On Sun, Mar 13, 2011 at 9:54 AM, Arnd Bergmann <arnd@xxxxxxxx> wrote: > On Sunday 13 March 2011 14:00:04 Andrei Warkentin wrote: >> > If the latter two are not much different on the toshiba card, that >> > would be a simpler implementation, and more likely to be useful on >> > other cards. >> > >> >> Revalidating the data now, along with some more tests, to get a better >> picture. It seems the more data I get, the less it makes sense :(. > > I was already fearing that the change would only benefit low-level > benchmarks. It certainly helps writing small chunks to the buffer > that is meant for FAT32 directories, but at some point, the card > will have to write back the entire logical erase block, so you > might not be able to gain much in real-world workloads. > Original data was collected by someone else on a 16G eMMC device. The data I have collected now for a 32G device of seemingly similar technology doesn't quite match the original data in terms of performance characteristics, so I am trying to collect the data on the 16G device and figure out what the difference is. The align change may only benefit certain types of accesses, but it for sure must not hurt the general case, otherwise there is really no point behind even considering it... It takes forever to get a good amount of data to see past the noise, too. My last set of tests has already been running for 10 hours... > Of course, I also have only a very limited understanding of the > algorithm used in the Toshiba chip. > Unfortunately, so do I. As far as manufacturer suggestions, we get a very controlled and limited view of things :). Thanks, A -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
