Thunderbolt and pricing,
Not really enticing me.
Galileo, Galileo, etc
Not so long ago, El Reg was pulling apart, in every sense of the phrase, Elgato’s Thunderbolt Drive+ SSD. A portable 512GB storage device with USB 3.0 and Thunderbolt interfacing. It put in a nifty performance, for sure, but man, it was pricey. GTech GDrive Mobile with Thunderbolt GTech's GDrive Mobile with Thunderbolt: 1TB …
Thunderbolt and pricing,
Not really enticing me.
Galileo, Galileo, etc
Is there something about Thunderbolt and pricing that you find particularly un-enticing, or do you habitually share your views about articles you find inapplicable?
I quite like the idea of signing onto a "Dancing with the Stars" board, and contributing "Didn't watch that episode" to every discussion, but it might eat into my day a bit.
I thought the original post was A Kind of Magic and didn't require any Back Chat.
But, Now I'm Here and you appear to be waiting for the other Hammer to Fall, might I suggest a Night at the Opera to provide the Flash of inspiration to Play the Game in future.
It's un-enticing because its so fucking expensive. Put down the gold leaf toilet paper and realise not everyone can pay 300% the market rate for a hard drive, especially when you usually need to buy 2. Once again Apple has taken a technology that already had a huge Intel-markup on it and put the huge Apple-markup on top of that. Oh good now if I sell my house I can buy like 5 thunderbolt drives with the money.... such good value for money, no?
When I can get a 1TB USB 3.0 drive for £49.99 on the high street I just don't see the reason to buy a thunderbolt drive. Certainly not for consumer use in any case!
Thunderbolt makes much more sense when you are using SSDs or arrays - it's technically superior to USB 3 but it's not cost competitive for low end devices.
Thunderbolt has some uses, it makes a great universal docking station connector for one.
Using it as an interface for a single portable hard drive via a SATA 3 interface is just stupid. In the future maybe we'll see some sort of PCI-E SSD connected over thunderbolt which would actually take advantage of the speed but ultimately very few people are going to benefit so the price will remain high.
In my case, Thunderbolt could be useful. The huge HP laptop I have does have a ESATA port, but the Macbook pro does not, so a SATA interface doesn't help there.
However, the problem with this device is having to UNPLUG the monitor to use it. FAIL, fail and fail again. If I have to choose between monitor and external disk, monitor wins every time.
To be fair - isnt it more realistic to expect the monitor to provide daisychained thunderbolt ports?
At the least its an equal fail.
Not if you are using a Mini Displayport to HDMI adapter for your monitor, or you connecting to a Displayport monitor. There are very few thunderbolt monitors around other than the model supplied by Apple.
Yes, it's a DisplayPort monitor. It's NOT a Thunderbolt monitor - it has a 17Gbit Display port input. It's the Thunderbolt peripheral that has to detect that there's a displayport down-stream and regenerate the 17Gbit DisplayPort data stream for it by demuxing it from the 10Gbit Thunderbolt datastream.
Looking at the simplistic benchmark screenshots there is no difference between USB3.0 and Thunderbolt, so in effect the money is wasted. Right?
Very happy to be corrected on this but isn't it the WAY data is transferred that is the difference? I tried using a USB3 drive with Final Cut Pro and FCP was not happy.
However, an external Firewire 800 drive + FCP and all's well. On paper, FW800 is a hell of a lot slower than USB3. They are different technologies designed for different uses.
And I still don't have a single Thunderbolt peripheral. I want one of these; it should improve the performance of Aperture (library currently on a 1TB USB2 disk) quite nicely.
Yes it would be much faster than a USB 2 drive but to really see a big difference pair it with a SSD drive.
NFG written on the case stands for "No F**king Good", which is the status of the drive after he destroyed it. It should never be that hard to take anything apart that's made of plastic. Always check the stickers!!
And here, I was taught that it stood for "No Factory Guarantee".
Having never heard of "No Factory Guarantee" my first thought as to what NFG stood for was actually "No F*cking Good". Perhaps its just the way some people are wired? :)
Can you do a review of this as well? You can swap the drive without assaulting it...
I have a couple of them and swap between Firewire 800 and USB 3 adapters as appropriate. I don't have anything to plug a thunderbolt adapter into, as my MacBook is the one that is Mini-Display port only with no Thunderbolt capabilities.
When I'm next in the market for an external drive, it will probably be another Go-Flex.
Wonderful video. Do you think he voided the warranty?
