In session 2 we made a comparative measurement between two readers, the one we are testing and the Sitecom's Internal Media Dock (compatible with the USB 2.0 and 1.1 standards as well); we connected them to a host with an Intel Pentium4 cpu and USB 1.1 ports.

In this session we evaluated the reader's performances with older computers to test the chance of buying this reader for the users who have obsolete pcs and even already a memory card reader.

3.2 Results understanding

The tipical output of tools that measure performances is the Combinated Index: the measuring take a number of steps, depending on the flash card storing space, and for each of them we evaluated the number of operations at minute in processing a file (here 512b, 32kb, 256kb, 2Mb, 64Mb files for testing the SD card).

For each file we got the output and on these partial results we operated an arithmetic mean, this one is the combinated index and it measures the system global response. Usually an output result strongly depends on many different conditions, and often a measurer can't control them unless making a statistical analysis of the results. For this reason we ran every test five times and then we calculated the combinated indices making the mean on the partial results. We will show it by Excel's histograms, without indicating the relative or absolute uncertainty. The benchmark tool we chose for testing, the Sandra Lite, showed itself very precise and accurate.

3.3 Session 1 Results Analysis

Connecting the device to different hosts, what can we say about its performances? Are they constant or do they change?

To give an answer to this question the reader and the flash memory cards (Compact Flash and Secure Digital as well) are been tested with three latest notebooks.The output resulting measures pointed out an essential uniformity when the reader worked with the CF memory; the Intel platform notebooks (Asus A2005H e Aspire 1682 WLMi) has just a little advantage, but as it is smaller than 1% (figure 6), that is not significative.

In processing a 512 byte file, the three systems performed, on average, 30545 op/min in reading and 193 op/min in writing: this means a transfer rate ( kb/s) of 255kb/s e 1.6Kb/s respectively. With a 64Mb file we got 6 op/min (6554kb/s) in reading and 3 op/min (3277kb/s) in writing.

Even though we stressed the distances among output measures of Combinated indices (figure 7), the Secure Digital histogram shows that the performances are quite the same for the three computers, though the one based on AMD Sempron Cpu has just a little advantage on the others.

In the next page we will session 2 testing results.

Visual Inspection

1 2 3 4

Session 2 Results