Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Just got a hold of the Seagate GoFlex For Mac 1TB kit, comes with the adapter and the thunderbolt cable. The cable is surprisingly short, at about twice the length of my palm.

The drive itself performs slightly better on the TB interface, measuring 110mb write and 110mb read. Comparatively my WD Black 1TB desktop drive gets 70mb read/write on the BlacX USB 2.0 interface.

For those in the US and Canada, you can price match/beat it at SIGelectronics.com as they are having a sale tomorrow for $128. Unbelievable price considering it costs $99 for the adapter and $50 for the cable itself.

it is going to sell out there are six left. it is a good price.
 
Just got a hold of the Seagate GoFlex For Mac 1TB kit, comes with the adapter and the thunderbolt cable. The cable is surprisingly short, at about twice the length of my palm.

Interesting. I have two Seagate GoFlex Thunderbolt adapters – one portable and one GoFlex Desk. Neither came with Thunderbolt cables. Every other GoFlex product I have – USB, USB 3.0, Firewire 800 – all shipped with short (8-10") cables. This includes GoFlex drive kits as well as separate adapters.

I've just run some speed tests, by the way. Doing a transfer of a 30GB file between an internal 750GB SSD and drives on the Thunderbolt portable adapter I get the following:

Write To WD 1TB 5400RPM WD10JPVT – 41 MB/sec
Read From WD 1TB 5400RPM WD10JPVT – 42 MB/sec
Write To Seagate 750GB 7200RPM ST9750420AS – 111 MB/sec
Read From Seagate 750GB 7200RPM ST9750420AS – 117 MB/sec

It's pretty clear that the Thunderbolt adapter performance is constrained by the drive it's connected to. The Thunderbolt bus itself is spec'd at 10Gbps (1280 MB/sec).
 
Interesting. I have two Seagate GoFlex Thunderbolt adapters – one portable and one GoFlex Desk. Neither came with Thunderbolt cables. Every other GoFlex product I have – USB, USB 3.0, Firewire 800 – all shipped with short (8-10") cables. This includes GoFlex drive kits as well as separate adapters.

I've just run some speed tests, by the way. Doing a transfer of a 30GB file between an internal 750GB SSD and drives on the Thunderbolt portable adapter I get the following:

Write To WD 1TB 5400RPM WD10JPVT – 41 MB/sec
Read From WD 1TB 5400RPM WD10JPVT – 42 MB/sec
Write To Seagate 750GB 7200RPM ST9750420AS – 111 MB/sec
Read From Seagate 750GB 7200RPM ST9750420AS – 117 MB/sec

It's pretty clear that the Thunderbolt adapter performance is constrained by the drive it's connected to. The Thunderbolt bus itself is spec'd at 10Gbps (1280 MB/sec).


This is the kit that comes with the 1tb drive and cable. Your WD is transferring only ~40mb/s? Min does 110mb/s, even if its a 5400rpm drive it seems off.
 
This is the kit that comes with the 1tb drive and cable. Your WD is transferring only ~40mb/s? Min does 110mb/s, even if its a 5400rpm drive it seems off.

To be fair, the 5400 RPM WD was formatted into three partitions, two of which had data on them. The transfer was onto the empty partition. The 7200 RPM Seagate, on the other hand, was a freshly formatted single partition drive. Other than that, everything was identical. I ejected the WD, plugged in the Seagate and transferred the same file to and from the internal SSD.

Still, there's a pretty huge difference in throughput. Even 111 MB/sec is an order of magnitude lower than the rated Thunderbolt spec.

I'm going to try to test with a clean 5400 RPM drive later today.
 
Got a hold of a second kit today and went about messing around with raid 0/1. In osx it performs as expected, doubling read and write performance to low 200s. The raid was not, however, detected when I booted into my windows 7 bootcamp.

Perhaps I'll delete the fat raid partition and set up the raid while in windows 7. Haven't tried that yet.

With mechanical drives would I be wrong to assume that although sustained read/writes benefit, random seek times will have increased rather than decreased?

