract installation. Learn more via the OpenBenchmarking.org test page.
OCRMyPDF OCRMyPDF is an optical character recognition (OCR) text layer to scanned PDF files, producing new PDFs with the text now selectable/searchable/copy-paste capable. OCRMyPDF leverages the Tesseract OCR engine and is written in Python. Learn more via the OpenBenchmarking.org test page.
Montage Astronomical Image Mosaic Engine Montage is an open-source astronomical image mosaic engine. This BSD-licensed astronomy software is developed by the California Institute of Technology, Pasadena. Learn more via the OpenBenchmarking.org test page.
Cost To Run Tests OpenBenchmarking.org Cost / Price Per Hour, Fewer Is Better Cost To Run Tests Cost / Price Per Hour c5ad.4xlarge c5a.4xlarge 2 4 6 8 10 8.64 1.66 1. c5ad.4xlarge: $0.688 reported cost per hour, running tests consumed 12 Hours, 33 Minutes, 42 Seconds: cost approximately 8.64 dollar. 2. c5a.4xlarge: $0.616 reported cost per hour, running tests consumed 2 Hours, 41 Minutes, 36 Seconds: cost approximately 1.66 dollar.
c5ad.4xlarge Processor: AMD EPYC 7R32 (8 Cores / 16 Threads), Motherboard: Amazon EC2 c5ad.4xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 32GB, Disk: 2 x 300GB Amazon EC2 NVMe Instance Storage + 9GB Amazon Elastic Block Store, Network: Amazon Elastic
OS: Ubuntu 18.04, Kernel: 5.3.0-1032-aws (x86_64), Compiler: GCC 7.5.0, File-System: ext4, System Layer: KVM
Environment Notes: COST_PERF_PER_HOUR=0.688Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / relatime,rwProcessor Notes: CPU Microcode: 0x8301025Python Notes: Python 2.7.17 + Python 3.6.9Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 14 August 2020 21:42 by user root.
c5a.4xlarge Processor: AMD EPYC 7R32 (8 Cores / 16 Threads), Motherboard: Amazon EC2 c5a.4xlarge (1.0 BIOS), Chipset: Intel 440FX 82441FX PMC, Memory: 32GB, Disk: 137GB Amazon Elastic Block Store, Network: Amazon Elastic
OS: Ubuntu 18.04, Kernel: 5.3.0-1032-aws (x86_64), Compiler: GCC 7.5.0, File-System: ext4, System Layer: KVM
Environment Notes: COST_PERF_PER_HOUR=0.616Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / discard,relatime,rwProcessor Notes: CPU Microcode: 0x8301025Python Notes: Python 2.7.17 + Python 3.6.9Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 15 August 2020 10:54 by user root.