Memory Scaling On The AMD Kaveri A10-7850K APU
Ryan Martin / 11 years ago
CPU Tests
Cinebench
Cinebench is a widely respected benchmark for testing the performance of x86 CPUs. The program allows you to test single and multi-threaded performance as well as GPU performance by rendering with Open GL. Download here.
Cinebench showed the kind of results that I was really expecting to see. In the CPU dominated single and multi-threaded tests we see almost identical results across all speeds, with the GPU-powered OpenGL test we see a huge boost in performance thanks to the increased bandwidth available to the GPU.
WPrime
wPrime is a leading multithreaded benchmark for x86 processors that tests your processor performance by calculating square roots with a recursive call of Newton’s method for estimating functions. Download here.
In WPrime there is a similar story – identical results within margin of error. In fact the fastest memory speed gave the slowest result suggesting the importance of timings in WPrime over pure memory bandwidth.
AIDA 64 Engineer
AIDA64 Engineer is a streamlined Windows diagnostic and benchmarking software for engineers. It has unique capabilities to assess the performance of the processor, system memory, and disk drives. AIDA64 is compatible with all current 32-bit and 64-bit Microsoft Windows operating systems, including Windows 8 and Windows Server 2012. AIDA64 implements a set of 64-bit benchmarks to measure how fast the computer performs various data processing tasks and mathematical calculations. Processor benchmarks utilize MMX, SSE, XOP, FMA, AVX and AVX2 instructions, and scale up to 32 processor threads. For legacy processors all benchmarks are available in 32-bit version as well. Download here.
On the CPU front a lot of AIDA64 benchmarks showed a small decline in performance, against suggesting that the slackened timings of the higher memory frequency cancelled out the benefits of more throughput. Overall the picture is one of minimal change with varying memory frequency.
The FPU benchmarks weren’t very different, again showing a broad insensitivity to changing memory frequencies.