Doing a 500GB Time Machine backup and notice the TB cable that comes with the kit is hotter than the one from Apple. The head on the Apple TB cable has a larger surface area, perhaps that contributes to it keeping cooler. None of the drives have flaked out disappeared on me like the Vertex 4 has. I would have thought the 1TB mechanical drive would consume more energy than an SSD would, but alas it has been stable thus far *knock on wood*.
 
Last edited:
...With mechanical drives would I be wrong to assume that although sustained read/writes benefit, random seek times will have increased rather than decreased?...

Maybe. It depends on the drives used.

I tested many raid0 hdd setups 2, 3 even 4 drives. Did it with different case different drives did it with fw800 sata and t-bolt. Some chipsets work better then others.


The best case of better seek time for random read write was with 1tb samsung 3.5 inch drives with owc's mercury pro Qx2 a raid0 with 4 hdds using fw800 gave good improvement. maybe 7 or 8 mbs which for hdds in random 4k tasks is not terrible. must are under 2 mbs when run as a single drive. this was far better with the 1tb samsung 3.5 inch hdds.
 
Get 330 MB/s Write and 364 MB/s Read speeds on my Seagate Thunderbolt adapter connected to a Samsung 830 series 256GB SSD
 

Attachments

  • thunderboltssd.jpg
    thunderboltssd.jpg
    63.9 KB · Views: 161
Get 330 MB/s Write and 364 MB/s Read speeds on my Seagate Thunderbolt adapter connected to a Samsung 830 series 256GB SSD

I see people getting near 500mb/s read on internal sata interface, surprised to see such a difference compared to thunderbolt.
 
I see people getting near 500mb/s read on internal sata interface, surprised to see such a difference compared to thunderbolt.

My thoughts exactly. I wonder why? TB was supposed to have no overhead (unlike USB) to make transfers as fast as internal connectors.
 
I see people getting near 500mb/s read on internal sata interface, surprised to see such a difference compared to thunderbolt.

Perhaps , the Seagate implementation of Thunderbolt in this adapter is a bit slower. I don't mind it actually as it requires no power supply.

And the speeds are crazy fast compared to ****** USB 2.0 LOL :D
 
Anyone reporting long-term success using a Crucial M4 256GB on the GoFlex STAE121?

plucky duck: Thanks for posting that deal! Just wish I'd pounced on it :)
 
Hi - I'm in UK and don't seem to be able to source the GoFlex STAE121 ... I am interested in this set-up as well as a cache drive on my iMac mid 2011 for After Effects coupled with a Crucial M4:

http://www.ebuyer.com/266526-crucial-128gb-m4-ssd-ct128m4ssd2?utm_source=google&utm_medium=products


... and a cable:

http://www.amazon.co.uk/Apple-Thunderbolt-cable-mini-DisplayPort/dp/B0058KHRPS/ref=sr_1_2?ie=UTF8&qid=1342095412&sr=8-2


... anyone any thoughts on this and whether I should continue looking for the GoFlex STAE121 or try an alternative ... any advice welcome. Regards Roger
 
As an Amazon Associate, MacRumors earns a commission from qualifying purchases made through links in this post.
Perhaps , the Seagate implementation of Thunderbolt in this adapter is a bit slower. I don't mind it actually as it requires no power supply.

And the speeds are crazy fast compared to ****** USB 2.0 LOL :D

The adapter is only SATA II. Internal is SATA III.

And anyway, the massive speed boost you get from running an SSD (as opposed to a hard disk) is in the random read performance (because the access times and max iops are so dramatically improved). None of this improvement is lost using the Seagate adapter - it's only the max sequential read which takes a small hit, and in practise that hardly matters at all.

The net result is that my Mac Mini with the SSD now installed internally is not even *slightly* faster than when the SSD was external on the Seagate adapter. No difference in real-world performance whatsoever.
 
The adapter is only SATA II. Internal is SATA III.

And anyway, the massive speed boost you get from running an SSD (as opposed to a hard disk) is in the random read performance (because the access times and max iops are so dramatically improved). None of this improvement is lost using the Seagate adapter - it's only the max sequential read which takes a small hit, and in practise that hardly matters at all.

