MSI must have provided a power supply for their expansion chassis.
This is the first TB expansion chassis I have seen that appears to provide a reasonably practical solution to the new Mini's limited graphics capability. Though not "cheap", the price looks to be much more reasonable than the alternative chassis which are of a lesser capability in any event.
Yeah, the GPU boxes have external PSUs - one of the reps picks up the PSU in one of the videos saying it can get quite hot. One thing I'd forgotten about MSI is they actually build the GPU boards too:
That gives them an advantage when it comes to price and also as far as drivers are concerned. I have a feeling that they are probably waiting on Intel's Redwood Ridge controller due next April-June. Right now Intel is supposedy charging about $20-25 per controller chip. While it sounds cheap, if you make 100,000 units, a price of say $5-10 looks a lot better. Then in 2014 with Falcon Ridge, they can double the bandwidth to 20Gbps to eliminate any real-world bottlenecks.
<snip>I have a feeling that they are probably waiting on Intel's Redwood Ridge controller due next April-June. Right now Intel is supposedy charging about $20-25 per controller chip. While it sounds cheap, if you make 100,000 units, a price of say $5-10 looks a lot better. Then in 2014 with Falcon Ridge, they can double the bandwidth to 20Gbps to eliminate any real-world bottlenecks.
I was aware that the new TB chips were supposed to be "cheaper", but had not picked up on the specific price point. The problem with the current chip set is that it has been about a $100 upcharge over essentially the same products w/o TB. That's just a big pill to swallow for most of the market when USB 3 is "good enough" for the overwhelming majority of people who are simply using an external hard drive either for backup or additional storage. At present, people who want an external RAID array and such either are continuing with one of the NAS systems or fibre for those who already are using it in a Map Pro. I had read that TB II, as I have been calling it, is supposed to be either announced or released late this year (which will have to happen very quickly) or early next year (2013) rather than 2014. I certainly would be willing to wait a little while to see if MSI were going to release this with TB II (and it would give them a bit of time to work on cooling the PS).
A second Thunderbolt port. A third Thunderbolt port if they insist on getting rid of Firewire 800. These ports are effectively like having external PCI and provide tremendous flexibility. Hook up an external PCI if you want discrete graphics, or put them whole thing in a rack mount casing for use as a server.
As many power consumption efficiencies as they can find.
One thing I haven't seen mentioned on wish lists is the drive bay height. With Western Digital pioneering a 15mm drive height with their 2TB 2.5 inch drive and others giving 12mm a shot, it's only logical that the Mac Mini should support these thicker drives. A double-2TB Mac Mini server in RAID-1 would be impressive indeed.
As more and more Mac Mini accessories come out, particularly accessories such as standard rack server cases that contain a Mac Mini, or hubs that you can put on top of a Mac Mini, it's important that Apple make a long-term public commitment to the Mini form factor. It would be a plus both for the product and the customer.
Make the server system more capable, perhaps. I get the feeling Mountain Lion server is dumbed-down too far. And we're getting to the point with miniaturization where this is the idea solution for a professional-grade server.
It adds up to the most flexible desktop computer on the market, fulfilling every role from home media center to supercomputer component, if they're able to incorporate all these changes.
In the case of the Mini embedded in a machine, it is in an interferometer of sorts. Interfacing is via Ethernet.
Speaking of which the suggested path for Mach (Mach the CNC software) under Windows is interfacing to smooth steeper over an Ehternet connection. The problem with USB being that it is unreliable.
I'm sure TB wold be an excellent solution for this sort of thing if somebody where to build the right hardware and software drivers. Actually isn't would be a good bipusiness opportunity for someone, but honestly I have no idea if TB even works under Linux. This sort of reminds me, what the world needs is a good clearing house or promotional site for TB based hardware. Lastly as far as I know the Mac never had Parallel ports.
I am currently constructing a CNC box with a breakout board, a cluster of Gecko drives, power supplies and an embedded computer.
The last of these is an Intel Atom D525MW board with an attached SSD with Ubuntu 10.04/LinuxCNC 2.5 on it. It never occurred to me to use a Mac Mini (even though it can run Linux) on account of the decade-old decision by Apple to drop the parallel port (other PC board makers seem to more and more be dropping it now, too).
I also noted yor comment about the Mini being used as part of embedded machine control, however, and that got me thinking a bit - presumably that's not for stuff that uses the old parallel port (which LinuuxCNC supports).
