Fusion Drive VS Thunderbolt SSD

[CLS7]

I have an iMac, 27 inch, late 2015, Intel i5, 16 GB memory, 1 TB Fusion Drive, MAC OS X Sierra.

If I connecting an external Thunderbolt SSD to my Mac and migrate the operating system from the Fusion disk to this disk. Will this speeding up the operating system?

 

[Fishrrman]

What speeds is the fusion drive delivering RIGHT NOW?

To find out, download AmorphousDiskMark from here:
http://www.katsurashareware.com/pgs/adm.html

Run it, and post the results here.

Realize that with a thunderbolt external drive, you will probably see write speeds around 400mbps and reads in the 250-350mbps range.

A USB3 drive may yield even faster speeds (430+/-mbps read speed is what you can get from USB3 in an enclosure that supports UASP).

But again -- what speeds are you getting NOW?

 

[SaSaSushi]

I ran a Late 2013 iMac that way for several years, using the Fusion Drive for storage. To answer your question, I experienced no perceptible speed difference in regular usage compared to a freshly formatted Fusion Drive.

Whether the Fusion Drive will win in benchmarks depends on how full it is.

Regarding the advice above, while it is true that USB3 might even be slightly faster (again, probably only discernible in benchmark tests) than Thunderbolt I would advise against that option for boot disks due to the inability to use TRIM commands (not to mention inability to run firmware updates).

 

[maflynn]

I'm running OS X off a USB SSD and its fairly quick, in many ways quicker then my Fusion drive. I have a 2TB Fusion which has a larger SSD then the 1TB model, so you should see a larger performance delta.

Also consider a USB SSD, the speed difference between TB and USB is negligible in real world uses and typically USB drives are less expensive.

 

[Fravin]

I'm running OS X off a USB SSD and its fairly quick, in many ways quicker then my Fusion drive. I have a 2TB Fusion which has a larger SSD then the 1TB model, so you should see a larger performance delta.

Also consider a USB SSD, the speed difference between TB and USB is negligible in real world uses and typically USB drives are less expensive.

What external SSD are you using?

It's quite interesting this option...

 

[maflynn]

What external SSD are you using?

Samsung T3, though the T5 is out and it appears to be a very fast drive.

 

[Fravin]

And what about writing/reading time?

 

[CLS7]

If I plug in a USB 3.0 external SSD into my iMac, is there any support for TRIM? I read that Thunderbolt has TRIM support but not for USB 3.0. Is TRIM actually important?

What is your thoughts about Samsung T5 SSD? Seems to be a very fast drive.

http://www.trustedreviews.com/reviews/samsung-t5

 

[SaSaSushi]

If I plug in a USB 3.0 external SSD into my iMac, is there any support for TRIM?

No

I read that Thunderbolt has TRIM support but not for USB 3.0. Is TRIM actually important?

Yes (see above)

 

[maflynn]

Yes (see above)

I could be wrong but most SSD drives have some form of garbage collection and in my own experience I've not incurred any performance penalty from running OS X on my Samsung T3. I think the benefits of TRIM are overstated.

 

[SaSaSushi]

I could be wrong but most SSD drives have some form of garbage collection and in my own experience I've not incurred any performance penalty from running OS X on my Samsung T3. I think the benefits of TRIM are overstated.

I didn't say garbage collection isn't useful. I just said that it is not a replacement for TRIM as if often stated incorrectly in here and elsewhere.

This page at Cindori.org explains it pretty well.

The TRIM command is sent by the OS to the SSD to identify what pages of data can be ignored during garbage collection. The SSD cannot tell what files have been deleted until the OS uses the same sectors to store new files, but by that time the SSD has already wasted cycles by garbage collecting data that was invalid, but known to the SSD.

 

[maflynn]

I just said that it is not a replacement for TRIM as if often stated incorrectly in here and elsewhere.

My point is that people get hung on TRIM and over emphasize the benefits.

 

[SaSaSushi]

My point is that people get hung on TRIM and over emphasize the benefits.

Well, there is no question that TRIM is beneficial to performance and that it significantly reduces write amplification.

 

[Fishrrman]

A USB3 SSD will run fine with the iMac.
Don't let the fulminations about TRIM bother you.

ALL my external SSD's are USB3, and drive performance has NEVER become an issue.

 

[SaSaSushi]

ALL my external SSD's are USB3, and drive performance has NEVER become an issue.

Is this anecdotal evidence meant to mean that you believe the fact that the lack of TRIM leads to increased write amplification is fake news?

As much as I'd like to take the credit for it, I didn't invent the phenomenon.

 

[cynics]

A USB3 SSD will run fine with the iMac.
Don't let the fulminations about TRIM bother you.

ALL my external SSD's are USB3, and drive performance has NEVER become an issue.

Is this anecdotal evidence meant to mean that you believe the fact that the lack of TRIM leads to increased write amplification is fake news?

As much as I'd like to take the credit for it, I didn't invent the phenomenon.

How aren't you guys ignoring each other yet?

At the first sign of USB + TRIM discussions I look for @Fishrrman, put in my input and PEACE OUT!

 

[SaSaSushi]

How aren't you guys ignoring each other yet?

I would never ignore Fishrrman!

I enjoy reading his contributions in here and really, how boring would it be if we all agreed on everything?

