AMD EPYC Rome Spectre Mitigation Benchmarks
Initial benchmarks for a future article.
HTML result view exported from: https://openbenchmarking.org/result/1908094-AS-MITIGATIO58&sro&grs.
MKL-DNN
Performance / Cost - Harness: Deconvolution Batch deconv_1d - Data Type: f32
Cpuminer-Opt
Performance / Cost - Algorithm: m7m
DaCapo Benchmark
Performance / Cost - Java Test: Tradebeans
Hackbench
Performance / Cost - Count: 32 - Type: Process
Cpuminer-Opt
Performance / Cost - Algorithm: deep
Stress-NG
Performance / Cost - Test: Crypto
C-Ray
Performance / Cost - Total Time - 4K, 16 Rays Per Pixel
MKL-DNN
Harness: Convolution Batch conv_googlenet_v3 - Data Type: f32
m-queens
Performance / Cost - Time To Solve
dav1d
Performance / Cost - Video Input: Summer Nature 1080p
MKL-DNN
Harness: Convolution Batch conv_alexnet - Data Type: f32
Sysbench
Performance / Cost - Test: CPU
CP2K Molecular Dynamics
Performance / Cost - Fayalite-FIST Data
John The Ripper
Performance / Cost - Test: Blowfish
Hackbench
Count: 32 - Type: Process
Novabench
Performance / Cost - Test: RAM
Cpuminer-Opt
Algorithm: m7m
Blender
Performance / Cost - Blend File: Classroom - Compute: CPU-Only
Himeno Benchmark
Poisson Pressure Solver
Stress-NG
Performance / Cost - Test: CPU Stress
Stockfish
Performance / Cost - Total Time
7-Zip Compression
Compress Speed Test
RAMspeed SMP
Performance / Cost - Type: Add - Benchmark: Floating Point
Stream
Performance / Cost - Type: Triad
Cpuminer-Opt
Algorithm: skein
SVT-AV1
1080p 8-bit YUV To AV1 Video Encode
OpenSSL
RSA 4096-bit Performance
OpenSSL
Performance / Cost - RSA 4096-bit Performance
Timed GCC Compilation
Performance / Cost - Time To Compile
Stream
Performance / Cost - Type: Add
Cpuminer-Opt
Performance / Cost - Algorithm: sha256t
Novabench
Performance / Cost - Test: CPU
NAS Parallel Benchmarks
Performance / Cost - Test / Class: EP.C
asmFish
Performance / Cost - 1024 Hash Memory, 26 Depth
NAS Parallel Benchmarks
Test / Class: EP.C
Stream
Performance / Cost - Type: Scale
Stress-NG
Test: Crypto
Geekbench
Performance / Cost - Test: Multi Core
Sysbench
Test: CPU
Stockfish
Total Time
dav1d
Performance / Cost - Video Input: Summer Nature 4K
Cpuminer-Opt
Performance / Cost - Algorithm: skein
RAMspeed SMP
Performance / Cost - Type: Average - Benchmark: Floating Point
GIMP
Performance / Cost - Test: rotate
Rodinia
Performance / Cost - Test: OpenMP LavaMD
NAS Parallel Benchmarks
Performance / Cost - Test / Class: LU.C
C-Ray
Total Time - 4K, 16 Rays Per Pixel
Blender
Performance / Cost - Blend File: Barbershop - Compute: CPU-Only
RAMspeed SMP
Performance / Cost - Type: Scale - Benchmark: Floating Point
m-queens
Time To Solve
RAMspeed SMP
Performance / Cost - Type: Add - Benchmark: Integer
Stress-NG
Test: CPU Stress
Blender
Performance / Cost - Blend File: BMW27 - Compute: CPU-Only
GIMP
Performance / Cost - Test: resize
Blender
Blend File: Barbershop - Compute: CPU-Only
7-Zip Compression
Performance / Cost - Compress Speed Test
POV-Ray
Performance / Cost - Trace Time
Stream
Performance / Cost - Type: Copy
OpenCV Benchmark
Performance / Cost -
Appleseed
Performance / Cost - Scene: Disney Material
asmFish
1024 Hash Memory, 26 Depth
Blender
Blend File: Classroom - Compute: CPU-Only
Timed Linux Kernel Compilation
Performance / Cost - Time To Compile
RAMspeed SMP
Performance / Cost - Type: Average - Benchmark: Integer
Geekbench
Test: Multi Core
Rust Mandelbrot
Performance / Cost - Time To Complete Serial/Parallel Mandelbrot
CP2K Molecular Dynamics
Fayalite-FIST Data
Blender
Performance / Cost - Blend File: Fishy Cat - Compute: CPU-Only
libjpeg-turbo tjbench
Performance / Cost - Test: Decompression Throughput
Himeno Benchmark
Performance / Cost - Poisson Pressure Solver
x264
Performance / Cost - H.264 Video Encoding
SVT-VP9
Performance / Cost - 1080p 8-bit YUV To VP9 Video Encode
SVT-HEVC
Performance / Cost - 1080p 8-bit YUV To HEVC Video Encode
ctx_clock
Performance / Cost - Context Switch Time
NAMD
Performance / Cost - ATPase Simulation - 327,506 Atoms
RAMspeed SMP
Performance / Cost - Type: Copy - Benchmark: Floating Point
Timed LLVM Compilation
Performance / Cost - Time To Compile
RAMspeed SMP
Performance / Cost - Type: Scale - Benchmark: Integer
dav1d
