LuxMark 2.0, OpenCL Benchmark for GPUs and CPUs



LuxMark 2.0, OpenCL Benchmark for GPUs and CPUs


Around one year after version 1.0, the new version of LuxMark, an OpenCL benchmark based on LuxRender, is available. LuxRender is a physically based and unbiased rendering engine. Based on state of the art algorithms, LuxRender simulates the flow of light according to physical equations, thus producing realistic images of photographic quality.

OpenCL logo



This new version of LuxMark 2.0 comes with a new rendering engine with multi-platform OpenCL support, two new benchmarks and the ability to run the benchmarks on selected compute devices (GPUs and / or CPUs).

Quick test:

Scene: Luxball

LuxMark 2.0, OpenCL Benchmark for GPUs and CPUs
Scene: Luxball – Score: 7872 points (Radeon HD 6970)


LuxMark 2.0, OpenCL Benchmark for GPUs and CPUs
Scene: Luxball – Score: 3120 points (GeForce GTX 460 + GeForce GT 520)



Scene: Sala

LuxMark 2.0, OpenCL Benchmark for GPUs and CPUs
Scene: Sala – Score: 844 points (Radeon HD 6970)


LuxMark 2.0, OpenCL Benchmark for GPUs and CPUs
Scene: Sala – Score: 456 points (GeForce GTX 460 + GeForce GT 520)



Source: Geeks3D forum






10 thoughts on “LuxMark 2.0, OpenCL Benchmark for GPUs and CPUs”

  1. fellix

    Sadly, Nvidia’s OpenCL 1.1 ICD performance is still broken and Fermi lags behind in this bench.

  2. applejack

    GTX 570 @ stock :

    Luxball scene – 5526 points.
    Sala scene – 820 points.

    seems to catch up with hd6970 as scene complexity / triangles count raise.

  3. zhaopeng

    GTX570 on openSUSE 12.1
    Room: 370
    Sala: 780
    LuxBall: 5228

    GTX570+GTX260 on openSUSE 12.1
    Room: 492
    Sala: 1046
    LuxBall: 7231

    The LuxMark 2.0 has a memory problem on Linux. After running benchmark several times, it ate more than 5G memory included swap which almost made my machine crash. I guess it allocates resources for each benchmark and releases them only after closing application.

    Nvidia is not active on OpenCL. And they care more about CUDA and latest openACC. I worry a little about the future of OpenCL support from Nvidia.

  4. DrBalthar

    Boy these scala images look extremely noisy definitely not enough samples per pixel you probably need at least 8 times more samples on these.

    Guess that’s the problem with brute-force MC algorithm

  5. komar

    c2q 12Mo l2 q9550 @ 3.75ghz
    4go PC 8500 cas 4
    AMD 5970@900/1200 OC 1.187Vcore
    SCORE: )
    room ; 777

    sala : 1748

Comments are closed.