160GB (X25-M G2) Short Stroke Testing

We’ve seen various review sites take a larger 2 or 3TB drive and ‘short stroke’ it to match the capacity of the VelociRaptor. The idea is that you can use only the higher throughput portion of the larger drive, coupled with the fact that you are using less swept area of the heads (and therefore lower seek times), you coan get really close to – if not beat a smaller faster disk with a larger slower one. This generally does (did) work out for most performance specs, the big exception being that rotational latency of the slower disk would make it impossible to beat the faster one on seek times. Clearly short stroking can’t break the laws of physics, but it does have its uses.

Here I did a different sort of short stroke test. Instead of trying to make a larger drive act like a VelociRaptor (a losing proposition right off as the VR has *much* higher sequential throughput than any current competition), I decided to short stroke all models. Yes, even the VelociRaptors. The goal was to see how *all* HDDs perform when using a fixed capacity point. For that point I will be doing out Iometer test but limiting the sector range to the exact range of the Intel X25-M G2 160GB SSD. 160GB makes for a nice round number that many might use for their primary partition should they have their HDD storage duties split between OS and Mass Storage duties. Having the primary partition as a fraction of the whole drive is a good way to reduce seek times and increase overall performance. Here’s how they all panned out:

When short stroked, the RE4 did amazingly well despite it only being a 7200 RPM drive. What we did not expect to see was the 600GB VelociRaptor climb so much higher – even surpassing the new 1TB model. This may be down to tuning issues, and that the 1TB model may be optimized for full-drive access, but thats only a best guess.

« PreviousNext »