With any system you will want to see a combination of synthetic testing and real-world. Synthetics give you a static, easily repeatable testing method that can be compared across multiple platforms. For our synthetic tests we use Everest Ultimate, Sisoft Sandra, Futuremark's 3DMark Vantage and PCMark Vantage, Cinebench as well as HyperPi. Each of these covers a different aspect of performance or a different angle of a certain type of performance.
Memory is a big part of current system performance. In most systems slow or flakey memory performance will impact almost every type of application you run. To test memory we use a combination of Sisoft Sandra, Everest and HyperPi 0.99.
Unfortunately at some point in time AMD lost the memory bandwidth war to Intel. This is very evident by the numbers we have seen on both 890GX boards we have tested so far.
I hope that we do see them come back in this area, though, as with the limited cache on the AMD CPUs this is a serious performance hindrance.
Everest Ultimate is a suite of tests and utilities that can be used for system diagnostics and testing. For our purposes here we use their memory bandwidth test and see what the theoretical performance is.
Everest backs up our observations from Sandra.
HyperPi is a front end for SuperPi that allows for multiple concurrent instances of SuperPi to be run on each core recognized by the system. It is very dependent on CPU to memory to HDD speed. The faster these components, the faster it is able to figure out the number Pi to the selected length.
For our testing we use the 32M run. This means that each of the four physical and four logical cores for the i7 and the four physical cores of the i5 is trying to calculate the number Pi out to 32 million decimal places. Each "run" is a comparative to ensure accuracy and any stability or performance issues in the loop mentioned above will cause errors in calculation.
Here we find an interesting combination of problems. The HDD speed issue (about 15-20MB/s slower than on an ICH10) combined with lower memory bandwidth brings our HyperPi times to a crawl. This is not a problem limited to the GIGABYTE board, but is something we have seen across the board with AMD CPUs. But there is hope as AMD works to correct some of the issues in the current design.