Edit by Sean Webster:
Thanks to everyone who has posted in this thread and spread the word on the matter, Samsung has taken note of the Samsung Evo Series SSDs degrading performance of older data. A firmware fix for the issue is scheduled to be released October 15th.
Furthermore, I would like to personally thank Techie007 and BrainSplatter for their developing of benchmark software to verify this issue.
Their latest testing software is located here: http://www.overclock.net/t/1512915/read-speeds-dropping-dramatically-on-older-files-benchmarks-needed-to-confirm-affected-ssds
Hi to all!
In my system I have 3 Samsung SSD drives:
1: 256 840 PRO with Windows 7 OS and programs
2: 500GB 840 EVO
3: 1TB 840 EVO
Both EVO drives have the same exact problem, read speed on old files drops with time as low as ~60Mb/sec. New written files instead are as fast as advertised read/write speed ~450Mb/sec and so if I start a new benchmark with any benchmarking app.
Please note that the data must be "old" in terms of weeks or months to show the problem. Newly-written data read speed is fast as advertised, this is why benckmarks shows always a fast drive..
After a reboot the read speed improves a bit, but not much (as also reported in another thread, don't remember where) and after some "on" time return to the pre-reboot speed.
This seems to me a common hardware problem of those cheap SSD...
Anyone with those SSD with old data on it, may do some read speed test?
One good test for this is using the good old HDTach v3 program. It will show very well the drop on read speed on disk zones where old data is, using the "long bench" option. It works very well on windows 7 in XP compatibility mode.
This seems to me a serious problem on these disks. I have also a 256 840 PRO which does not have the problem (same PC and all SSD connected to INTEL AHCI ports, well configured).
To illustrate, here is a HDTach bitmap of my 500 EVO, where the last part is "free" space:
And to compare, here is my 840 PRO, same pc, same AHCI Intel ports:
And as I said, posting a new benchmark is useless as it will be exactly as advertised... The problem is reading old data and it gets worse over time.