Oh, well, this tip is just to enable ReadyBoost for flash drives that don't pass the test. Frankly, ReadyBoost is a bit odd. Even fast USB flash drives can fail the test. I really have no idea what performance criteria Microsoft is basing its tests on.
The thing is, the restriction was put so that only flash drives which passed the test would provide a more noticeable 'boost' compared to flash drives that fail the test and if enabled, probably wouldn't give such a noticeable 'boost' Adrian, the fast they're referring to are random reads and writes, not the sequential reads or writes we usually consider while transferring large files(e.g. photos from digital cameras) http://blogs.msdn.com/tomarcher/archive/2006/06/02/615199.aspx I'm not sure whether this article explains it or not, but I found it to be very useful to see which flash drive provides the best ReadyBoost performance(well in the sense that it's random reads and writes are faster than the competition). http://www.xbitlabs.com/articles/memory/display/4gb-usbflash-roundup.html http://www.xbitlabs.com/articles/memory/display/12usbflash-roundup.html http://www.xbitlabs.com/articles/memory/display/10usbflash-roundup.html (and no I don't own a PC with vista, so I've no idea what noticeable 'boost' ReadyBoost gives)
Actually, flash memory, by design, is always fast at random reads and writes. Why? There's no mechanical read/write head to move around. Random access in flash drives is virtually instant, which makes it more like RAM, rather than a hard drive. Every cell in the flash memory requires the same small latency to access.
Adrian, that's the latency, it's true that the latency for a flash drive is very very small, but if you look at the random writes for this set of 4gb flash drives, you'll notice a huge variation between the random write speeds. yes, I do realise that it's still huge compared to a hard drive, but I do wonder whether you're actually able to notice the difference if the random reads and writes were below their set benchmark(on a typical new PC with 1gb ram and blah)(anyone willing to participate in a double blind test?) http://www.xbitlabs.com/articles/memory/display/4gb-usbflash-roundup_12.html
Well, the effect would depend on the system. The more RAM you have, the less effect ReadyBoost will have.
Just curious, I'm pretty sure that one RAM stick is much faster than one flash stick. A flash stick is only USB 2.0 and maybe the data transfer is much lesser...
Actually, I can see why they didn't notice any performance difference. With 2GB of RAM, the system will not really use much virtual memory, unless they test memory-guzzling games like CoH. TBH, the Photoshop loading tests isn't going to show anything. ReadyBoost is not a hard drive cache. Photoshop will not load faster in subsequent loadings with ReadyBoost. If they want to really find out what ReadyBoost can do, strip the system down to 1GB or better, 512MB. Then run a benchmark suite like Winstone. Still, it would be hard to truly gauge the performance improvement from ReadyBoost. IMHO, memory is cheap now. Buy REAL memory. If I buy a flash drive, I want to use it for flash storage, not dangling off my LCD or the case trying to be what it isn't - fast RAM.