Bonnie Results Interpretation

We compare tests run on different kernels, using one filesystem type (ext2) Tests run on one kernel can be used to compare filesystem performance. For each metric, we calculate the average score and the standard deviation for the five kernels. Then to scale the results, we calculate the ratio between the standard deviation and the average, expressed as percent ( (StdDev/Average) * 100 ) This comparison revels which test are sensitive to kernel changes. In the following charts, we have color coded the results and grouped them into three areas:

Results Comparison