The net result is that my Mac Mini with the SSD now installed internally is not even *slightly* faster than when the SSD was external on the Seagate adapter. No difference in real-world performance whatsoever.

not correct the seagate t-bolt is sata iii . it has latency due to the 2 chips needed at each end of the cable. it can push 380 mbs read write on a long clone if the mac mini has a ssd mounted on the inside.

the rest of your statement is correct.


to the op the internal hdd is slow with reads and writes at the random 4k level t-bold ssds are 20 to 25x faster.
 
not correct the seagate t-bolt is sata iii . it has latency due to the 2 chips needed at each end of the cable. it can push 380 mbs read write on a long clone if the mac mini has a ssd mounted on the inside.

the rest of your statement is correct.


to the op the internal hdd is slow with reads and writes at the random 4k level t-bold ssds are 20 to 25x faster.

To further philipma's comments,,, Thunderbolt is an exposed PCIe lane. Thus it requires a Thunderbolt controller at each end, plus a controller for the device connecting to Thunderbolt. In this case a SATA III controller. Its well known that Intel Sandy Bridge and now Ivy Bridge SATA III controllers are the fastest on the market. Marvell and others are far slower. That's the real difference between the Seagate GoFlex Thunderbolt adapter and an internal drive. Even when these other SATA III controllers are integrated into ATX motherboards, they are slower and prone to performance issues.
 
not correct the seagate t-bolt is sata iii . it has latency due to the 2 chips needed at each end of the cable.

Regards SATA III - my apologies, I stand corrected.

However, the slower performance over TB has nothing to do with "the chips at the end". The TB interface supports 2 x 10Gb/s bi-directionally, even through the "chips at the end". Just using 1 10Gb channel means it would max out at over 1Gbyte/s not 380MB/s.

380MB/s is clearly slower than the 500MB/s+ you get if you install the drive internally so this must be down to the slower performance of the Seagate controller.
 
Last edited:
I have 3 different t bolt devices all have latency when compared to a direct connection inside the mini.

the cable is part of the reason it is simply more work to be done then a direct connection. Not disagreeing that the controller in the adapter is part of the latency cause. It is as are the two chips in the cable and the interface in the mini to the t-bolt jack.

basically 4 'chips' and some wire vs 1 chip and no wire. The speed delays are real.

but the trade off is portable osx, no warranty risk to your mac and instant replacement if the t-bolt breaks (and your choose to have a spare).
 
I have 3 different t bolt devices all have latency when compared to a direct connection inside the mini.

the cable is part of the reason it is simply more work to be done then a direct connection. Not disagreeing that the controller in the adapter is part of the latency cause. It is as are the two chips in the cable and the interface in the mini to the t-bolt jack.

basically 4 'chips' and some wire vs 1 chip and no wire. The speed delays are real.

but the trade off is portable osx, no warranty risk to your mac and instant replacement if the t-bolt breaks (and your choose to have a spare).

I think you use the word "latency" too much. Latency and throughput are not the same thing!

When you say "I have 3 different t bolt devices all have latency when compared to a direct connection inside the mini."

I think you mean "I have 3 different t bolt devices all are slower when compared to a direct connection inside the mini."

Just because they are slower does not mean latency issues.

Dictionary:

la·ten·cy   [leyt-n-see] Show IPA
noun, plural la·ten·cies.
1. the state of being latent.
2. Computers . the time required to locate the first bit or character in a storage location, expressed as access time minus word time.
3. latent period.
 
slower works for me. all t-bolt devices I own are 10 to 15 percent slower then a direct connection to my minis sata bus.

Having got the semantics out of the way, that's an interesting comment. And a rather surprising and disappointing one in fact. If a 20Gbit bandwidth connection cannot keep up with a 600Mbit interface (at best) then that is concerning.
 
