Now Reading
13 Flash Memory Card Roundup: Does High Speed Really Matter?

13 Flash Memory Card Roundup: Does High Speed Really Matter?

Sam ChenDecember 19, 2012

13 Flash Memory Cards Tested

memory-card-roundup-2012-custom-pc-review-5A couple weeks ago, I was talking with a couple colleagues about DSLR cameras, and invariably the topic about high speed SD and CF memory cards came up. What we wanted to know is if a high speed memory card such as the Lexar Professional 1000x CF (Compact Flash) card will actually perform better than a SanDisk Extreme 400x CF card under typical real world conditions. When looking at the price difference between the two cards, the SanDisk Extreme 400X 32GB CF card runs for about $88 while the Lexar Professional 1000x 32GB CF card runs for about $135, that’s a difference of nearly $47!

Since testing only two memory cards seems almost wasteful, I decided to test a couple more memory cards in today’s roundup, so today we’ll be checking out the performance from 13 different memory cards from a number of different popular manufacturers such as SanDisk, Kingston, Lexar, Patriot, and ADATA to see what kind of performance each memory card offers and what can be expected from some synthetic and real world testing.

Speeds, Classes, and Ratings

Before we dive into our roundup, let’s first talk a little about speed class ratings on both SD and CF cards. With SD cards, manufacturers generally rate the cards in three different ways.

  • The first method is through the use of class ratings. For example, most SD cards will be rated at Class 2, Class 4, Class 6 or Class 10. Each of these class ratings also translate into minimum transfer speed ratings whereas a Class 2 card has a minimum of 2 MB/s transfer rate, a Class 10 card has a minimum of 10 MB/s transfer rate.
  • The second method is a newer method of the class ratings called UHS (Ultra High Speed). UHS-I cards are much, much faster and can support up to 104 MB/s.
  • The third method is through the use of “x” ratings. For most computer enthusiasts, this should be quite familiar since it’s just a multiple of 150kB/s, the speed of the standard CD-ROM drive (Oh yeah, that old thing…). Therefore, an SD card with a 133x speed rating should be able to achieve a transfer rate of 19,950 kBytes/s, which is about 20 MB/s. High speed SD cards will sometimes use the x rating as class ratings max out at Class 10, which is a paltry 10 MB/s minimum.

Moving onto CF cards, the rating system is much simpler. CF cards generally just use the “x” system, which is the same “x” system used on CD/DVD drives and SD cards. Some CF cards will also use UDMA (Ultra Direct Memory Access) rating, which is actually an interface used in computers prior to the introduction of SATA. UDMA currently ranges from 0 to 7 with UDMA 7 currently topping out at 167 MB/s.

Before we delve any further, let’s take a look at the different SD and CF memory cards we’ll be testing today.


TypeSpeed RatingCapacityPurchase Link
ADATASDClass 616GBClick Here
Patriot LXSDClass 1016GBClick Here
TranscendSDClass 1016GBClick Here
KingstonSDClass 1064GBClick Here
Kingston Ultimate XXSDClass 10/UHS-I32GBClick Here
SanDisk ExtremeSDClass 10/UHS-I32GBClick Here
SanDisk Extreme ProSDClass 10/UHS-I16GBClick Here
Patriot EPSDClass 10/UHS-I32GBClick Here
Patriot EP ProSDClass 10/UHS-I32GBClick Here
SanDisk ExtremeCF400x/UDMA 532GBClick Here
Kingston UltimateCF600x/UDMA 632GBClick Here
SanDisk Extreme ProCF600x/UDMA 632GBClick Here
Lexar ProfessionalCF1000x/UDMA 716GBClick Here
About The Author
Sam Chen
Hardware and Technology Enthusiast. SSD Evangelist. Editor-in-Chief. You can find Sam's full biography here
  • Russell Robinson

    This is a pretty good comparison, some people still think all sd cards are equal

    • Brett A. Wheeler

      All SD cards are equal, although some are more equal than others.

  • Greg Zeng

    In my experiences, I’ve found flash card performance varies between individual items which seem similar: brand, model, size, speed, etc: batch & random variations.

    Speed often lessens, depending on hubs, expansion cables, and even which hardware port on my PC is used. It’d be interesting to see if bench testing within similar batch are consistent too.