Login Register
MacInTouch Home Page MacInTouch Discussions

MacInTouch Amazon link...
FireWire, USB, Thunderbolt and SSD
 


2017-10-08 at 02:08 #26562   (41)
(2017-10-07 at 21:21)iansax08 wrote:  ... For the last few days my Thunderbolt drive either won't mount, or will unmount itself shortly after appearing on my desktop. I receive error messages informing me that the drive wasn't ejected properly, even though I haven't touched it. ...
I wonder if your cabling has a glitch in it. If you have similar cables, try swapping them around. Just an idea.


2017-10-08 at 11:08 #26570   (42)
(2017-10-08 at 02:08)Rich-S*. wrote:  
(2017-10-07 at 21:21)iansax08 wrote:  ... For the last few days my Thunderbolt drive either won't mount, or will unmount itself shortly after appearing on my desktop. I receive error messages informing me that the drive wasn't ejected properly, even though I haven't touched it. ...
I wonder if your cabling has a glitch in it. If you have similar cables, try swapping them around. Just an idea.
I tried 3 separate cables to the drive, in case that was the problem. No solution, sorry to say. Thanks, though!


2017-10-08 at 17:32 #26574   (43)
(2017-10-08 at 02:08)Rich-S*. wrote:  
(2017-10-07 at 21:21)iansax08 wrote:  ... For the last few days my Thunderbolt drive either won't mount, or will unmount itself shortly after appearing on my desktop. I receive error messages informing me that the drive wasn't ejected properly, even though I haven't touched it. ...
I wonder if your cabling has a glitch in it. If you have similar cables, try swapping them around. Just an idea.
Yes, I tried that. I have 4 separate cables that I used on the SSD, and I tried inserting each of the cables into each of the 4 Thunderbolt ports on my laptop. Thanks for the suggestion! I'm runnung out of options. I'll try contacting the manufacturer after the weekend; perhaps they can offer some ideas/relief/ replacement drive...


2017-10-12 at 14:20 #26876   (44)
(2017-10-07 at 21:21)iansax08 wrote:  ... I am a composer, so I store most of my samples on a 1TB External SSD, an Oyen Digital U32 Shadow. For the last few days my Thunderbolt drive either won't mount, or will unmount itself shortly after appearing on my desktop. I receive error messages informing me that the drive wasn't ejected properly, even though I haven't touched it....
You say your Thunderbolt drive won't mount. Are you referring to your Oyen Digital U32 Shadow? Because I can only find it on the internet as a USB 3.0 or 3.1 drive, not Thunderbolt. The pictures of the drive labelled as USB 3.1 show a USB 3.0 port, presumably so it is backwards-compatible with USB 3.0.

In reviews for an SSD version on Amazon, at least one reviewer posts that it keeps disconnecting. So you're not alone. A replacement may make a difference, and maybe Oyen will replace it.


2018-02-05 at 22:54 #32943   (45)
I'm seeing shockingly bad performance for Thunderbolt Target Disk Mode, and I'm wondering if anyone else is seeing the same.

I'm working with a 2011 MacBook Pro 13" (Thunderbolt 1) and a 2015 MacBook Pro 15" (Thunderbolt 2), both equipped with SSDs. I'm using Blackmagic Disk Speed Test to check performance.

With the 2011 MacBook Pro in Thunderbolt Target Disk Mode, I see performance around 100Mbps - about 1/4 what it should be and about 1/10 of what the bandwidth should be. This is hard disk speed, not SSD speed.

Swapping roles, so the 2015 MacBook Pro is in Thunderbolt Target Disk Mode, I see better performance, but it's still less than half what the bandwidth should allow and less than a quarter of the speed the SSD offers natively.

Things I've tried - all failing to resolve the problem:
  • running on battery vs. a/c adapter
  • 0.5m Akitio cable vs. 6' OWC cable
  • FileVault vs. unencrypted HFS
Anyone else seeing this massive bottleneck?


2018-02-06 at 06:42 #32952   (46)
I have seen similar performance issues with an older MacBook Pro model in Target Disk Mode when I was attempting to copy its hard disk contents to a newer MacBook Pro. After some experimentation I came to the conclusion that it's related to the model (or a combination of them) and the direction (which one being in Target Disk Mode).

