Benchmarks I
Performance is heavily reliant on just how much usable bandwidth you have at your disposal. Ideally, we'd be wanting to look at unbuffered performance, such that prefetching of data is disabled. That'll tell us just how good the memory controller is. Looking at SiSoft SANDRA's unbuffered benchmark.
We'd expect the double-pumped Barton to be languishing near the bottom. What we wouldn't expect is to see the Springdale eclipse the Caterwood's unbuffered results. Where's the much-vaunted PAT technology when you need it most ?. Almost 3GB/s of unbuffered goodness. The above result is contrary to what we'd expect. We do state it like it is, unexpected or not. Historically, a decent showing in the unbuffered benchmark has translated into excellent benchmarking performance.
Running Pifast v41 next. It simply calculates the constant Pi to the desired number of places. 10 million is chosen for this test. If you want to run it for yourself, click here for the benchmark standings and download link. Just unzip and click on the .bat file.
The astonishing unbuffered bandwidth of this ASUS P4P800 translates into the fastest, stock Pifast time on record. It seems as if I mentioning this fact in almost every new motherboard review. The top Pifast slot has been keenly contested of late. Pifast is all about usable bandwidth. It thrives on the ability to shift masses of data from the RAM to the MCH and then to the CPU. Its speed of shifting is necessitated by low latency access. The time of 63.95s highlights that aspect. PAT, who needs it ?.
Pretty much what we expected here.
Next is an activity similar in usage to Pifast but far longer in duration. SETI loves low latency bandwidth as much as any program. A .0417WU is used.
Another extremely impressing showing by the bandwidth monster. We're steadily closing in on the 2-hour barrier.