A quick check of the Mini's specs didn't enlighten me wrt the presence or otherwise of a PCI slot into which one might be able to plug a parallel port card and thus be able to use the Mini for LinuxCNC control.
Does the Mini in fact have such a slot buried in its not-very-deep depths? Or is there another way around this that I don't know about (e.g., via TB)?
Thanks, Wizard . I think you're right about the pport not being on Macs - I was probably thinking of the old ADB, which in any case is a completely different bucket o' snakes from the pport.
I hadn't thought about TB drivers for Linux - yet another snag. I should ask on the LinuxCNC forum. I like your idea of getting TB out into the wild, as it were. At this juncture, a Mac Mini with an extra chassis and a PCI pport card would be unfeasible financially, and probably wouldn't fit in the enclosure even if I removed the Mini from its case. Anyway, as I said, thanks.
A second Thunderbolt port. A third Thunderbolt port if they insist on getting rid of Firewire 800. These ports are effectively like having external PCI and provide tremendous flexibility. Hook up an external PCI if you want discrete graphics, or put them whole thing in a rack mount casing for use as a server.
The current chips support a total of 2 ports, so that is the most you'll see for now. People are always mentioning the concept of eGPUs. It's not as simple as plug and play. Marvin linked a couple before that were working quite effectively, but someone has to do the driver work. A lot of cards you see run in such things are hacked or rely on ones used in the mac pro. The mac pro hasn't seen any updates, so I would worry about that potential source. It would be a better solution if companies that sell GPU cards came out with their own eGPU designs with the inclusion of appropriate drivers. Otherwise this is nothing like a plug and play kind of experience.
If MSI sold a GUS box at $199, they'd probably have to miss out the TB cable. Each cable has two microchips inside so the cost to manufacturer them must be quite significant. A company called Gennum (now owned by Semtech) makes Apple's chips:
Even if Redwood Ridge lowers the cost of the chips, the cables might still create a premium over passive USB 3 cables. I wish the cables had been designed as purely passive optical cables with electrical power capability on short cables and Intel could have dealt with the cost of the controller. The bandwidth should be as high as possible and it can be shared between all the ports. 50-100Gbps between 2-4 ports is fine because if you absolutey need the throughput for one application, you can give it the full bandwidth.
I had read that TB II, as I have been calling it, is supposed to be either announced or released late this year (which will have to happen very quickly) or early next year (2013) rather than 2014.
This gen (2nd gen) is Cactus Ridge, next is Redwood Ridge and then Falcon Ridge:
A second Thunderbolt port. A third Thunderbolt port if they insist on getting rid of Firewire 800.
I was surprised they kept FW800 on the Mini instead of adding a second TB port - they dropped it from the iMac. Adding a 3rd would add 10W power allocation though, which might be a bit much for the 85W Mini. They have a 45W CPU, 10W TB, 4.5W per USB 3 port x 4 = 73W. Firewire 800 supposedly provides up to 45W but obviously the system will limit the power and FW bus-powered devices are around 7W. They need two TB though.
Speaking of the iMac, they are reported to be shipping to distribution hubs now:
"There are three possible solutions for printing to an Apple LaserWriter printer using TCP/IP:
Configure the LaserWriter for TCP/IP printing (LPR)
TCP/IP - AppleTalk conversion (LPR-PAP)
Direct Serial or parallel connections"
There is a USB adaptor here but that won't work in all cases:
If MSI sold a GUS box at $199, they'd probably have to miss out the TB cable. Each cable has two microchips inside so the cost to manufacturer them must be quite significant. A company called Gennum (now owned by Semtech) makes Apple's chips:
Even if Redwood Ridge lowers the cost of the chips, the cables might still create a premium over passive USB 3 cables. I wish the cables had been designed as purely passive optical cables with electrical power capability on short cables and Intel could have dealt with the cost of the controller. The bandwidth should be as high as possible and it can be shared between all the ports. 50-100Gbps between 2-4 ports is fine because if you absolutey need the throughput for one application, you can give it the full bandwidth.
This gen (2nd gen) is Cactus Ridge, next is Redwood Ridge and then Falcon Ridge:
If they can get Falcon Ridge ready for Haswell, that would be great but I suspect they'll drag it out.
