Review-AMD “Llano” A6-3650 and A8-3850 Performance Assessment
Test Results
Blender 2.49b
Blender is an open-source 3D-modelling software. Performance is measured through the amount of time needed to render a 3D object. Lower figures represent better results.
The Core i3-2100 leads the pack in this 3D-modelling application. It’s around 3% faster than the Phenom II X4 840. Llano A8-3850 comes in at third place. Phenom II X4 840 has the advantage of higher clockspeed. When lowered down to the same frequency as A8-3850, however, it’s actually slower by around 4% compared to AMD’s latest APU.
Cinebench R11.5
Cinebench benchmark calculates hardware performance in rendering a 3D scene. The final results are measured in points. Higher points mean better performance.
Again, Phenom II X4 840 surpasses the newer A8-A3850 APU, this time by around 6%. It must be noted that the Phenom holds the upper hand due to its higher clockspeed. Without that advantage, Phenom II X4 840 is actually a tad slower than the A8-A3850.
Excel 2010 – Montecarlo
Surely you are well familiar with this popular spreadsheet application from Microsoft. Here, we see how fast each processor could finish the “Monte Carlo Simulation Table” in Excel 2010. The final scores are measured in seconds. Smaller figures indicate better results.
When it comes to complex Montecarlo simulation, both the A8-3850 and A6-A3650, and the Phenom II X4 840 are no match for Intel’s Core i3-2100. The 40% performance gap between the Core i3 and Llano A8-3850 is due to the lack of L3 cache on the latter processor. Phenom II X4 840 lags behind both Llanos because it has smaller L2 cache.
7-Zip
Another file compression software that we use as a benchmarking tool is the 7-Zip (version 9.20). This particular application is able to utilize all available threads, making it ideal to test the performance of multi-core processors. The same set of files as used in the Winzip test are also employed in this test. For our testing purposes, the file compression is carried out under three different scenarios: utilizing 1, 2, and four threads. Results are measured in the amount of time needed to finish the process. Lower figures are better.
Core i3-2100 has a significant lead over the AMD processors until the software is configured to utilize 4 threads, in which case one of the Llanos take over the top position. However, the performance difference is very small.
FastStone Photo Resizer
FastStone PhotoResizer is an image converter/ resize software capable of applying image manipulations such as resizing, cropping, rotating, and others on multiple files through batch operation. For the purpose of this test, we convert a collection of TIFF images (66 files; 527 MB total size) to JPEG format (quality level: 90). Each picture is also resized to 600 pixels on the longest side. The entire process only utilizes one processor core (thread). Therefore, processors with high clockspeeds would have a slight advantage over slower CPUs. Results are measured in the amount of time needed to finish the resizing and converting process. Lower scores are better.
In this single-threading application, Core i3-2100 manages to outperform the A8-A3850, although not by much (6%). It seems like Sandy Bridge’s per-core performance is still beyond Llano’s.