I don't think there's anything particularly noteworthy here other than old fashioned hard drives are slow.
I dont have $700 for a 1TB SSD though
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
nMP Disk Speed Test
- Thread starter rasputin666
- Start date
- Sort by reaction score
I dont have $700 for a 1TB SSD though
very tempting, links? I am getting ready to close on another project and that be a nice way to spend the deposit. I was also using samsungs evo 840 as a price point.
As an Amazon Associate, MacRumors earns a commission from qualifying purchases made through links in this post.
CalDigit T3
I just got a 3GB T3 yesterday and I am very happy with the speed results:
Using the blackmagic test I am getting 511.8 Write and 492.8 Read.
The only problem is that with the nMP there is no noise. When I am running the T3 Raid it is loud enough to hear anywhere in my suite.
I just got a 3GB T3 yesterday and I am very happy with the speed results:
Using the blackmagic test I am getting 511.8 Write and 492.8 Read.
The only problem is that with the nMP there is no noise. When I am running the T3 Raid it is loud enough to hear anywhere in my suite.
i know...this is the problem now. any enclosure is noisier than the machine itself.
i'm running all SSDs so thinking about completely disabling fans.
A single mechanical hard drive isn't going to show much difference between a TB enclosure and a USB 3 enclosure. It's only going to be as fast as the mechanical drive can be - which is much slower than either TB or USB 3 can handle.
Run 4+ fast drives in a RAID 0 and you can potentially saturate a USB 3.0 bus in the right enclosure, though likely not in the Qx2.
I also have a Qx2 (two, actually) attached to a LaCie eSATA hub and I get about the same speeds as you. It's about all we can expect out of a single eSATA channel I think. I got the same speeds off the Qx2 when it was connected directly via eSATA on my old Mac Pro 1,1. It's no slower via the LaCie TB-eSATA hub. No faster, either, but no slower.
Either way, it's faster than the other Qx2 connection options (Firewire 800 and UBS 2.0 for the older model Qx2's).
i've been impressed with Qx2 via Lacie hub. At first i had an issue but i resolved that because of my RAID setup. My other issue, from above, is my STAE129 adapter. The 3TB drive in that Thunderbolt adapter is getting in the 80's for Read/Write. and i put multiple drives in there. I have a SSD drive in my 3,1 MP and i'm moving data around to connect that to that adapter to see if SSD gets what it should. I've put 4 different 7200 RPM drives in that adapter and they are all well under 100 MB/s, usually around 80. From what i've seen, that's underperforming.
Curious if anybody has experience with the newer Qx2 with USB 3.0 to a nMP
I know OWC released their TB version this week which has no built in RAID controllers like the Qx2.
Like you said with 7200 spinners and USB 3.0, adding TB doesn't bring much to the table unless you have a bunch of USB goodies vs TB goodies and you happen to be saturating one protocol over the other.
(I have the older fw800/USB2/estata Qx2 Raid 5 (4-3TB disks) via eSata to a MP2.1 or to the LaCie TB hub which I connect to my 2012 MBP also (waiting for February to end like a lot of people) Thinking of using the old Qx2 for backups and replace with a new one but get the USB 3.0 version vs the TB version.
I have the new Qx2 (plus 2 older ones) and am running it on a 5,1 2012 Mac Pro. I didn't notice much difference running from USB 3.0 vs. eSATA (both from CalDigit card in Mac Pro). I do like that when the computer sleeps, the new Qx2 also sleeps.
OWC also sells a NewerTech "USB 3.0-to-eSATA adapter" which works with the older Qx2 RAID systems, but one wouldn't expect any speed increase from using that.
-howard
Akitio/Monoprice 2.5" dual drive Thunderbolt enclosure
My Akitio 2.5" dual drive Thunderbolt enclosure was delivered today.
I installed my two 250GB Samsung 840 EVO drives in to the enclosure (pulled them out of the OWC Mercury mini USB 3.0 cases). The Akitio drive case came with an approximately 18" Thunderbolt cable.
Attached is a Blackmagic speed test of the drives in the TB case running OS X's software RAID 0. The speed test of the previous dual-channel USB 3.0, dual-enclosure RAID 0 is attached to an earlier post.
Read speed is very slightly slower than the two-channel USB 3.0 RAID 0 I had been using (734.5MB/sec on TB versus 770.3MB/sec via dual USB 3.0). Write speed is measurably slower (514.0MB/sec on Thunderbolt versus 732.0MB/sec on dual-channel USB 3.0).