And I also believe (it's been a while when I did this) that it made a difference whether I booted the other Mac from its Rescue system or from a regular macOS installation.

I am not sure of the MacBook Pro's exact model, but it was a model that contained a DVD drive, non-Retina (like the Mid-2012).


2018-02-07 at 05:38 #33001   (47)
Markh
I recently set up a new/used 2012 MacBook Pro. The new 2012 had a mechanical hard disk, so I bought a new SSD to replace it.

I cloned the hard disk to SSD as follows:
Booted it into target disk mode. I connected it to my 2011 17" MacBook Pro via Thunderbolt. I put the SSD in a Newer Tech Voyager drive dock, connected to my 17" via an eSATA card. I used Carbon Copy Cloner to clone the hard disk to the new SSD. Transplant the SSD into the 2012. I did not notice any particularly slow performance, but I wasn't benchmarking it either.


2018-02-07 at 10:30 #33007   (48)
(2018-02-07 at 05:38)Markh wrote:  I cloned the hard disk to SSD as follows... I did not notice any particularly slow performance, but I wasn't benchmarking it either.
I did something similar quite a while ago and was surprised at how long it took, since I was going internal SSD-to-SSD (Mac-to-Mac) over Thunderbolt, but I hadn't benchmarked it either.

If you could run Blackmagic Disk Speed Test or a similar benchmark, I'd love to hear how fast any SSD is in Thunderbolt Target Disk Mode on your Macs.


2018-02-07 at 15:47 #33027   (49)
markh
(2018-02-07 at 10:30)Ric Ford wrote:  I did something similar quite a while ago and was surprised at how long it took, since I was going internal SSD-to-SSD (Mac-to-Mac) over Thunderbolt, but I hadn't benchmarked it either.
   If you could run Blackmagic Disk Speed Test or a similar benchmark, I'd love to hear how fast any SSD is in Thunderbolt Target Disk Mode on your Macs.
I just ran Blackmagic Disk Speed Test with my early 2011 17" MacBook Pro (8,3) running in target disk mode, connected to my other 17" MacBook Pro (late 2011) - internally it has a 960GB OWC Mercury Extreme Pro 6G SSD.

I'm getting about 100 MB/s write and 180 MB/s read. That is much slower than the maximum theoretical speed possible, but much faster than hard drive speed. It is also slower than an SSD connected via the eSATA card, which is in the range of 200-300 MB/s. The maximum should be in the 400-500 MB/s range.


2018-02-07 at 20:53 #33043   (50)
(2018-02-07 at 10:30)Ric Ford wrote:  If you could run Blackmagic Disk Speed Test or a similar benchmark, I'd love to hear how fast any SSD is in Thunderbolt Target Disk Mode on your Macs.
For what it's worth, here are a couple tests using adapters:

MacBook Pro (17-inch, Mid 2010) with 1TB 3GB SSD in Target Disk Mode
to
MacBook Pro (15-inch, 2016) using Apple Thunderbolt 3 (USB-C) to Thunderbolt 2 Adapter [to Thunderbolt-FireWire adapter...]

37 MB/s Read
30 MB/s Write

MacBook Air (13-inch,Mid 2013) in Target Disk Mode
to
MacBook Pro (15-inch, 2016) using Apple Thunderbolt 3 (USB-C) to Thunderbolt 2 Adapter

198 MB/s Read
125 MB/s Write


lichinho Show this Post
2018-02-08 at 16:27 #33083   (51)
lichinho
(2018-02-07 at 20:53)Scott wrote:  MacBook Pro (17-inch, Mid 2010) with 1TB 3GB SSD in Target Disk Mode
to
MacBook Pro (15-inch, 2016) using Apple Thunderbolt 3 (USB-C) to Thunderbolt 2 Adapter

37 MB/s Read
30 MB/s Write
How did you connect the 2010 MacBook Pro? That Mac does not have Thunderbolt, so I can only assume you did it via FireWire, which would require not just TB3->TB2 adapter but combining that with TB2->FW.

Here, using a 2014 Retina MacBook Pro (TB2) in Target Disk Mode, connected to a 2011 iMac (only TB1), I got over 300MB/s, both read and write (reads just a little higher than writes).


