On Thursday 05 October 2006 08:55, Vitaly Wool wrote:
> Here the basic performance measurements for JFFS2 and YAFFS2
> as of 5/10/2006 are presented. The measurements were performed
> on the PNX4008 platform which is a 208 MHz CPU ARM926-based
> Philips board. The 2k page flash is the real NAND flash (20MiB
> partition); the 512b flash is nandsim (8 MB partition).
What is the performance of the underlying flash i/o -- it varies
wildly based on platform, hook-up, bus timing, NAND chip,
mtd-interface (so called 'mapping' driver). Without figures for
base NAND i/o (page write, page read, oob-read, block erase, etc)
higher level figures for filesystem performance are unhelpful.
Because JFFS compresses data and YAFFS does not, this also has to
be factored into any figures. How does the test data compress
etc., who much real flash i/o is (was) performed?
The fact that your numbers show that Yaffs is faster in the
nandsim test, indicates to me that the underlying performance of
the real NAND device and hook-up is a large factor in any
measured performance. nandsim really doesn't simulate NAND.
-imcd
(text/plain)