52 128 GB & 256 GB SSDs tested and compared

Here you can find out which SSD is the best for you

By ,


Atto Disk Benchmark

The Atto Disk Benchmark reads and writes data blocks of different sizes as quickly as possible to random places on the SSD. We used the 4kB and 1MB tests. The 4kB tests reflect working with very small files, resembling what happens when starting up software. The 1MB test are a best-case scenario for SSDs and indicate what type of performance you can expect when working with large files, such as copying digital photos, MP3s, and films.

Four commands are simultaneously sent to the disks each time. Since the flash chips in SSDs are parallel controlled, smart controllers can execute multiple commands at once. This improves the performance. In real-life, Windows and other software will also simultaneously send multiple commands to the drives.

Please note: Atto Disk Benchmark uses compressible data. Many files - especially program files - are like this. However, files such as digital photos, MP3s, or films are barely compressible or not compressible at all. The AS SSD benchmark on the next page provides a clearer picture.

The 1MB read and write tests give a good indication of the performance of an SSD. In the 1MB read test most SSDs end up around  550 MB/s, the limit for Serial ATA 600. The OCZ Petrol and the Crucial v4s are exceptions here due to their SATA300 bus. The Agility 4 falls a bit short here as well. For the 1MB write tests all SandForce SSDs end up at around 530 MB/s, but that's because Atto's data is compressible. The fact that the Samsung 840 Pro 256GB and OCZ Vector 256GB achieve the same performance without compression tricks is impressive. A number of SSDs don't write data as fast, such as the Crucial m4 and v4, OCZ Petrol and Samsung 840. 

In the 4k read tests the Samsung 840, Samsung 840 Pro and OCZ Vector score the best. In the 4k write tests the 840 Pro and Vector also do well, but the scores are closer to each other.

4k read

4k write

1MB read

1MB write


The Hardware.Info website uses cookies. Read our cookie policy.