SSD Read Speed Tester

http://www.techspot.com/downloads/6712-ssd-read-speed-tester.html

TLC NAND based SSDs have recently shown a trend of having crippled performance when reading old files (2 to 3 months old data for example) that could make read speeds go as low as 50 MB/S, which is about an 80% of a drop of their normal performance that can reach up to around 500 MB/S.

Since synthetic benchmarks perform tests by writing a new fresh temp file and then trying to read it, this deterioration in performance is very hard to distinguish due to the nature of how synthetic benchmarks work.

If you have a TLC NAND based SSD such as the Samsung 840, Samsung 840 EVO, SanDisk Ultra II, etc. you may want to run this tool that will read your current data to show you the real world performance when it comes to read speeds to see if you have an issue with your TLC NAND based SSD or not. Samsung recently released a firmware update to fix the issue of the 840 EVO 2.5" (not the mSATA version yet) read speed deterioration by having an algorithm in the SSD's controller that will periodically refresh old data to bring its speeds back to normal.

Here are my 840 EVO mSATA Read Speed results of 2 months old data. I have never experienced this slow down really, either the mSATA doesn't experience this slowdown or I got a very lucky sample as its performance has always been higher than the 2.5" 840 EVO counterpart in my testing:

 

I think one of the reasons people have different experiences with 840 EVO is temperatures of their system: ideally, contrary to what you may think, for best data retention, you want high temperatures when writing the data and low temperatures when system is shut down.
Here is some info from Intel, but probably relates to others as well:
http://www.anandtech.com/show/9248/the-truth-about-ssd-data-retention

 

Wow. Very interesting.



So the hotter it runs and the cooler it's stored, the longer the data lasts. Just 5C difference can double or halve the storage life. Very interesing. So we actually want hot running SSD's then (or at least hot NAND, not controller).

 

When you really think about it, it actually makes sense: when you're writing, you want to "push" as many electrons inside memory cell as possible, hotter temperature makes electrons "jump" inside easier. While in storage you want to limit electron leakage from the cell and lower temperatures are better to do so. This would also suggest rewriting old data on regular basis would be necessary on all SSD drives, not only Samsung 840 evo, for example if you run your system 24/7 at somehow steady 40C, you would need to refresh the data every 14 weeks or so, according to this table. That begs the question WTH did Samsung did to their drivers to create that 840 evo mess, if they were properly rewriting old data from the beginning, all would be fine . Keep in mind the data comes from Intel and there is probably more to it, that they don't say, but still, general idea stands.

 

No. You don't. This is the same misinterpretation of the data that tech news jumped on a few weeks ago. That data, and the paper associated with it, is for flash-based SSDs that are worn out and close to failing due to wear:

http://www.pcworld.com/article/2925...nt-lose-data-if-left-unplugged-after-all.html

 

Of course this data is for worse case scenario where drive had 100TB of writes and it's from 5 yrs ago, but what PCWorld article you linked to forgot to mention is the fact that since then memory cells are much smaller, each memory cell is stacked, holding more data and therefore more sensitive to error and I'm assuming the relations of writing/storing temperatures still holds true (best to write hot and store cold) even if not exactly following numbers in the table (how could they with so many variations in technology between different drives). I wrote this in context of 840 evo slow downs, as possible explanation of different experience by different users. As a matter of fact, looking at this again I wonder if 840 evo meets JEDEC specification of data retention for 1 year under 30c storage at the EOL, if brand new drives start showing reading issues. As a matter of fact I would love to see retention data for brand new, current SSD drives, not some ambiguous "normal use", "typical consumer" terms.

 

Well i have an 512GB SLC SSD, see my sig, and i ran the test, and the old files read considerably slow than the newer one, my SDD has had around 25TB of writes.

My latest firmware reduce my ssd controller from SATA600 to SATA300, I dont know why HP did this, but i dont really seen any real world difference in performance.

John.

 

No, it's a different cause. Wear is actual physical damage: the electron tunneling mechanism causes the flash cells to deteriorate. When they deteriorate to the point of structural near-failure they can lose bits in high temperature conditions.

The 840 EVO problem is not due to physical deterioration. If it were then rewriting would not help; it would make the problem worse by causing further deterioration.

 

Ok, I did a secure erase using my plextools app it took approx 10 seconds yesterday, I did a clean install of win8.1 and below is the results.

I wonder if the spike at 93 weeks is some old software i downloaded, but this would mean the benchmark might have a problem with reading the original creation date, and not my download/install date???

John.

 

Electron leakage from memory cell happens at all times, it's faster as the cell wears off, but it happens from brand new, as a matter of fact the tables for data deterioration are for cells around 100TB mark and we know from tests many SSD will survive 10 times as much, but if you know more than technical writers at Anandtech and technicians at Samsung, who I'm assuming provided the writer with technical info, then no point in arguing.

 

Well, gee, according to Samsung the cause of the problem with the 840 EVO series -- only the 840 EVO series -- is a bug in the original firmware versions that causes a problem that triggers the error correction mechanisms. It's not deterioration of the flash cells which is the point of the presentation cited in the article that you yourself linked, an article among the many that grossly misinterpreted the presentation, a presentation addressing extreme conditions that you are unlikely to encounter unless you are in the habit of breaking down while driving a truck full of worn-out SSDs across the Arizona desert at mid-day in July.

Or, you know, you can just believe the hype and wear out your SSDs with excessive "protection".

 

840 Evo 2.5" SATA had a new firmware that looks like it's fixed this read speed of old data issue once & for all. I'm using the latest firmware since the day it launched 8 weeks ago, and read speeds are exactly the same at 523MB/s for the OS Partition, and 543MB/s for the Games Partition (+/- 2MB/s). Previously to the latest firmware data was slowing down at 8 weeks. I think it's fixed now. Various threads have now gone quiet on the 840 Evo since the release of latest firmware - probably because there's nothing to complain about anymore after most people's initial reports that the latest firmware has sorted their read speeds. At least Samsung finally managed to fix it - yes I'm being positive, but with the way the new firmware approaches the problem I'd be extremely surprised if it ever slows down again.