Originally Posted By: Mac
For those interested, here is a webpublished ongoing durability test project of SSDs going thru torture-test Write Cycles that supports what Matt is doingand advocating:

http://ssdendurancetest.com/

Pitting the Samsung 640 Evo, Intel SSD 520. SanDisk Ultra Plus and Kingston SSDNow V300 against one another.

So far, the Kingston has failed, at only 444.9 Write Cycles. Kingston has stated that they are going to replace under warranty.

Of the other three, the Samsung and Intel have just crossed the 3,000 cycle upper end recommendation point, the Sandisk has doubled that and is just over 6,000 at the moment.

Quote:

Is this really normal?

No. The tests preformed here must not be confused with normal use. They are designed to stress SSDs and simulate a really busy environment.
Normal workstation use would be more like 10-20GB written daily.


they also state:

Quote:

Please note that this unit is heavily stressed and in steady state most of the time.


Which begs the question, how would normal heatup/cooldown cycles, typical of normal use powerup/powerdown, along with the Writes, stress the drives? Likely not in a good way, expansion and contraction testing of other electronics bears that out. Nothing is a single-input problem in my field.

I would add the caveat that true testing should not involve a single example from each source.

I would like to have multiple examples from each source, side by side under same environment and conditions, enough multiple examples to provide some modicum of statistical substantiation. As any practicing test engineer should.

As usual for the intertubes, not. a. real. test.


--Mac



Mac,

That's it! You've put the fear of Silicon in me! I'm going to remove my SSD and replace it with a Rotating Hard Drive, and put my operating system there and everything else in my other hard drives.:-)

Seriously, if you think about it, if you just put the BiaB application on the SSD, there isn't much writing going on just reading. I'm not a software engineer by profession, but I have written several C programs with about 11K lines of code for GPIB control of my electronic lab test equipment.

I need to look at what's in the BiaB hard drive, but I'm assuming the BiaB application program is a .exe file. As far as I know that means it executes commands. You can't write to the .exe file and modify it. I will store samples, music files, etc. on the hard drive.

I just got off the phone with another drive designer with one of the largest drive maker on the east coast that I worked with for 18 years. He confirmed that the SSD should contain the OS and your applications for fastest throughput.

Just my 2 cents.

Wally G

Yamaha Montage 6 Roland J80, D50, Integra 7, BK-7m, FR-8X, 1967 Fender Jaguar, Fender Strat, Fender 1965 Twin Reverb reissue, Selmer Trumpet, Akai EWI, Studio One 4

My WEB Site: www.gontowski.com/music