2018-02-10 at 17:47 #33153   (52)
(2018-02-08 at 16:27)lichinho wrote:  How did you connect the 2010 MacBook Pro? That Mac does not have Thunderbolt, so I can only assume you did it via FireWire, which would require not just TB3->TB2 adapter but combining that with TB2->FW.
Yes, thank you for the correction.

Thunderbolt 3 (USB-C) to Thunderbolt 2 Adapter + Thunderbolt 2 to FireWire 800 Adapter.


2018-02-14 at 11:45 #33321   (53)
(2018-02-05 at 22:54)Ric Ford wrote:  I'm seeing shockingly bad performance for Thunderbolt Target Disk Mode, and I'm wondering if anyone else is seeing the same. I'm working with a 2011 MacBook Pro 13" (Thunderbolt 1) and a 2015 MacBook Pro 15" (Thunderbolt 2), both equipped with SSDs. I'm using Blackmagic Disk Speed Test to check performance.
   With the 2011 MacBook Pro in Thunderbolt Target Disk Mode, I see performance around 100Mbps - about 1/4 what it should be and about 1/10 of what the bandwidth should be. This is hard disk speed, not SSD speed....
I ran some tests on the 2011 MacBook Pro with an Akitio SATAgo Thunderbolt-to-eSATA adapter connected via eSATA cable to a CRU RAID storage system with two Seagate Barracuda drives in a RAID-0 configuration, which produced speeds over 200 MBps. So, it's not Thunderbolt, per se, that's the problem. It seems more like Target Disk Mode may be a bottleneck for some reason.


2018-04-12 at 20:25 #36644   (54)
I stumbled across the following regarding Target Disk Mode:

Apple wrote:How to use target disk mode to move files to another computer
It seems to indicate that the only Thunderbolt varieties to support TDM are 2 and 3. What about those of us using the original version, such as my Late 2012 iMac?

Is this just an accidental omission on Apple's part?


2018-04-12 at 21:33 #36645   (55)
(2018-04-12 at 20:25)anon9678 wrote:  It seems to indicate that the only Thunderbolt varieties to support TDM are 2 and 3. What about those of us using the original version, such as my Late 2012 iMac? Is this just an accidental omission on Apple's part?
It's an omission. I use target mode regularly to sync between my Late 2012 iMac and Early 2013 MacBook Pro. I've also used it with other Thunderbolt Macs.


2018-04-12 at 21:37 #36646   (56)
(2018-04-12 at 21:33)CyborgSam wrote:  
(2018-04-12 at 20:25)anon9678 wrote:  It seems to indicate that the only Thunderbolt varieties to support TDM are 2 and 3. What about those of us using the original version, such as my Late 2012 iMac? Is this just an accidental omission on Apple's part?
It's an omission. I use target mode regularly to sync between my Late 2012 iMac and Early 2013 MacBook Pro. I've also used it with other Thunderbolt Macs.
I wonder if it might be an intentional omission, due to the terribly poor performance (documented above with Thunderbolt 1 Target Disk Mode.


2018-04-12 at 23:36 #36647   (57)
(2018-04-12 at 21:37)Ric Ford wrote:  I wonder if it might be an intentional omission, due to the terribly poor performance (documented above with Thunderbolt 1 Target Disk Mode.
FWIW: I always get fast transfer speeds between two Thunderbolt 1 Macs using an Apple cable.


2018-04-12 at 23:57 #36648   (58)
(2018-04-12 at 23:36)CyborgSam wrote:  FWIW: I always get fast transfer speeds between two Thunderbolt 1 Macs using an Apple cable.
Could you try running Blackmagic Disk Speed Test and let us know what kind of transfer rates you get? Thanks!


2018-04-13 at 02:44 #36654   (59)
(2018-04-12 at 21:37)Ric Ford wrote:  I wonder if it might be an intentional omission, due to the terribly poor performance (documented above with Thunderbolt 1 Target Disk Mode.
If that were the case, I'm sure they would have omitted FireWire at the same time. Isn't even FW800 slower than any version of Thunderbolt?