 

[cynics]

I would never ignore Fishrrman!

I enjoy reading his contributions in here and really, how boring would it be if we all agreed on everything?

Hah I was mostly joking. However this is a sensitive topic. Lol

For me I 100% agree with you. However the fact remains, myself and many others have used USB externals as boot drives without any noticeable slow downs so it can't be discounted entirely.

 

[SaSaSushi]

For me I 100% agree with you. However the fact remains, myself and many others have used USB externals as boot drives without any noticeable slow downs so it can't be discounted entirely.

It is possible enough time has not passed yet for the long term effects of running without TRIM to become apparent. In another thread a user mentioned that their heavily-used 3.5 year old external, non-TRIMed USB3-connected Samsung 850 Pro SSD shows around 50% SSD lifetime left indicator while my own nearly 4 year old, always TRIMed Samsung 840 EVO still shows 92%.

I submit that at the very least there have to be negative effects from the increased write amplification in the long term.

 

[Fishrrman]

SaSa wrote:
"It is possible enough time has not passed yet for the long term effects of running without TRIM to become apparent. In another thread a user mentioned that their heavily-used 3.5 year old external, non-TRIMed USB3-connected Samsung 850 Pro SSD shows around 50% SSD lifetime left indicator while my own nearly 4 year old, always TRIMed Samsung 840 EVO still shows 92%.
I submit that at the very least there have to be negative effects from the increased write amplification in the long term."

Once again, I must respond.

I'm out-of-state and cannot test my 2012 Mac Mini and post the results, but I actually just tried it yesterday (using the AmorphousDiskMark utility) and got reads of 429mbps and writes around 260mbps (IIRC).
(By contrast, when brand new, the drive benchmarked at 431mbps read and 273mbps write, using BlackMagic.)

Regarding "write amplification", does that impact read/write speeds?
My read speeds are all-but identical to when the drives I've used were new; write speeds have dropped very slightly (see above), but nothing I'm concerned about.

Insofar as "lifetime left" is concerned, I'm not sure what can be used to check that. Can you recommend an application that can query this via USB?

I've booted and run (late-2012 i7 Mini) via USB/SSD since January 2013 and the drive performs as fast (at least from my own perception) as when it was new.

Perhaps Samsung SSD's have something in their controllers or circuitry that are vulnerable to write amplification. It's also possible that the Crucial and Sandisk drives that I use have something in -their- controller circuitry that deals with the drives while they sit "at idle". I do leave the Mini idling for a couple of hours each afternoon. Perhaps the SSD is doing some "internal cleanup" during those times.

After 4.5 years, if I was going to experience problems, I would think that they would have manifested by now. They haven't.
I can easily see getting 3-4 more years out of the current setup without problems...

 

[SaSaSushi]

Regarding "write amplification", does that impact read/write speeds?

Write amplification can affect speeds negatively. The bigger concern is that with an SSD you've got a limited number of writes per sector. Without TRIM, increased write amplification leads to lots more writes as data is moved around unnecessarily.

Insofar as "lifetime left" is concerned, I'm not sure what can be used to check that. Can you recommend an application that can query this via USB?

I use DriveDX and you can install it for free on a 10-day trial. This will show the SSD Lifetime Left Indicator.

It's also possible that the Crucial and Sandisk drives that I use have something in -their- controller circuitry that deals with the drives while they sit "at idle". I do leave the Mini idling for a couple of hours each afternoon. Perhaps the SSD is doing some "internal cleanup" during those times.

I believe the point with TRIM is that it is a line of communication between the OS and the SSD regarding the data structure of the drive so that there is no on-drive-only technology that can replace it.

After 4.5 years, if I was going to experience problems, I would think that they would have manifested by now. They haven't.
I can easily see getting 3-4 more years out of the current setup without problems...

It's possible but it's also possible you could have gotten twice that longevity with TRIM enabled.

 

[rapicell]

I've been using a thunderbolt driven SSD for months now and with the one I have, I get anywhere from 300-450 write and usually 400-450 read. I just leave the stock fusion drive as a beta tester for OS updates now, but I haven't had any issues so far.

 

[alien3dx]

I have an iMac, 27 inch, late 2015, Intel i5, 16 GB memory, 1 TB Fusion Drive, MAC OS X Sierra.

If I connecting an external Thunderbolt SSD to my Mac and migrate the operating system from the Fusion disk to this disk. Will this speeding up the operating system?

What my testing internal disk 5400 RPM, USB 3.1 pen drive , USB 3.0 external hardisk .. all same average 100 to 150 .
The only diff i see just loading windows in my old both windows laptop.,Else non.

Fusion is great idea to cut cost.

** if i had time i want to test usb 3.1 gen 2 western digital ssd once to see how much improve

 

[Fravin]

I have changed my HDD by a Sandisk's SSD, just oppened iMac and replaced the drives.

I've made a Fusion Drive with Apple's SSD, should you consider it? The Apple SSD is so small that I can't imagine running it alone.

Actually I've made a test installing only the OS on it and using the Sandisk SSD for homefolder and Apps, but mosto of the Apps (those purchased in AppStore) juts can't run in another folder. So I've decided to do a Fusion. The iMac is superfast now.

What do you think about it?