System Benchmarks
LAME first, which is pure CPU. The timings probably could have been set to the worst possible allowed in the BIOS, and we'd still have produced the graph above. 2.2GHz of fast FPU'd goodness to give performance that could shame almost any shipping Intel system on the market today. SETI shows us two things. Firstly, the Time is configured well, and performs as well as my hand tuned system. Secondly, without the helping hand of better memory timings, performance certainly does drop off slightly. SETI is all about good memory subsystem performance, and while it's working well at DDR400 speeds, there's certainly some extra performance to be had by tweaking our 4 magic BIOS numbers. We see exactly the same thing with Pifast. Comparative performance to the other test rig, and a drop off when using the shipped timings. I know it's not too much of a big deal, but why ship Corsair in that case? It's certainly more expensive, even to a vendor with big pockets that can make big bulk discount orders, so why run it at the timings you could get out of much cheaper modules? |