Question Unusual SSD (SATA) read performance?

[tcsenter]

Running latest CrystalMark bench on my PC, using relatively low hour PNY CS1211 480GB SSD (SATA). System is fresh install, all drivers up to date. The random 4K Q1/T1 write speeds are a multiple greater than read?




I tested again after restart and letting the system sit for about five minutes after loading the desktop. Similar result. Is this suggest bad mojo with the flash memory or controller?

 

[Insert_Nickname]

Nothing weird in that. It's quite normal.

The controller can parse random writes faster then it can fetch a random block from the NAND. If it gets a write command, it can just write to a random free block. Finding, then selecting, a specific block, then reading the content is a bit slower.

I hope that makes sense.

 

[tcsenter]

Nothing weird in that. It's quite normal.

The controller can parse random writes faster then it can fetch a random block from the NAND. If it gets a write command, it can just write to a random free block. Finding, then selecting, a specific block, then reading the content is a bit slower.

OK thanks, I had never seen such a big disparity before.

 

[solidsnake1298]

Phoronix benchmarked this drive and found a similar behavior. How a drive behaves is affected by a lot more than you think. You have to take into consideration the controller, firmware, whether it has DRAM or not, the kind of NAND (SLC, MLC, etc.), how many NAND chips, how many channels each chip has.

PNY CS1211 120GB SSD Tests On Linux Review - Phoronix

www.phoronix.com