Every demolisher knows there is always a hidden screw under the label, because the warranty is void if you remove or damage the label.
BTW I always use a blunt Stanley knife to open the box. It is safer and more effective than a screwdriver.
See http://www.youtube.com/watch?v=iiyGVvA6KfU if you want to see blood.
Wow - that was painful to watch. Almost like an electronics horror movie. At one point when he went away I was waiting for him to come back with a dremel tool and a hammer. Might have been less painful all around. lol
this disassembler guy must have sated the blood lust of all his tools* very recently. i kept thinking as i watched "there will be blood."
* all tools demand a blood sacrifice . . . sooner or later
There are a lot of portable drives on the market, some bigger, some just as fast. Unless I'm missing something here, this is nothing that special. Why bother devoting an article to it? It's just a rather boring mechanical drive in a not so special case with a couple of slightly interesting interface choices. So what?
Thunderbolt does for PCI Express and DisplayPort buses what eSATA does for SATA: It's basically PCIe and DisplayPort on a wire, so you no longer need to provide space inside a computer's case for traditional expansion slots. (Hence the recent Mac Pro redesign.)
For external storage, PCIe offers another advantage: it operates entirely independently of the CPU. USB keeps its price low by making the CPU to do much of the heavy lifting, taking valuable processing power away from your applications.
What the benchmarks in the article don't show – and they should – is the additional processing overheads imposed by using USB. Yes: the raw speed looks identical, but if you're having to give up a CPU core to achieve it, it's going to brutally hammer the performance of any high-end video editing suite you're using at the same time. Even Aperture and Lightroom will be noticeably more sluggish.
Trust me: if you're working in a field where processing large lumps of media is a core activity, you will care about this. It's why Apple redesigned their Mac Pros the way they did: that machine has the equivalent of 18 PCIe expansion slots. (Or 15 + 3 x 4K displays if you prefer.) All that's changed is that those slots are now on the outside of the machine, allowing the engineers to optimise the hell out of the arrangement of the core components inside the case.
That is what Thunderbolt can do that USB cannot. USB isn't even playing the same game, let alone in the same league.
Thunderbolt comes into its own in the high-end professional markets, where the cost of the actual computer itself is tiny compared to all the storage and other peripherals you need to connect to it. No, most readers here won't need that level of power, but it most definitely has a market.
Just a quick point here (my desktop PC has no 3.0 ports, and I am running everything from a USB 2.0 memory stick, you may fear the lag, but I have been working in it!). You are talking about the CPU waiting for IO things and dealing with reads and writes directly.
This is why we have DMA controllers, and living without one makes me miss it, rather than the CPU dealing with stuff it goes "DMA controller, interrupt me when it is done, write this much here starting from here" and it does! Now the PCIe bus can talk to RAM (I remember reading a hardware hack that could dump the first 4gb of ram, interesting stuff) which is why PCIe stuff is great at this, it can do it without the CPU's intervention at the cost of security (it's on a system bus!)
That aside, my computer is FINE until something starts thrashing and the cores spend most of their time waiting for I/O. Usually this is Firefox going "Another GB please" and the resulting swap activity is what I am waiting for.
BUT Wait time is time the CPU can do other stuff in! It is not actually halted, the reason my system freezes for a... I don't want to admit how long is because the SCHEDULER has buggered up, gone are the days of the CPU literally waiting for something to complete. IOWait now means (I believe from experience) time spent waiting where the system was actually waiting and had nothing else it could do because everything was waiting for data.
So how CPU intensive is it actually going to be? Not very. Our critical sections are small, and we have stacks that can actually fit things in them.
So yeah... I want my USB bus not on my system bus, and also I don't really mind the few 1000 operations (maybe? It wont be very many) my CPU spends doing them. BUT I can tell you, running an entire system from USB 2.0 sucks. Please send help.
(Another good example for this but slightly higher up is MySQL from a few years ago (I've been trying to think of the exact thing) but basically a really stupid scheduler was like "I'm going to flush EVERYTHING! NAOW!" and it did. This would cripple the server for a while. A smarter scheduler doesn't do this. The same problem (silly scheduler) is why my system freezes when Firefox or Thunderbird (the two big ones) want more RAM, the OS makes some poor choices and gets stuck waiting)
I wonder how many people here bothered to click on the 'Show More' under the into, to discover the category is Comedy!
Should have waited a few days for April 1st, some of you might have twigged!
fscked by SHA-1 collision? Not so fast, says Linus Torvalds