Crucial MX100 128GB: A Cheap But Good SSD For Linux Systems
Phoronix: Crucial MX100 128GB: A Cheap But Good SSD For Linux Systems
Crucial is out with a new solid-state drive line-up that's generating a lot of interest due to its lower price-per-Gigabyte than competing drives or even their former drives. The Crucial MX100 is the new SSD series and today we're testing out the Crucial MX100 128GB SSD, which costs just $80 USD (or about $0.62 per GB while the higher-capacity MX100 SSDs are comparatively even cheaper with the 512GB version costing less than $0.50 per GB).
Would be nice to know if this SSD also claims to support queued trim but fails rather horribly killing your data when you actually try it like the M500/M550 series. (Worse it seems Crucial claimed the bug is fixed with a newer firmware hence for a time the feature was reenabled in the kernel but apparently the bug was not actually fixed...)
I think the MX100 isn't blacklisted in the kernel but I believe that's because it's just too new rather than it is known to work...
In any case the 128GB version is indeed not very cost effective, since the 256GB version costs maybe 50% more and has 100% more capacity and on top of that for some usage scenarios the performance suffers quite a bit on the 128GB model. In contrast from 256GB to 512GB the price increase is proportional to the capacity and the performance difference is rather minimal.
Controller and Flash geometry correction
First, since the Vertex 3 uses a sandforce controller--which uses compression to enhance performance--I'm curious to know how compressable the data is that was used in those tests which show the Vertex 3 as having high performance.
Secondly, according to other reviewers, the 128GB version uses older 20nm Flash, not the 16nm which the 256 and 512 GB models use:
"At the other end of the spectrum, we should note that the MX100 128GB actually uses older 20-nm NAND. Only the 256GB and 512GB versions have the latest 16-nm flash." from http://techreport.com/review/26532/c...drive-reviewed
Does anyone have data on sustained write performance for this disk? I'm curious about whether it falls apart under a sustained load like the 840 unfortunately does.
Must say, I'm very surprised you people seem to have issues with Crucial SSDs, my Crucial M4 is running great.
Some random thoughts.
Some random thoughts
1) Vertex SSD seems to use built-in compression. Hence nice results in benchmarks. However, built-in compression also means that filesystems with compression would not gain speed benefit on such SSDs. So it can make sense to try some filesystems with compression to see how it performs on various SSDs. This would make results more in line with actual hardware abilities.
2) If something claimed to be "good for Linux", it can be really great idea to see if TRIM works and take a look how SSD options of various filesystems are performing with particular device.
I never said I have a problem - the kernel usually blacklists the affected models so it doesn't use queued trim (just the old non queued version) so you don't see a problem (I believe just about the only benefit of queued trim is that if you issue a trim command when there's other i/o happening it would be faster - of course if there's no other i/o you don't get any errors when you issue a queued trim neither). I'm pretty sure your m4 didn't support that feature in the first place. I read about this problem more by accident I was never affected by it as I never used a kernel which didn't blacklist my m500. Bugs leading to data loss are very very bad for a SSD though imho, it's nice the kernel works around it but really Crucial should do something about it.
Originally Posted by xeekei
But I think it's a valid question that the mx100 could also be affected - with the extreme similarity of it with the m550 / m500 (sharing the exact same controller as the former even) and Crucial's apparent inability or unwillingness to fix the firmware for the other two it isn't much of a stretch to imagine it could be affected too (and right now it would not be blacklisted).
According their support forums it seems they are selling SSDs with very buggy firmware.
It seems Intel does same.
Is there a firmware update to allow this device to support Queued TRIM or is it a hardware limitation?
Originally Posted by mczak