Skip navigation

Standard Performance Evaluation Corporation

Facebook logo LinkedIn logo Twitter logo Google+ logo
 
 

SPECworkstationTM 3 workloads

Media and Entertainment

Blender

This workload uses the open-source Blender (blender.org) application to measure the performance of content-creation workflows. It contains the following tests:

  • _start_3BMWs.blend: This is a scene with ~470K vertices and ~900K triangles. The scene includes three identical car models. It is rendered with the OpenGL renderer as the camera pans through 360 degrees.
  • _start_BMW1M.blend: This is a scene with ~91K vertices and ~172K triangles. It uses a single car model. A single frame is rendered with the Blender Cycles ray-tracing renderer.
  • _start_island.blend: This is a scene with ~16K vertices and 32K triangles. It includes a particle system to represent a large quantity of birds.
  • _start_BMW27.blend: This scene depicts two cars rendered with the Blender Cycles ray-tracing renderer. It contains 99K vertices and 187K triangles.
  • _start_classroom.blend: An interior scene with lots of light sources and bounced lights, including dust particles suspended in the air and light passing through a window and emanating from a lamp.

Handbrake

This workload is based on the open-source media encoder handbrake (www.handbrake.fr). It takes a 4K mp4 file and transcodes it with an H.265 encoder at two different resolutions. The first transcoding down-samples the file from 4K to 2K. The second trans coding goes from H264 4K to H265 4K.

LuxRender

LuxRender uses LuxMark, a benchmark based on the new LuxCore physically based renderer, to render a chrome sphere resting on a grid of numbers in a beach scene.

Maya

This is the maya-05 viewset from the SPECviewperf 13 graphics performance benchmark. It was created from traces of the graphics workload generated by the Maya 2017 application from Autodesk.

3ds Max

This is the 3dsmax-06 viewset from the SPECviewperf 13 graphics performance benchmark. It was created from traces of the graphics workload generated by 3ds Max 2016 using the default Nitrous DX11 driver.


Classroom image from the Blender workload

Product Development

Rodinia (CFD)

Rodinia is a benchmark suite developed by Professor Kevin Skadron, et al, at the University of Virginia, to benchmark heterogeneous platforms. The CFD solver is an unstructured-grid, finite-volume solver for 3D Euler equations representing compressible flow.

WPCcfd

This workload is based on the OpenFOAM open-source CFD solver. The OpenFOAM combustion solver XiFoam is used to solve compressible premixed/partially-premixed combustion with turbulence modeling.

CalculiX

This workload is based on the open-source finite element program Calculix (www.calculix.de). It models the internal temperature of a jet engine turbine.

Catia

This is the catia-05 viewset from SPECviewperf 13 for testing graphics performance. It was created from traces of the graphics workload generated by the CATIA V6 R2012 application from Dassault Systémes. Model sizes range from 5.1 to 21 million vertices.

Creo

This is the creo-02 viewset from the SPECviewperf 13 graphics performance benchmark. It was created from traces of the graphics workload generated by the Creo 3 and Creo 4 applications from PTC. Model sizes range from 20 to 48 million vertices.

NX

This is the snx-03 viewset from the SPECviewperf 13 graphics performance benchmark. It was created from traces of the graphics workload generated by the NX 8.0 application from Siemens PLM. Model sizes range from 7.15 to 8.45 million vertices.

Solidworks

This is the sw-04 viewset from the SPECviewperf 13 graphics performance benchmark. It was created from traces of Dassault Systémes’ SolidWorks 2013 SP1 application. Models used in the viewset range in size from 2.1 to 21 million vertices.

Showcase

This is the showcase-02 viewset from the SPECviewperf 13 graphics performance benchmark. It was created from traces of Autodesk’s Showcase 2013 application. The model used in the viewset contains 8 million vertices.


Shaded model from the catia-05 viewset

Life Sciences

LAMMPS

LAMMPS is a molecular dynamics simulator developed at Sandia National Labs (http://lammps.sandia.gov/bench.html). The workload consists of five tests that simulate a variety of molecular properties. LAMMPS is run in parallel using MPI in the SPECworkstation benchmark.