I'm a little disappointed that the write speed isn't faster, but really I'm splitting hairs here, really. For the work I do, the real-world difference will negligible to none. And I gain back two precious USB 3.0 bus channels, too. The drive is also at the end of the TB chain, plugged in to the outbound port of my LaCie eSATA hub, though I expect that makes little difference in performance.
The case itself is quite small and compact, much smaller than I expected it to be, and seemingly of sturdy quality. Stylistically it looks almost exactly like a (much) smaller OWC Qx2. Same kind of perforated PowerMac G5/Mac Pro style aluminum front panel and two engraved stripes running down the bottom side edges of the case, just like the Qx2's. And virtually the same back printed front panel iconography and LEDs as the new version of the Qx2. I would't be surprised if someone told me they came from the same Asian OEM.
Each drive slides in on a set of removable rails. I cannot hear the very small fan the Akitio enclosure has unless I put my ear right to the unit. But then again, it's sitting directly on top of a pair of Qx2's with eight 7200RPM drives between the two of them, so whatever noise the Akitio makes is going to be totally drowned out anyway.
The power supply is an external brick. This might bother some people who want to use a unit like this as a portable drive, though makes no difference for me, and I'm sure it helps keep the size and heat of the unit down to a minimum. The unit is also nice in that it has a second TB port for daisy chaining additional devices, unlike the portable TB drives made by Buffalo and LaCie.
I'll try to post a couple pictures later tonight.
So far, I give it two thumbs up.
My Akitio 2.5" dual drive Thunderbolt enclosure was delivered today.
I installed my two 250GB Samsung 840 EVO drives in to the enclosure (pulled them out of the OWC Mercury mini USB 3.0 cases). The Akitio drive case came with an approximately 18" Thunderbolt cable.
Attached is a Blackmagic speed test of the drives in the TB case running OS X's software RAID 0. The speed test of the previous dual-channel USB 3.0, dual-enclosure RAID 0 is attached to an earlier post.
Read speed is very slightly slower than the two-channel USB 3.0 RAID 0 I had been using (734.5MB/sec on TB versus 770.3MB/sec via dual USB 3.0). Write speed is measurably slower (514.0MB/sec on Thunderbolt versus 732.0MB/sec on dual-channel USB 3.0).
I'm a little disappointed that the write speed isn't faster, but really I'm splitting hairs here, really. For the work I do, the real-world difference will negligible to none. And I gain back two precious USB 3.0 bus channels, too. The drive is also at the end of the TB chain, plugged in to the outbound port of my LaCie eSATA hub, though I expect that makes little difference in performance.
The case itself is quite small and compact, much smaller than I expected it to be, and seemingly of sturdy quality. Stylistically it looks almost exactly like a (much) smaller OWC Qx2. Same kind of perforated PowerMac G5/Mac Pro style aluminum front panel and two engraved stripes running down the bottom side edges of the case, just like the Qx2's. And virtually the same back printed front panel iconography and LEDs as the new version of the Qx2. I would't be surprised if someone told me they came from the same Asian OEM.
Each drive slides in on a set of removable rails. I cannot hear the very small fan the Akitio enclosure has unless I put my ear right to the unit. But then again, it's sitting directly on top of a pair of Qx2's with eight 7200RPM drives between the two of them, so whatever noise the Akitio makes is going to be totally drowned out anyway.
The power supply is an external brick. This might bother some people who want to use a unit like this as a portable drive, though makes no difference for me, and I'm sure it helps keep the size and heat of the unit down to a minimum. The unit is also nice in that it has a second TB port for daisy chaining additional devices, unlike the portable TB drives made by Buffalo and LaCie.
I'll try to post a couple pictures later tonight.
So far, I give it two thumbs up.
Attachments
so far, i have tested every conceivable option. different ports. different cables. with other drives connected. with all other connections removed. i have a SSD drive in my 3,1 MP that i am moving the data from so i can attach it to the thunderbolt adapter and see if that shows what it should. Seagate email support has been responsive but claiming 'that's an expected range of transfer speed'.....i emailed them some screen shots of USB2, USB3, TB tests with multiple drives and scenarios. can't dispute the data. main goal is to determine if it's an adapter issue or a nMP issue, TB bus or something.
