Let’s look at just the CPU Mark 99 results using both CAS 2 and CAS 2.5 with 200MHz and 266MHz FSB, along with all the different BIOS timing settings:
CPU Mark 99 does more than just benchmark the CPU. It tests the CPU, all cache levels and the main memory. Since the CPU and caches have not changed, we know that any difference in the results will be due to the changes made in the memory settings and nothing else, a good useful tool to use.
A couple of things stand out, with the first being the ‘Ultra 2’ results, which are the same as the ‘Fail safe’ setting. I don’t know for sure (yet) but suspect it’s just a BIOS issue. It is not uncommon for an evaluation BIOS to have some features disabled or to not be fully optimized yet. Looking at the ‘Fast’ and ‘Ultra settings I would not expect the ‘Ultra 2’ setting to give a significant gain, but only time will tell for sure…
The other thing that caught my attention while looking at the graph is that the PC1600 results show a nice gradual gain from the lowest to the highest. But the PC2100 does not. The PC2100 takes a big jump when going to ‘Slow’ from ‘Fail Safe’ and then levels off with smaller gains from there. There is about a 4.5% difference between the highest and lowest scores using PC2100 and a 266MHz FSB, but a 6% when using PC1600 and a 200MHz FSB.
You can also see clearly that there just isn’t a big difference between CAS 2 and CAS 2.5. You can also see that when allowing the BIOS to use the SPD on the module we have pretty optimal performance (that will vary with different modules due to different spec’s on the SPD).
Be the first to discuss this article!