Samsung's Portable SSD T1 Tested

MojoKid writes The bulk of today's high-capacity external storage devices still rely on mechanical hard disk drives with spinning media and other delicate parts. Solid state drives are much faster and less susceptible to damage from vibration, of course. That being the case, Samsung saw an opportunity to capitalize on a market segment that hasn't seen enough development it seems--external SSDs. There are already external storage devices that use full-sized SSDs, but Samsung's new Portable SSD T1 is more akin to a thumb drive, only a little wider and typically much faster. Utilizing Samsung's 3D Vertical NAND (V-NAND) technology and a SuperSpeed USB 3.0 interface, the Portable SSD T1 redlines at up to 450MB/s when reading or writing data sequentially, claims Samsung. For random read and write activities, Samsung rates the drive at up to 8,000 IOPS and 21,000 IOPS, respectively. Pricing is more in-line with high-performance standalone SSDs, with this 1TB model reviewed here arriving at about $579. In testing, the drive did live up to its performance and bandwidth claims as well.

Comments

no

[TheGratefulNet]

regular ssd, usb3 interface, UASP (scsi over usb, new standard) and you have all the speed of native sata (that the drives can put out) and are still vendor neutral.

I try to avoid samsung products these days. after the fiasco with the evo drives, I'll look for another vendor.

and then there is always the worry that samsung will insert commercials between disk block seeks (inside joke, sorry if that does not make immediate sense to you).

Re:no

[ihtoit]

I'm more concerned with Samsung uploading an incremental mirror of your hard drive to its cloud.

(ditto inside joke thing).

Re:no

[ihtoit]

even further back. I have three Adaptec SCSI-to-USB adapters - actually physical pin-compatibility adapters. I've had those since probably 2005 or even before. They'll mount on pretty much anything I plug them into, from Windows ME through 7, OSX from Tiger/PPC (the one I've tried it on), and several flavours of Linux from around Knoppix 5.1.1 and I can still read every hard drive I still own from a 10MB 40-pin Winchester through the pile of 500GB Deskstars, several Seagate 9.1GB UW ans a good few 50-pin random and various capacity drives - not forgetting of course, the takep drives, slot loading and cassette DVD/R/RW/RAM drives and my pride and joy of MO gear that still works: a custom cased LS120/Zip100 triple threat (it reads 3.5" floppies, too!). All USB mass storage is really just SCSI layer on the USB stack.

NSA Backdoor preinstalled?

[Anonymous Coward]

NSA Backdoor preinstalled?

Danger of SSDs

[bradgoodman]

for a ton of technical reasons I won't get into right now (remapping/wear leveling) SSDs aren't usually able to handle power faults like regular HDs. Too often, taking an unexpected power hit can easily result of massive amounts of lost data, or even loss of the device itself. I've seen this happen at least 20 times. Thete are allegedly some "enterprise grade SSDs" which may or may not mitigate this issue. I'm tired of seeing articles citing all kinds of performance tests that go into absolutely no detail on if you are going to lose all your data the next time you lose power, or have to force-off your laptop because it locked-up on you.

Re: Danger of SSDs

[dpidcoe]

I thought this was the reason a lot of SSDs now have a collection of capacitors to finish out the writes with in the event of a power loss?

Re:Danger of SSDs

[mlts]

I wonder if the latest generation of filesystems like ZFS, btrfs, and ReFS would be useful, so a corrupt file that wasn't completely written would be detected by the FS during a background scrub or garbage collection task. With RAID-Z, the corruption can be found. Z2, the corruption likely can be fixed.

Re: Danger of SSDs

[Bengie]

SSDs with "power loss protection" store enough power to write out all of their cache, which is something like 1GB now days. Like we've mentioned, we don't care about caches not being flushed, but how to the internal mapping tables hold up without "power loss protected". My hope would be that modern controllers can handle keeping internal state and just screw the data in cache.

I was reading about Samsung's "RAPID Mode" that uses system memory as a write cache to speed up writes to the SSD. One of the topics about "RAPID Mode", which is even more sensitive to power loss because of increase caching, is that it handles power loss "well". They have done extensive testing with "RAPID Mode" and power loss. I figure if they can offer 10 year warranties and feel confident about these issues, I'll trust them until proven otherwise. They have a great track record. I still wouldn't put all of my eggs in one basket.