Having got the semantics out of the way, that's an interesting comment. And a rather surprising and disappointing one in fact. If a 20Gbit bandwidth connection cannot keep up with a 600Mbit interface (at best) then that is concerning.

well that is on 1 drive so if i run my pegasus r6 with 6 drives as jbod they are slower ,

but if I raid them they are smoking fast. I run it with 2 ssds in raid0 2 two 2tb hdds in raid0 and two 2tb hdds as jbod.


the raid 0 of 4tb holds movies.

the raid0 of 2 ssds is my osx.

the jbod of 1 2tb disk is a time machine the

2tb jbod of the other 2tb is a clone back up.

1 wire 6 hdds and if i am doing a few jobs at the same time I can read write at 700mbs.

the pegasus r6 lights up and flashes like crazy.

my mini is tossing around a lot of data in conjunction with the pegasus r6. I got very lucky with the pegasus r6. I purchased it new on sale for 1320 with 6 1tb hdds. the day it came to my house the western digital plant flooded I sold the 6 1tb hdds for 750 bucks. so I got the case for under 600. i had a few older 2tb drives on hand and added 4 of them (cost of 400)

and I got 2 big ssds on sale (cost of 600). so my setup on paper is more then 2500 and I paid 1600.

right now 1600 does not get a new pegasus r6 with 6 1tb hdds it is 1799 at apple

http://store.apple.com/us/product/H5186VC/A/Thunderbolt?


no less my case with 4 2tb hdds and 2 500gb ssds at just about 1600.

same goes for the lacie lbd it will run 2 ssds at raid0 very quickly I run 2 of them with 3 256gb ssds raid0 I get 640 mbs read write, even though each ssd is at sata II speed.


These units are good I really like running the lacie lbd's with 2 of these



ssd's



http://www.ebay.com/itm/Samsung-MZ7...Disk_Drives&hash=item2a1fa7caa9#ht_884wt_1160


the ssds are about 390


the little big disk is 235 the seller is legit


http://www.ebay.com/itm/LaCie-90001...isk_Drives&hash=item3a7418b294#ht_1392wt_1393
 
Elgato Cable and Seagate Thunderbolt Adapter

yeah big transfers can be a problem and there seems to be some cable issues on apples part. I went back to the lacie little big disks. I also sit the gear on a laptop cooler to manage heat better.

Hi,

Figured I'd check back in on this. I replaced my Apple Thunderbolt cable with the short Elgato Thunderbolt cable and have yet to have any more disconnect issues. It's been plenty hot in the office where the machine is, and I've done a couple of large (100gb+) transfers with no issues. This is with a 27" iMac, STAE121, Elgato Thunderbolt cable, and Corsair Force 240GT SSD running Lion.

Here's to hoping it lasts!

Frank
 
Hi,

Figured I'd check back in on this. I replaced my Apple Thunderbolt cable with the short Elgato Thunderbolt cable and have yet to have any more disconnect issues. It's been plenty hot in the office where the machine is, and I've done a couple of large (100gb+) transfers with no issues. This is with a 27" iMac, STAE121, Elgato Thunderbolt cable, and Corsair Force 240GT SSD running Lion.

Here's to hoping it lasts!

Frank

Same here. The Elgato Thunderbolt cable fixed the 480GB and 512GB SATA III drive disconnects as well.
 
Hi,

Figured I'd check back in on this. I replaced my Apple Thunderbolt cable with the short Elgato Thunderbolt cable and have yet to have any more disconnect issues. It's been plenty hot in the office where the machine is, and I've done a couple of large (100gb+) transfers with no issues. This is with a 27" iMac, STAE121, Elgato Thunderbolt cable, and Corsair Force 240GT SSD running Lion.

Here's to hoping it lasts!

Frank

Same here. The Elgato Thunderbolt cable fixed the 480GB and 512GB SATA III drive disconnects as well.

good I am glad we have 2 with better results using the short elgato cable.


I am running;

mini
lbd
lbd
seagate


the pair of lbd's have a 768gb raid0 = osx

the seagate has a 1tb toshiba = clone of osx


all using apple cables this is working well
 
Does anyone have any comparison of the Goflex Desk and the portable one?

I am considering this setup for an iMac and am wondering if the desk version is worth double the price.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.