2018-04-13 at 08:41 #36659   (60)
(2018-04-13 at 02:44)anon9678 wrote:  If that were the case, I'm sure they would have omitted FireWire at the same time. Isn't even FW800 slower than any version of Thunderbolt?
The point is that Target Disk Mode performance for Thunderbolt 1 is many times slower than its 10Gbps bandwidth, which means something's very wrong with Apple's implementation. FireWire 800 Target Disk Mode performance, by contrast, seems normal - on par with regular FireWire performance and bandwidth. Thunderbolt 1 is supposed to have 12.5 times the bandwidth of FireWire 800, but you don't see anything remotely like that kind of improvement when switching from FireWire Target Disk Mode to Thunderbolt Target Disk Mode and actually transferring files.


2018-04-13 at 11:57 #36666   (61)
I wasn't able to use Thunderbolt Target Disk Mode between an early 2011 MacBookPro (Thunderbolt 1) and a 2017 MacBook Pro (Thunderbolt 3) with a Thunderbolt 2 cable and the Thunderbolt 3-to-Thunderbolt 2 Apple adapter.

The computers could not be detected.

It worked fine with a combination of the Thunderbolt 3-to-Thunderbolt 2 adapter, a Thunderbolt 2-to-FireWire adapter and a standard FireWire 800 cable.


2018-04-13 at 12:59 #36669   (62)
(2018-04-13 at 08:41)Ric Ford wrote:  The point is that Target Disk Mode performance for Thunderbolt 1 is many times slower than its 10Gbps bandwidth, which means something's very wrong with Apple's implementation. FireWire 800 Target Disk Mode performance, by contrast, seems normal - on par with regular FireWire performance and bandwidth. Thunderbolt 1 is supposed to have 12.5 times the bandwidth of FireWire 800, but you don't see anything remotely like that kind of improvement when switching from FireWire Target Disk Mode to Thunderbolt Target Disk Mode and actually transferring files.
[In certain cases], the limiting factor is the speed of the drive, not Thunderbolt. For example, the hard drive in my 2011 iMac can reach about 120MBps throughput, which is some ten times less than Thunderbolt 1 and only about 1.5 times the capability of FireWire. That is a limitation of the hard drive itself, no matter how it is connected. So at best, going from FireWire to Thunderbolt with such a drive in TDM could improve 1.5 times.

The scaling is better if the drive is an SSD, but even then, in most cases, even the SSD can not saturate the Thunderbolt 1 bus. For example, the internal SSD in my MacBook Pro shows about 700MBps in the BlackMagic test - which is, again, less than the theoretical 10Gbps of Thunderbolt 1.

That said, I have just tried both the above drives using Target Disk Mode in each direction (the 2011 iMac has only Thunderbolt 1, the MacBook Pro has Thunderbolt 2). The throughput of both the HD and SSD when used via target disk mode is about 2/3 (HD) or 1/2 (SSD) of the throughput of the same drive when accessed directly in the machine where it is installed. A little penalty would seem normal, but this seems a little bit poor performance of Target Disk Mode.


2018-04-13 at 13:36 #36672   (63)
(2018-04-13 at 12:59)lichinho wrote:  The scaling is better if the drive is an SSD, but even then, in most cases, even the SSD can not saturate the Thunderbolt 1 bus. For example, the internal SSD in my MacBook Pro shows about 700MBps in the BlackMagic test - which is, again, less than the theoretical 10Gbps of Thunderbolt 1.
Well, of course a single hard drive isn't going to saturate a Thunderbolt bus (unless Apple is doing something very wrong with Thunderbolt), but Apple promises 3.2GB/s SSD performance for its current MacBook Pro and I don't think you'll see anything close to Thunderbolt 1 saturation in Target Disk Mode, even though the MacBook Pro SSD is 25 times faster than Thunderbolt 1....


2018-04-13 at 19:11 #36680   (64)
(2018-04-13 at 08:41)Ric Ford wrote:  The point is that Target Disk Mode performance for Thunderbolt 1 is many times slower than its 10Gbps bandwidth, which means something's very wrong with Apple's implementation. FireWire 800 Target Disk Mode performance, by contrast, seems normal - on par with regular FireWire performance and bandwidth. Thunderbolt 1 is supposed to have 12.5 times the bandwidth of FireWire 800, but you don't see anything remotely like that kind of improvement when switching from FireWire Target Disk Mode to Thunderbolt Target Disk Mode and actually transferring files.
Wow. Is this just with Target Disk Mode, or is FireWire 800 still superior in general usage as well? ie., is the problem with the overall implementation of Thunderbolt 1?


