Kurt Shintaku's Blog

I have been searching for ages for some benchmarks on flash memory to see which cards perform the best under specific configurations.  Ultimately, all I want is someone to tell me:

• The precise configuration of their PC hardware
• The exact type/speed/brand of flash memory they used
• The method by which they connected the memory to their PC
• The software used to benchmark the memory’s performance

It would seem that no one – not even the digital photography industry – appears to have done any recent tests on flash card performance for an assortment of flash cards and I just don’t understand that.  I would think digital photographers would be all over flash memory performance to maximize their ability to take photos very rapidly. 

NOTE:  Speed is very important to me.  Why?  Because flash to me as a laptop user is the ultimate backup store.  You can keep it in your computer and always have high speed backup storage without attaching another hard drive.  Or better yet, you can use the card as an overflow drive for digital content like movies and stuff that you don’t access very often.  Or even better yet, you can use it as a ReadyBoost drive under Windows Vista.  Or maybe all 3.

The only article that I’ve been able to find on the topic is from Digital Photography NOW.com:
http://dpnow.com/4445.html

It’s a phenomenal article (if you ignore the inanities of a jerk named StuartB in the comments) that explains the nuances of top performing flash cards and all the factors that go into “judging” the performance of a flash storage medium including:

• What does ‘100x’ really mean?
  This is the data transfer performance of the memory relative to a CD music player which is .15MB/sec.  So a 100x card should be able to transfer 15MB/sec.
• Impact of the data path on accessing Flash storage
  Everything from the flash adapter to the connectivity (USB, Firewire) to the bus (PCI Express) impacts performance.  Just because a card is rated 133x doesn’t mean you’ll get 133x speed.  For example, USB 2.0 erodes access performance of flash due to the introduction of protocol latency.  A 133x card performs at 17MB/second on a high-end PC, which equates to only 113x.
• Dimensions in testing: read/write, file size, etc.
  Rarely is write performance the same speed as read performance.  And additionally, most flash memory is accessed faster initially compared to later in the read due to the fact that flash doesn’t have a “seek” time relative to disk drives.  So a 1K read might be extremely fast compared to a 100k read which might be just moderately fast.

In any case, I’m going to go purchasing some of the fastest memory I can find and see how they perform as extra storage mediums in my laptop.  I’ll post the results.