[SATLUG] Benchmarking SSD Drives
bruce.dubbs at gmail.com
Tue Jan 31 16:40:24 CST 2012
I had a friend loan me an SSD device. It is a card that mounts in a PCI
slot, but does not actually use the PCI connector. The card has a power
connector and a SATA connector.
I'm trying to check it out. It formatted fine (ext3 for now).
The card inserted into this device is a SanDisk SDCFH-008G.
Testing with hdparm:
hdparm -tT /dev/sdc
Timing cached reads: 14842 MB in 2.00 seconds = 7430.48 MB/sec
Timing buffered disk reads: 130 MB in 3.01 seconds = 43.16 MB/sec
Timing my conventional drives:
hdparm -tT /dev/sda
Timing cached reads: 14796 MB in 2.00 seconds = 7407.85 MB/sec
Timing buffered disk reads: 332 MB in 3.01 seconds = 110.32 MB/sec
hdparm -tT /dev/sdb
Timing cached reads: 14984 MB in 2.00 seconds = 7502.19 MB/sec
Timing buffered disk reads: 330 MB in 3.01 seconds = 109.70 MB/sec
So the cached reads are all basically the same. This seems reasonable
since the system is actually bypassing the drive.
On the other hand, the SSD device says 30GB on the card, but the
buffered reads are only half that of the standard drives.
OTOH, a copy of a 541M file gives:
time cp Fedora-16-x86_64-Live-LXDE.iso xxx
On SSD SanDisk SDCFH-008G : real 0m5.386s (100 MB/s)
On Seagate ST3750330AS : real 0m3.257s (166 MB/s)
Does anyone have any experience with these. It's mysterious to me that
the SSD is only about half the speed of the standard SATA drives. Are
there some configuration items that need to be set? BIOS? Kernel?
Or is this just the socket card or the SSD device itself?
More information about the SATLUG