The GPGPU Benchmark delves into the world of General-Purpose computing on Graphics Processing Units (GPUs). This panel provides a suite of OpenCL benchmarks designed to assess your GPU's performance in tasks beyond just graphics rendering. OpenCL is a powerful language that allows programmers to leverage the parallel processing power of GPUs for various applications – from video editing and scientific computing to even some games.
Multi-GPU Support
AIDA64 caters to a wide range of systems. The GPGPU Benchmark can handle up to 16 GPUs simultaneously, including AMD, Intel, and NVIDIA models. This makes it ideal for benchmarking multi-GPU setups like CrossFire and SLI configurations, as well as systems with both dedicated GPUs (dGPUs) and integrated graphics (APUs). Currently, there is only preliminary support for HSA configurations. Basically, any computing device that is listed as a GPU among the OpenCL devices will be benchmarked.
Benchmark Flexibility
The GPGPU Benchmark offers flexibility in choosing which components to test. By default, it focuses on GPU performance. However, you can also include CPU measurements for comparison purposes. It's important to note that the CPU benchmarks use native machine code optimized (utilizing available instruction set extensions such as SSE, AVX, AVX2, FMA and XOP) for your specific CPU architecture, while the GPU benchmarks rely on OpenCL, a more general-purpose approach. The CPU benchmarks are heavily multi-threaded, and are optimized for each CPU architecture introduced since the first Pentium.
Optimized for Accuracy
AIDA64 prioritizes accurate results. The OpenCL kernels used in the benchmarks are compiled in real-time using your GPU's latest drivers. This ensures the benchmarks leverage the full potential of your hardware. To achieve this, AIDA64 utilizes specific compiler options like "-cl-fast-relaxed-math" and "-cl-mad-enable" for optimal performance.
By utilizing the GPGPU Benchmark effectively, you can gain valuable insights into your system's graphics processing capabilities and its suitability for various GPGPU workloads.
