From The Front End - ZDNet Edition
My ZDNet version of my blog. Will feature news my mad opinions and relevant posts or partial quotes from my blog - http://www.fromthefrontend.co.uk
Topics: Wed Standards, Tools and Services for Designers, Photography and graphics, Accessibility and other Front-End Webdesign related info.
Friday 15 February 2008, 12:30 PM
ReadyBoost or ReadyDrag
The problem appeared to be with (1) Their system setup (2) The tasks they were performing and (3) the quality of the memory stick/card being used.
Out of these I've found the most significant factor to be the memory stick. I tested 6 different memory sticks/cards on my laptop and the difference in how effective ReadyBoost is huge.

Starting with the bottom of the scale I had an old 2GB SD memory card. I inserted this into the built-in memory card reader and found a major reduction in system performance. Opening Apps or even switching between Apps would often cause lock-ups and pauses. At the time I had no idea what was causing the problem and began to think my system was under performing. I remove the memory card and things return to normal. The speed of the memory card was far too slow for ReadyBoost causing delays while data was read/writing.
I next used a 512mb SD card with "x66" boldly displayed across the front. This didn't have the major performance dips caused by the first memory card but I also couldn't say I noticed a performance boost. After doing a little research I found that memory cards are designed for sequential writes like storing photos as they are taken on a digital camera. Hard disk drives sequential read and write speed is significantly faster than anything you will get on a memory card and so using a memory card provides no real benefit.
ReadyBoost relies on memory sticks faster random access speed. The hard disk has to physically move parts to access data across different parts of the harddrive. The Solid State memory on a memory stick has a faster response time, doesn't have to spin up from idle and accesses random data significantly faster. So would the memory sticks fair better.
This time I got 2 ReadyBoost logo'd memory sticks for about £17-20 each. The first was a 2GB one from Argos from a brand I hadn't heard of or remember. The performance increase was noticeable especially in memory hogging programs like Photoshop and games. The 2nd stick was a 4GB Sandisk cruzer u3. There was no notable performance improvement. What was the difference?
I looked up what was the requirements for a device to be classed as ReadyBoost capable. It was a read speed of 5mb/s and a write speed of just 3mb/s. The sandisk must have been pretty close to the bottom end as every now and then when I'd insert it I wouldn't get the option to use it "to speed up my system" (not you have to diable ReadyBoost or delete the ReadyBoost file and reinsert the stick for the option to show up otherwise it will automatically use the stick for ReadyBoost). When you compare this to some of the better memory sticks like Corsair's GT range that read at 34mb/s and write at 28mb/s you can begin to see why there is such a difference in how effective ReadyBoost is. I lost the former 2 memory sticks and so I'm now on to a Corsair Flash Voyager. Not the super-fast GT version but still a lot faster than the average memory stick. It isn't even ReadyBoost logo'd but according to the specs it should be a lot better than the sandisk cruzer I was using. If you live in the US Kingston do a memory stick called HyperX that is exclusive to Amazon.com and is super-fast and good value.
The good news is that the speeds and capacities of these memory sticks is improving all the time so ReadyBoost will continue to be more and more effective.
The bad news is that 1) the ReadyBoost logo is given out too easy and consumers could end up with inferior memory sticks that don't help them 2) with Solid State Hard drives getting closer to usable capacities will we even need ReadyBoost?
For now check out some of the fastest USB flash drives here:
everythingusb.com
Comments on this post
Thanks for a good article, David.
Did the testing you describe happen before or after you installed Vista SP1, or both? I am curious, because my latest Fujitsu laptop came with 1 Gb of "Intel Turbo Memory" pre-installed; I didn't notice any speed improvement while I was using it with Vista, and I haven't notice any speed loss since I went back to XP Professional (and thus can't use the Turbo Memory).
Now I have to go digging around to try to find the read/write specs of the memory, and I've put this at the top of the list of things to test when I put the Vista disk back in and upgrade to SP1.
Thanks again,
jw
All tests were conducted before SP1 and some were repeated post SP1 install. Results pretty much the same.
I've started seeing SSD in ExpressCard format with capacitys from 4GB to 32GB which are also touted as suitable for ReadyBoost. I was considering getting one and installing apps (that would benefit from faster response times and random access) directly onto it and using the rest of the space as ReadyBoost. However, price is still incredibly high.
A 32GB Transcend card is £184 on amazon.co.uk
A 16GB Lexar version costs £143
When I compare this to the £15 it cost me for my 4GB Corsair USB stick I think I'll wait for a price drop. The faster speed of an express card slot compared to USB2.0 will make installing apps on it more realistic than on a USB stick (despite what portableApps and U3 may suggest) but the bottleneck will still be the read/write time of the flash memory which Transcend quote at:
Read up to 18MB/s, Write up to 8MB/s
Pretty much the same as the USB stick I have.