Video Input: Summer Nature 1080p
Scikit-Learn
Performance / Cost -
Blender
Blend File: BMW27 - Compute: CPU-Only
RAMspeed SMP
Performance / Cost - Type: Copy - Benchmark: Integer
MKL-DNN
Performance / Cost - Harness: Convolution Batch conv_alexnet - Data Type: f32
Timed PHP Compilation
Performance / Cost - Time To Compile
Geekbench
Performance / Cost - Test: Single Core
Numenta Anomaly Benchmark
Performance / Cost - Time To Completion
glibc bench
Performance / Cost - Benchmark: sin
MKL-DNN
Performance / Cost - Harness: Convolution Batch conv_googlenet_v3 - Data Type: f32
glibc bench
Performance / Cost - Benchmark: cos
GIMP
Performance / Cost - Test: auto-levels
Radiance Benchmark
Performance / Cost - Test: SMP Parallel
Novabench
Test: CPU
Redis
Performance / Cost - Test: GET
dav1d
Video Input: Summer Nature 4K
Tesseract OCR
Performance / Cost - Time To OCR 7 Images
NAS Parallel Benchmarks
Test / Class: LU.C
Timed LLVM Compilation
Time To Compile
NAMD
ATPase Simulation - 327,506 Atoms
Radiance Benchmark
Performance / Cost - Test: Serial
GIMP
Performance / Cost - Test: unsharp-mask
DaCapo Benchmark
Performance / Cost - Java Test: Jython
Numpy Benchmark
RAMspeed SMP
Type: Add - Benchmark: Floating Point
PHPBench
PHP Benchmark Suite
SPECjbb 2015
Performance / Cost - SPECjbb2015-Composite max-jOPS
Renaissance
Performance / Cost - Test: Scala Dotty
RAMspeed SMP
Type: Scale - Benchmark: Integer
Blender
Blend File: Fishy Cat - Compute: CPU-Only
RAMspeed SMP
Type: Scale - Benchmark: Floating Point
glibc bench
Performance / Cost - Benchmark: pthread_once
Zstd Compression
Compressing ubuntu-16.04.3-server-i386.img, Compression Level 19
Cython benchmark
Performance / Cost -
PyBench
Performance / Cost - Total For Average Test Times
glibc bench
Performance / Cost - Benchmark: sqrt
Rodinia
Test: OpenMP LavaMD
POV-Ray
Trace Time
RAMspeed SMP
Type: Average - Benchmark: Floating Point
Appleseed
Scene: Emily
RAMspeed SMP
Type: Average - Benchmark: Integer
Apache Siege
Performance / Cost - Concurrent Users: 250
RAMspeed SMP
Type: Add - Benchmark: Integer
Redis
Performance / Cost - Test: SET
Redis
Test: SET
MariaDB
Performance / Cost - Clients: 128
glibc bench
Benchmark: sqrt
RAMspeed SMP
Type: Copy - Benchmark: Floating Point
Renaissance
Performance / Cost - Test: Apache Spark PageRank
Appleseed
Performance / Cost - Scene: Emily
GIMP
Test: unsharp-mask
PyBench
Total For Average Test Times
Cython benchmark
Zstd Compression
Performance / Cost - Compressing ubuntu-16.04.3-server-i386.img, Compression Level 19
glibc bench
Benchmark: pthread_once
SPECjbb 2015
Performance / Cost - SPECjbb2015-Composite critical-jOPS
RAMspeed SMP
Type: Copy - Benchmark: Integer
PHPBench
Performance / Cost - PHP Benchmark Suite
libjpeg-turbo tjbench
Test: Decompression Throughput
Numpy Benchmark
Performance / Cost -
DaCapo Benchmark
Java Test: Jython
Renaissance
Performance / Cost - Test: Savina Reactors.IO
Radiance Benchmark
Test: Serial
x265
H.265 1080p Video Encoding
GIMP
Test: auto-levels
Numenta Anomaly Benchmark
Time To Completion
Radiance Benchmark
Test: SMP Parallel
Tesseract OCR
Time To OCR 7 Images
x264
H.264 Video Encoding
glibc bench
Benchmark: cos
Novabench
Test: RAM
glibc bench
Benchmark: sin
Stream
Type: Add
Geekbench
Test: Single Core
Timed GCC Compilation
Time To Compile
Timed PHP Compilation
Time To Compile
OpenCV Benchmark
Scikit-Learn
Rust Mandelbrot
Time To Complete Serial/Parallel Mandelbrot
Stream
Type: Scale
Appleseed
Scene: Disney Material
ctx_clock
Context Switch Time
GIMP
Test: resize
GIMP
Test: rotate
Novabench
Test: RAM
MariaDB
Performance / Cost - Clients: 256
SVT-AV1
Performance / Cost - 1080p 8-bit YUV To AV1 Video Encode
Stream
Type: Triad
Stream
Type: Copy
Cpuminer-Opt
Algorithm: sha256t
Cpuminer-Opt
Algorithm: deep
Apache Siege
Concurrent Users: 250
Redis
Test: GET
Timed Linux Kernel Compilation
Time To Compile
SVT-VP9
1080p 8-bit YUV To VP9 Video Encode
SVT-HEVC
1080p 8-bit YUV To HEVC Video Encode
MKL-DNN
Harness: Deconvolution Batch deconv_1d - Data Type: f32
John The Ripper
Test: Blowfish
Renaissance
Test: Apache Spark PageRank
Renaissance
Test: Savina Reactors.IO
Renaissance
Test: Scala Dotty
DaCapo Benchmark
Java Test: Tradebeans
Phoronix Test Suite v10.8.4