How big of a deal is the Yosemite -TRIM-SSD drama?

[NeilHD]

I've yet to install my SSD (just got a 250Gb Samsung 840 to install in my new 2012 mini), and have done a bit of reading around TRIM as I hadn't heard of it before (my first SSD. Well my work PC has one but it came like that).

TRIM, as we all know by now, improves garbage collection by the OS sending a "Hey, clean this up will ya?" message to the SSD when a file is modified/deleted. The SSD then immediately does so, rather than waiting for normal garbage collection to run. This is important because of the way the data is split into pages on the drive.

What this means is you see a big increase in read/write speeds when running performance tests. But here's what I think is the key point that most people miss - performance tests by their very nature do a whole bunch of writing/deleting/reading very quickly to measure the throughput and therefore benefit most from TRIM being enabled. I think in normal day-to-day use the difference would not be that noticeable as you simply do not run often the same scenario in real-life usage.

Just my opinion. If anyone has any actual research to prove otherwise then I'll gladly admit to being wrong.

 

[qcmacmini]

I have read that even the latest SSDs benefit massively with TRIM, but only once the drive is nearing capacity. So, if you have a small drive you'll notice it a lot sooner than with a big one.

 

[Ethosik]

I've yet to install my SSD (just got a 250Gb Samsung 840 to install in my new 2012 mini), and have done a bit of reading around TRIM as I hadn't heard of it before (my first SSD. Well my work PC has one but it came like that).

TRIM, as we all know by now, improves garbage collection by the OS sending a "Hey, clean this up will ya?" message to the SSD when a file is modified/deleted. The SSD then immediately does so, rather than waiting for normal garbage collection to run. This is important because of the way the data is split into pages on the drive.

What this means is you see a big increase in read/write speeds when running performance tests. But here's what I think is the key point that most people miss - performance tests by their very nature do a whole bunch of writing/deleting/reading very quickly to measure the throughput and therefore benefit most from TRIM being enabled. I think in normal day-to-day use the difference would not be that noticeable as you simply do not run often the same scenario in real-life usage.

Just my opinion. If anyone has any actual research to prove otherwise then I'll gladly admit to being wrong.

Agreed. And I do not know why benchmarks are almost double on some user's results IMMEDIATELY after activating TRIM. it doesn't work like that. TRIM is not something that run in the background all the time, it only happens when you modify or delete a file.

My old Mac Pro is stuck on SATA 2. Brand new drives with and without TRIM can only get 250 MB/s read and write in both OS X and Windows. That is far more of a performance loss than TRIM. It is still ridiculously fast.

 

[qcmacmini]

TRIM is simply a software based solution to tell the drive that data is now deleted and those sectors can be written to again. It only takes one pass apparently to regain performance. Apparently once a drive is 'full', well the SSD thinks it's full, write speeds can decrease to less than HDD speeds.

 

[KrisLord]

Does this mean that I don't have to worry about TRIM if I installed an SSD (in addition to a HDD) in order to create a DIY Fusion drive?

FYI, I bought a 2012 Mini (i7 quad & 1TB HDD) and was planning to install an SSD (Intel 730 [SSDSC2BP480G4R5]) in addition to the HDD to create a fusion drive. Guess that I won't need TRIM enabler, correct?

BTW, I am planning to stay on Mavericks for the time being, also for the reason mentioned on the following website.
http://www.cindori.org/trim-enabler-and-yosemite/

I think that quote meant having a fusion drive doesn't effect the trim/no trim question. You can still enable trim with a DIY fusion drive.

I'd actually suggest that in a fusion setup trim is more important than a non fusion setup. The reason being that with fusion drives, your SSD only has 4GB of free space (assuming the fusion drive has more data than the SSD capacity) and the difference in performance between trim vs no trim is more pronounced when a drive is full.