NAMD

NAMD is a scalable molecular dynamics simulator 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 (http://www.ks.uiuc.edu/Research/namd/). The workload consists of three tests that simulate a variety of molecular interactions.

Rodinia (heartwall, lavamd, hotspot, srad)

Rodinia is a benchmark suite developed by Professor Kevin Skadron, et al, at the University of Virginia, to benchmark heterogeneous platforms. It contains the following tests:

  • Heartwall: A medical imaging process used to track movements of a set of sample points in the image of a beating heart.
  • Lavamd: The code calculates particle potential and relocation due to mutual forces between particles within a large 3D space.
  • Hotspot: A thermal simulation tool to estimate processor temperature.
  • SRAD: Speckle Reducing Anisotropic Diffusion, an algorithm to remove speckles in an image without sacrificing features.

Medical

This is the medical-02 viewset from the SPECviewperf 13 graphics performance benchmark. It uses the Tuvok rendering core of the ImageVis3D (http://www.sci.utah.edu/software/imagevis3d.html) volume visualization program to render a 2D projection of a 3D volumetric grid. A typical 3D grid in this viewset is a group of 3D slices acquired by a scanner (such as CT or MRI).


Scan image from medical-02 viewset

Financial Services

This section uses three popular algorithms from the financial services industry.

  • Monte Carlo simulation, or probability simulation, is a technique used to understand the impact of risk and uncertainty in financial, project management, cost, and other forecasting models.
  • Black-Scholes is a pricing model used to determine the fair price or theoretical value for a call or a put option based on variables such as volatility, type of option, underlying stock price, time, strike price, and risk-free rate.
  • Binomial options pricing model is a generalized numerical method for valuation of options. It uses a latticed-based model of the varying price over time of the underlying financial instrument.

Energy (oil and gas)

FFTW

This workload computes the discrete Fourier transform for 1D, 2D and 3D transforms. The 1D transform uses a 16MB vector, the 2D transform uses a 4Kx4K matrix, and the 3D transform uses a 256x256x256 matrix.

Convolution

This workload computes a convolution with a random filter of 100x100 pixels on an image that is 20,000x20,000 pixels.

SRMP

This workload addresses seismic data processing. It implements a Surface-Related Multiples Prediction (SRMP) algorithm written by Evgeny Kurin of GEOLAB Ltd. to remove multiples from seismic data.

Kirchhoff Migration

This workload uses a Kirchhoff equation form of the wave equation to calculate the back propagation of the seismic wavefield.

Poisson

This workload solves the Poisson equation using OpenMP for parallel execution.

Energy

This is the energy-02 viewset from the SPECviewperf 13 graphics performance benchmark. It is based on rendering techniques used by the open-source OpendTect seismic visualization application.


Topographical model from the energy-02 viewset

General Operations

This section benchmarks several applications that are in widespread use in the workstation market.

7zip

The freeware 7zip program offers good scaling performance between processors to compress and extract a hierarchy of files.

Python

This workload uses Python 3.6 to benchmark a variety of math operations, including those using the numpy and scipy math libraries, and the make charts and multithreaded matrix functionalities.

Octave

This workload uses Octave, a programming language for scientific computing, to compute a variety of math operations.

Storage

The storage workload is based on storage transaction traces from a wide variety of professional applications engaged in real work. These are grouped according to market segments for scoring purposes.

GPU Compute

LuxRender

LuxRender uses LuxMark, a benchmark based on the new LuxCore physically based renderer, to render a chrome sphere resting on a grid of numbers in a beach scene.

Caffe

Caffe is a deep-learning framework developed by Berkeley AI Research (BAIR) and by community contributors. Yangqing Jia created the project during his PhD studies at UC Berkeley.

Folding@Home

Folding@home (FAH or F@h) is a distributed computing project for disease research that simulates protein folding, computational drug design, and other types of molecular dynamics. 


Scene from the LuxRender workload