I was surprised they kept FW800 on the Mini instead of adding a second TB port - they dropped it from the iMac. Adding a 3rd would add 10W power allocation though, which might be a bit much for the 85W Mini. They have a 45W CPU, 10W TB, 4.5W per USB 3 port x 4 = 73W. Firewire 800 supposedly provides up to 45W but obviously the system will limit the power and FW bus-powered devices are around 7W. They need two TB though.
Speaking of the iMac, they are reported to be shipping to distribution hubs now:
They'd have to be really, we are nearing the end of November.
Haswell is supposed to be out somewhere around March/April 2013. If Apple were to incorporate the next gen TB (I'm afraid I got lost on which gen TB chipset is the presently shipping one...I thought this was a 1st gen chipset and so I am referring to whatever is "next" as a candidate for a 2013 Mac Mini) into the next release of Macs it would both improve performance potential and possibly have a lesser cost chip. I fully agree about the cabling. I fear that, having made the mistake of putting chips in the cabling in the 1st gen TB, they will continue to do so for backwards compatibility reasons. That would be bad all the way around IMO because they simply must get the cost down if TB is to succeed at all, even if it means that the current gen TB is the red-headed step child of the TB family. I think it's that important.
I don't know how Intel would get the Falcon Ridge (Gen 4) chips ready next year unless they simply chose to skip the intervening generations of chips which would assume that they have had remarkable success implementing both the design and fabrication of that chip set. I would certainly welcome a deviation from the usual incremental upgrades. I really don't think the cost of optical fiber, surrounded by some sort of protective casing to keep it from getting kinked and broken, would cost that much if there were not those chips... which also contribute to overhead in the transmission of data.
If Apple needed to go to a 95 or 100 W power supply in order to support 2 TB ports I don't see that as a particular problem. There probably would need to be a few tweaks in the system cooling, maybe a second fan or a revised air inlet/outlet system ("in with the good air, out with the bad air"), but it shouldn't be that big a deal. Even if the form factor of the Mini needed to be revised to a slightly larger package, so what. It's still smaller than most anything other than a Raspberry Pi.
On the other hand, Haswell is supposed to have a lower power draw than the current Ivy Bridge CPUs so it might make up the difference and allow the use of the same rated power supply and cooling system.
Haswell is supposed to be out somewhere around March/April 2013. If Apple were to incorporate the next gen TB (I'm afraid I got lost on which gen TB chipset is the presently shipping one...I thought this was a 1st gen chipset and so I am referring to whatever is "next" as a candidate for a 2013 Mac Mini) into the next release of Macs it would both improve performance potential and possibly have a lesser cost chip. I fully agree about the cabling. I fear that, having made the mistake of putting chips in the cabling in the 1st gen TB, they will continue to do so for backwards compatibility reasons.
I'm not sure where this idea comes from but putting electronics in the cable is not a mistake at all. Please stop all comparisons with USB 3 as the two ports aren't even remotely similar nor are the targeted at similar usages.
In any event the chips in those cables perform very important functions and allow an easy transition to fiber optics for those that need it. In effect Apple/Intel have decoupled the physical media from the current port standard. This gives them real flexibility in the future.
That would be bad all the way around IMO because they simply must get the cost down if TB is to succeed at all, even if it means that the current gen TB is the red-headed step child of the TB family. I think it's that important.
TB is already a huge success. I suspect it is being used exactly as Apple has imagined, that is as a single cable docking solution for laptops. Intel may not like that, you may not like that and frankly I might not like it but for Apple they have solved one very significant issue for themselves.
I don't know how Intel would get the Falcon Ridge (Gen 4) chips ready next year unless they simply chose to skip the intervening generations of chips which would assume that they have had remarkable success implementing both the design and fabrication of that chip set. I would certainly welcome a deviation from the usual incremental upgrades. I really don't think the cost of optical fiber, surrounded by some sort of protective casing to keep it from getting kinked and broken, would cost that much if there were not those chips... which also contribute to overhead in the transmission of data.
What chips! Data cables, that is fiber based data cables have been terminated for years with the electronic interface in the cable. The other option is an optical connector built into your Mac which isn't exactly a user friendly termination technique. Remember we are talking high performance data transmissions here, this is not the same class of performance as seen in your stereo system.