People are looking at the wrong specs

[Solandri]

The problem is that MB/s is the inverse of the figure we're actually interested in. When you're sitting at a computer waiting for it to finish doing something, you measure it in seconds. But HDD manufacturers chose to rate their drives in terms of disk speeds (MB/s), rather than wait times (sec/MB)

The tl;dr version if you want to skip everything below is that because MB/s is the inverse, the bigger the MB/s gets, the less difference it makes. Let me repeat because it's so counter-intuitive: The bigger the MB/s gets, the less difference it makes.

There are two effects going on here. First , because MB/s is the inverse of what you're really interested in (sec), as it gets bigger its benefit to wait times is actually becoming smaller. Imagine you need to read 1000 MB of data.

• 10 sec = 100 MB/s HDD
• 4 sec = 250 MB/s SATA2 SSD; a 150 MB/s increase yields a 6 sec time savings
• 2 sec = 500 MB/s SATA3 SSD; a 250 MB/s increase yields a 2 sec time savings
• 1 sec = 1000 MB/s PCIe SSD; a 500 MB/s increase yields a 1 sec time savings
• 0.1 sec = 10 GB/s 20-drive RAID 0 SSD; a 9000 MB/s increase yields a 0.9 sec time savings

Or in terms of time savings per constant MB/s:

• 4 sec saved per 100 MB/s increase (100 -> 250 MB/s, HDD to SATA2)
• 0.8 sec saved per 100 MB/s increase (250 -> 500 MB/s, SATA2 to SATA3)
• 0.2 sec saved per 100 MB/s increase (500 -> 1000 MB/s, SATA3 to PCIe)
• 0.01 sec saved per 100 MB/s increase (1000 MB/s -> 10 GB/s, PCIe to something ridiculous)

See how the bigger the drive's MB/s gets, the less time savings you get per 100 MB/s increase? That's because MB/s the inverse of what we're really interested in - sec/MB. Beyond about SATA3 or even SATA2 speeds, you've pretty much passed the point of diminishing returns and there's very little meaningful time saved by going faster (for tasks
The second effect has to do with which disk operations take more time. Everyone looks at sequential read/write speeds, when they should actually be concentrating on the 4k random speeds. Why? Say your SATA3 drive gets 500 MB/s sequential speeds, and 50 MB/s 4k speeds. Because the 500 MB/s figure is 10x bigger than the 50 MB/s figure, people assume that means it's 10x more important.. In fact, the opposite is true - the smaller MB/s figure matters 10x more. Imagine this drive has to read 1000 MB of sequential data + 1000 MB of 4k data.

• 2 sec = 1000 MB sequential data @ 500 MB/s
• 20 sec = 1000 MB 4k data @ 50 MB/s
• 22 sec = total read time

Surprise! Your drive spends 91% of its time doing the 4k read task, and only 9% of its time doing the sequential read task. Consequently, the smaller MB/s figure matters a lot more in terms of how long you wait for the drive to finish doing something. Which is just another way of saying: The bigger the MB/s gets, the less difference it makes.

So why do HDD manufacturers and review sites insist on using MB/s? Marketing. Because it exaggerates the impact of faster drives, it makes you want to pay more money for the latest and greatest drive. The jump from 500 MB/s to 1000 MB/s by going to PCIe SSDs seems like a huge increase, when in fact it's a miniscule tiptoe. And review sites are complicit because if word got out that there's really not much difference between using a budget SSD and the latest and greatest PCIe SSD, people would stop reading reviews and they'd lose advertising revenue. (That's not to say this is always true - certain specialized tasks benefit disproportionately from the high sequential speeds on large files. Real time video editing is a good example.)

So if you want a fast SSD, you should be obsessing over the smallest MB/s figures, not the biggest ones; the 4k speeds, not the sequential speeds. 10 MB/s faster 4k speeds makes about 10x as much difference as 100 MB/s faster sequential speeds. (The same thing is true for car MPG. MPG i