1 – GpuTest 0.7.0 Downloads
Windows 64-bit (XP, Vista, 7, 8, 10)
Mac OS X 10.7, 10.8, 10.9, 10.10
Linux 64-bit (Ubuntu-based, openSUSE)
2 – GpuTest 0.7.0 Release Highlights
OpenCL lets you tap into the parallel computing power of modern GPUs and multicore CPUs to accelerate compute-intensive tasks in your Mac apps.Use OpenCL to incorporate advanced numerical and data analytics features, perform cutting-edge image and media processing, and deliver accurate physics and AI simulation in games. @jsteube I've encountered this issue as well on macOS 10.14 Mojave using hash mode 11300 and benchmark reports issues with 1700, 2500 (that one gets cl build program failure), 1800, 7100 and 11300. The common factor in most of those seems to be SHA-512. If this isn't something that can be fixed in code is the solution just using-self-test-disable.
This update of GpuTest is available with two new features. The first one is a new OpenGL benchmark that draws the popular Julia fractal in a GLSL pixel shader. The interesting thing is that the Julia fractal is available in two flavors: FP32 and FP64. The Julia FP32 (fp32 = single precision floating point) test should run on any OpenGL 2 or OpenGL 3 capable hardware. The Julia FP64 is limited to OpenGL 4.0 capable hardware and draws the Julia fractal using FP64 (fp64 = double precision floating point) variables. Some scores are available in this article.
- Hello All, I don't know if this is the right place to ask. In my project (OSX) I have to execute a lot of FFTs (1-dimensional, radix-2): up to 60 times per second.
- CompuBench measures the compute performance of your OpenCL and CUDA device.
On NVIDIA GPUs, the Julia FP64 test runs perfectly on all platforms (Windows, Linux and OS X 10.9). On AMD GPUs, the Julia FP64 test runs fine on Windows, has some trouble on Linux (tested with FGLRX drivers on Linux Mint 16), and does not work on OS X…
I didn’t tested the FP64 demo on Intel GPUs.
GpuTest 0.7.0, Julia FP64 OpenGL 4.0 test – GeForce GT 650M – Mac OS X 10.9
The second feature is the possibility to submit scores to an online GPU database. This database stores all possible (platform-dependent) GPU data (device IDs, clock speeds, temperature, drivers, etc.) as well as OpenGL and OpenCL information (extensions, version, etc.). This GPU database is also used by the latest GPU Caps Viewer.
Ranking / comparative tables will be available shortly!
GpuTest 0.7.0 comes with new textures to change a bit the look of the other 3D tests:
GpuTest 0.7.0, FurMark test
GpuTest 0.7.0, TessMark X32 test
GpuTest 0.7.0, GiMark test
For Linux users, an Python/Tkinter interface is available. To launch it, justr open a terminal in GpuTest folder and type:
Depending on the state of your distro, maybe you will need to install the python-tk package:
Do not hesitate to improve this interface and send me your work. I will add it in GpuTest (with your credits!).
GpuTest 0.7.0, Linux Python/tkinter GUI
In the Windows and Linux versions, I added the temperature graph when GPU core temperature sensor is available. The temperature graph can be displayed with the OSI (On Screen Info) checkbox or via the command line:
The temp graph is not available in benchmark mode for performance reasons.
Does Android Support OpenCL?
Update (2014.04.25): for Linux users, here is an alternative front-end (gtk+/glade/python):
The front-end require GTK+ 3.10.
Opencl Test For Mac Os
A forum is available HERE for feedbacks or bug-reports.
3 – GpuTest 0.7.0 Changelog
- added online score database.
- added temperature graph (OSI) when available.
- added new Julia FP64 fractal test (OpenGL 4.0).
- added new Julia FP32 fractal test (OpenGL 2.1 / 3.0).
- updated with latest 3d framework.