If Apple needed to go to a 95 or 100 W power supply in order to support 2 TB ports I don't see that as a particular problem. There probably would need to be a few tweaks in the system cooling, maybe a second fan or a revised air inlet/outlet system ("in with the good air, out with the bad air"), but it shouldn't be that big a deal. Even if the form factor of the Mini needed to be revised to a slightly larger package, so what. It's still smaller than most anything other than a Raspberry Pi.
Well this is one thing I can agree with Apple needs to have a power supply option in the Mini that allows a true distinction between the base Mini and an uprated model.
On the other hand, Haswell is supposed to have a lower power draw than the current Ivy Bridge CPUs so it might make up the difference and allow the use of the same rated power supply and cooling system.
Cheers
Ideally any power savings in Haswell will be reallocated to higher performance graphics. Thus I'm really hoping for GT3 class GPUs in the next upper end Mini. Again the base model would likely be more run of the mill but that is really the point. If I buy a $799 Mini I really want to get very good GPU performance for that extra Money.
I'm not sure where this idea comes from but putting electronics in the cable is not a mistake at all. Please stop all comparisons with USB 3 as the two ports aren't even remotely similar nor are the targeted at similar usages.
In any event the chips in those cables perform very important functions and allow an easy transition to fiber optics for those that need it. In effect Apple/Intel have decoupled the physical media from the current port standard. This gives them real flexibility in the future.
TB is already a huge success. I suspect it is being used exactly as Apple has imagined, that is as a single cable docking solution for laptops. Intel may not like that, you may not like that and frankly I might not like it but for Apple they have solved one very significant issue for themselves.
What chips! Data cables, that is fiber based data cables have been terminated for years with the electronic interface in the cable. The other option is an optical connector built into your Mac which isn't exactly a user friendly termination technique. Remember we are talking high performance data transmissions here, this is not the same class of performance as seen in your stereo system.
Well this is one thing I can agree with Apple needs to have a power supply option in the Mini that allows a true distinction between the base Mini and an uprated model.
Ideally any power savings in Haswell will be reallocated to higher performance graphics. Thus I'm really hoping for GT3 class GPUs in the next upper end Mini. Again the base model would likely be more run of the mill but that is really the point. If I buy a $799 Mini I really want to get very good GPU performance for that extra Money.
I disagree in part. The idea of "not putting chips in cables" is, quite simply, a cost issue which goes to the very heart of TB. Whether TB is a "huge success" or not must certainly depend upon one's perspective and what one considers "success". In my view TB is not yet a failure and neither is it a success at this time. Unless something is done about the cost issues, I question whether it will be broadly adopted. It could easily become a niche product.
As far as putting decent graphics in the Mini, I could not agree more. The power supply is only a part of the problem. The other problem is thermal design/management and the question of the size of the box. Experience has shown that Ive places a greater premium on appearance and size than functionality. I would gleefully take a Mini that was a bit bigger if I could get decent graphics in it and improve the thermal design a bit while they were at it.
I do grasp that the entire reason for being of TB is high speed data transfer. Quite frankly, that has been one of the big limitations of external storage. Apple, for reasons best known to them, entirely skipped eSATA. In my experience, it was a welcome option for data backup as well as external data storage.
Out of curiosity, not to challenge you, but just how much utilization of TB as a single cable docking solution for laptops have you seen? I have seen very little utilization of TB for anything, but you probably are exposed to a wider group of users than I.
As an aside, I am just waiting for OWC to come out with some TB storage products. Even if the current TB "isn't [edit] all that fast", it is a great deal faster than any current alternative and would do nicely with the Mini I just pulled the trigger on because I just could not wait any longer.
TB is already a huge success. I suspect it is being used exactly as Apple has imagined, that is as a single cable docking solution for laptops. Intel may not like that, you may not like that and frankly I might not like it but for Apple they have solved one very significant issue for themselves.
I've said it before. Apple was able to implement a display/docking station and storage solution with the launch of thunderbolt. Prior to that they had mini displayport there anyway. It's not like anything was really compromised. The disconnect was in the lord of the rings complex projected on it. I remain curious where Apple is going with their gpu solutions. Memory access seems to be the main limiting factor at the moment.
I read somewhere, I think either Ars or Anandtech, that the reason for the high cost cabling on TB was due to design issues with copper cabling and that they had to scramble (I think) to solve the issue resulting in a two chip solution. So there are four chips in a TB cable, two in each end. While combing those two into one would save costs, due to design issues with the current 10G spec, they can't really do that in a cost effective way. The plan is apparently to do so, with the TB II (20G) version when it comes out in 2014. That presents issues, in that there is a fix for it, but it won't be available for around two years. Therefore, people are basically stuck with the expensive cables until the next gen comes out.
