Benchmarking and Burn In Testing SSD's

[MrPisky]

I have recently built a new system using an OCZ Solid3 120GB SSD as the Win7 OS drive. I have been reading some issues that pertain to OCZ drives (maybe most SSD's in general?) and the writing of non-compressible data during benchmarking/testing. It is my understanding that OCZ uses ATTO and IOMETER to obtain their own performance benchmarks and marketable drive stats.

So, my question is...are there any known issues or precautions in using Passmark Benchmarking or Burn In Test with SSD's in general or OCZ drives in particular?

Thanks!

[David (PassMark)]

None that we are aware of.

What we have observed is that many SSD do poorly when reading and writing small chunks of data. And as a result never get near their max claimed speeds.

But you can use the advanced disk test in PerformanceTest to try out various scenarios.

[MrPisky]

Yeah, there's a whole debate on the issue of compressible/non-compressible data used for the benchmarks. OCZ's Sandforce-based drives definitely handle things differently than their competitors. The Sandforce drives benefit greatly from highly-compressible data and actually "throttle down" to preserve nand lifespan when excessive non-compressible writes hit the drive from excessive benchmarking. There are some other technical factors involved, such as "queue depth", that I'm just beginning to learn about. I'm definitely a noob where SSD's are concerned.

All I know is I'm happy with my OCZ Solid3 so far. On a native SATA3 mobo with the OCZ I'm getting sustained 100-125MB/s throughput on my Acronis backups. It would be even faster if I were backing up to another native SATA3 device instead of my eSATA2 external. That's a real-time, real-world user result.

Thanks for the info. I ran Performance Test and this new 890FX SATA3 build with SSD blows away my old HP SATA2 machine with a WD spinner on all tests. My best HP score was 1142 and the new build ran 2685. Suits my needs for sure.