----------
That is the one thing i have learned moving from my 3,1 MP to the nMP, Thunderbolt is the shiny toy and TB2 has people buzzing, but unless you're using non-mechanical drives, not much difference from USB3 [obviously that will change when TB2 is available to the masses]. I'd be happy if i had consistent speeds of 120 MB/s, my issue is most times i test, my thunderbolt drive is in the 80's or maybe right at 90. From everything i've read, TB and USB3 should be 3 x of USB2, which has been consistently around 40 for me, so 120 MB/s should be the baseline for a TB drive with a non-flash drive attached. I am just not seeing that right now from my gear.
I recall that your SeaGate GoFlex Thunderbolt adapter is the powered desktop version and uses 7200rpm 3.5" disk drives.
Just for comparison, here is a test from the portable SeaGate GoFlex Thunderbolt bus-powered adapter:
1) 105/102 with standard GoFlex 1TB 5400rpm 2.5" hard disk
2) 260/386 with Crucial M4 SSD 256GB
(I discovered that the 512GB M4 SSD draws too much power for the GoFLex adapter and "bricks" the drive ... 2 different drives, 2 different adapters)
-howard
.
Just for comparison, here is a test from the portable SeaGate GoFlex Thunderbolt bus-powered adapter:
1) 105/102 with standard GoFlex 1TB 5400rpm 2.5" hard disk
2) 260/386 with Crucial M4 SSD 256GB
(I discovered that the 512GB M4 SSD draws too much power for the GoFLex adapter and "bricks" the drive ... 2 different drives, 2 different adapters)
-howard
.
Attachments
thunderbolt does not magically take a slow drive and make it faster. a mechanical HDD is limited by its own interface, so neither USB3 nor TB (1 or 2) would be a limiter.
when you have faster drives and/or combine HDDs in a RAID array there will be a difference...with a good enclosure.
tested a few enclosures--my nMP wrang a bit more performance out of drives connected to a TB1 enclosure (vs a 2013 iMac).
tested with a few enclosures and results were the same. i was surprised to see a boost with a TB1 device.
tested with a few enclosures and results were the same. i was surprised to see a boost with a TB1 device.
Nice. I've done some more testing and have concluded only one thing. I HAVE REALLY SLOW EXTERNAL HARD DRIVES! i thought my TB adapters were having issues [they do get very very warm] but i just pulled my 128GB SSD from my 3,1 MP, put it in both my GoFlex portable TB adapter [STAE121] and my GoFlex Desk TB adapter [STAE129] and speed tests were at or over 300 MB/s. Conclusions? my externals must be all 5400 RPM drives? actually wouldn't mind finding that out. So for now, since i bought the nMP, i'll have to take the wife on vacation as pay-back, so faster drives will wait. I may wait till TB2, thoughts on timing of widespread TB2 drives? 12-18 months? longer? Thx to all for the dialogue, awesome forum.
It isn't the Thunderbolt 2 interface that is the problem, it is the drives that you connect on it that are often the speed bottleneck.
But Intel are having a hard time trying to get Thunderbolt adopted on the PC so prices will still remain much higher than say USB3 devices.
So for most it goes like this...
1. Buy a Thunderbolt 2 compatible enclosure that will suit your data rate demands.
e.g
- Single drive enclosure
- Dual drive enclosure
- PCI-e based expansion enclosures
Or consider daisy chaining. e.g. Linking 2 x single drive enclosures and setting them up as raid 0.
You can find products already available that do all the above.
2. Choose a drive setup.
From cheap to expensive:
5400 RPM drive = slow
Best use for low power demands (e.g. laptops) and for non critical work loads like archives and backups, photos and music.
7200 RPM drive = below average
Best used for standard office type work loads, but have been used for years as acceptable OS drives until a few years ago.
SSD drive = above average
Best used for demanding workloads, video editing and currently the preferred way for OS installs due to the overall speed increases from SSD. SSD also has no moving parts so should be more reliable than 5400 and 7200 platter drives.
PCI-e SSD = fast
Best used for realtime 4K/6K video editing or very demanding workloads.
EDIT: You can also raid 0 the above by buying 2 of them which often gives 80%+ speed boost.
So correct me if I am wrong and without going into too much technical discussion on drive reliability and interfaces, that is how it stands from my understanding at this point in time.
Hope I have been informative.
Anim
Last edited:
Ooops, for some reason i was thinking about raid 1. Must have been looking at my nas box when i wrote that. Thanks for the catch.