2018-04-13 at 20:21 #36682   (65)
(2018-04-12 at 23:57)Ric Ford wrote:  Could you try running Blackmagic Disk Speed Test and let us know what kind of transfer rates you get? Thanks!
I measured using Blackmagic, my Thunderbolt speeds were slower than I expected. I felt it was fast because it's twice as fast as FireWire. I guess with my old hardware my expectations are low...

Blackmagic Disk Speed Test was run on my Late 2012 iMac 21.5". My Early 2013 MacBook Pro Retina 15" with an Apple 750GB SSD was used for Target Mode testing.

I also measured my iMac's internal OWC 6G SSD plus two external 4TB hard drives in OWC USB-3 cases with internal SATA 6G. The slower hard drive is SATA 3G, the faster is SATA 6G.

Target Mode: Thunderbolt 1 to Thunderbolt 1:
Write: 144 MB/sec
Read: 172 MB/sec

Target Mode: FireWire to FireWire using two Thunderbolt dongles:
Write: 59 MB/sec
Read: 54 MB/sec

Internal OWC 1TB 6G SSD:
Write: 429 MB/sec
Read: 505 MB/sec

External Hitachi HDE721010SLA330 4TB in OWC USB-3:
Write: 69 MB/sec
Read: 71 MB/sec

External HGST HDN724040ALE640 4TB in OWC USB-3:
Write: 136 MB/sec
Read: 135 MB/sec


2018-04-14 at 00:56 #36686   (66)
(2018-04-13 at 19:11)anon9678 wrote:  
(2018-04-13 at 08:41)Ric Ford wrote:  The point is that Target Disk Mode performance for Thunderbolt 1 is many times slower than its 10Gbps bandwidth...
Wow. Is this just with Target Disk Mode, or is FireWire 800 still superior in general usage as well? ie., is the problem with the overall implementation of Thunderbolt 1?
Thunderbolt 1 was slower than USB 3.0 in an older MacBook Pro, despite having double the bandwidth (theoretically):
Post #18809 in "SSD and USB flash benchmarks."


2018-04-14 at 20:34 #36703   (67)
A while back I tested a 2011 MacBook Pro in target disk mode and connected via the Thunderbolt port. I think Blackmagic speed test reported in the neighborhood of 100 MB/sec write and 200 MB/sec read. The Mac has an SSD. That supports other observations that the Thunderbolt port is slow in target disk mode.

I don't have any external Thunderbolt drives with an SSD, only mechanical drives, so I can't test the max speed of the port with a normal drive. I have used Carbon Copy Cloner to clone my internal drive to external drives, both FIreWire and Thunderbolt. I have observed that the Thunderbolt port is much faster. It can also boot from a Thunderbolt drive and be usable. Booted from a FireWire drive is so slow, it is futile.

On the other hand, I have not noticed much difference with Time Machine backups. I suspect that Time Machine has lots of overhead that negates the speed advantage of the Thunderbolt port. I have since been using FireWire drives exclusively for Time Machine backups. Part of that is driven by the fact that I only have 1 Thunderbolt port, which usually has an external monitor connected.


2018-04-16 at 00:49 #36716   (68)
(2018-04-14 at 00:56)Ric Ford wrote:  Thunderbolt 1 was slower than USB 3.0 in an older MacBook Pro, despite having double the bandwidth (theoretically)...
That's amazing!
This makes me wonder if I could see even better performance from my G-Drive with Thunderbolt drives if I switch the connection from Thunderbolt 1 to USB3!


2018-04-16 at 20:34 #36736   (69)
If I restart my system, my external hard drives connected via Thunderbolt 1 spin down, then immediately spin back up even before they've finished spinning down. Isn't this bad for them? It seems like the electrical equivalent of whiplash.

If I worked at Apple, I would make it so that external hard drives kept spinning across restarts.

I've been cautious, shutting down the computer completely and waiting 10 seconds before booting back up, in instances where most people would issue a Restart command, in an effort to be "kind" to the drives.