Unfortunately, I don't have the link, nor do I remember exactly where I read that Article.
How long do you figure it takes for a Mac mini redesign? I like the way it is now though knowing Apple, I think it is only a matter of time before it gets smaller and/or thinner.
I do grasp that the entire reason for being of TB is high speed data transfer.
Except that "high speed data transfer" is not at all Thunderbolt's reason for being.
Its reason for being is a universal, simultaneously multi-protocol expansion port for the current and future generations of computers. As an embodiment of PCI-E, it can carry any current or future device that would be feasible with a PCI-based architecture, including multiple devices at the same time.
High-speed data transfer is simply a capability that is needed to meet this goal.
Quote:
Apple, for reasons best known to them, entirely skipped eSATA.
Apple skipped eSATA because it is a non-universal, single-protocol port that is likely only desired by a tiny percentace of their customers. I've never seen an eSATA drive in use, despite plenty of laptops with ports. USB 2/3 drives are commonplace and cheap, and serve most people's needs. For special storage solutions, Thunderbolt-connected drives exist. If you really *must* have that eSATA plug, Thunderbolt can do that: http://www.lacie.com/products/product.htm?id=10574 (I won't call it cheap, though...)
Knew a guy with one. It was an eSATA+USB port, which I thought was nifty, if nothing else. Also probably the only way they could get anyone to actually use the port, since he never used it for anything but USB devices. Claimed it was "faster than the others".
Knew a guy with one. It was an eSATA+USB port, which I thought was nifty, if nothing else. Also probably the only way they could get anyone to actually use the port, since he never used it for anything but USB devices. Claimed it was "faster than the others".
The eSATA standard can be faster than usb. It depends on the implementation, but you can get SATA bandwidth with minimal overhead. Some cards are much faster than others. I've never seen one with a usb connector as opposed to an eSATA connector though.
The eSATA standard can be faster than usb. It depends on the implementation, but you can get SATA bandwidth with minimal overhead. Some cards are much faster than others. I've never seen one with a usb connector as opposed to an eSATA connector though.
Precisely! The iMacs had an extra internal SATA controller port. You could run some wiring from it to an eSATA port (OWC did this) and get the benefit of an external drive at essentially the same speed an an internal SATA drive. It was faster than FW800 in my experience.
There seems to be some confusion about it not being a standard. SATA is a standard. It is just that there are three generations of it and a SATA 3 drive on a SATA 1 controller is only going to transfer data at SATA 1 rates.
There was a combination port I have sen on a few PCs which could physically accept either a USB plug or an eSATA plug. I have not first hand experience with them however.
Comments
Yeah, the GPU boxes have external PSUs - one of the reps picks up the PSU in one of the videos saying it can get quite hot. One thing I'd forgotten about MSI is they actually build the GPU boards too:
http://uk.msi.com/product/vga/R6770-MD1GD5.html
That gives them an advantage when it comes to price and also as far as drivers are concerned. I have a feeling that they are probably waiting on Intel's Redwood Ridge controller due next April-June. Right now Intel is supposedy charging about $20-25 per controller chip. While it sounds cheap, if you make 100,000 units, a price of say $5-10 looks a lot better. Then in 2014 with Falcon Ridge, they can double the bandwidth to 20Gbps to eliminate any real-world bottlenecks.
Quote:
Originally Posted by Marvin
<snip>I have a feeling that they are probably waiting on Intel's Redwood Ridge controller due next April-June. Right now Intel is supposedy charging about $20-25 per controller chip. While it sounds cheap, if you make 100,000 units, a price of say $5-10 looks a lot better. Then in 2014 with Falcon Ridge, they can double the bandwidth to 20Gbps to eliminate any real-world bottlenecks.
I was aware that the new TB chips were supposed to be "cheaper", but had not picked up on the specific price point. The problem with the current chip set is that it has been about a $100 upcharge over essentially the same products w/o TB. That's just a big pill to swallow for most of the market when USB 3 is "good enough" for the overwhelming majority of people who are simply using an external hard drive either for backup or additional storage. At present, people who want an external RAID array and such either are continuing with one of the NAS systems or fibre for those who already are using it in a Map Pro. I had read that TB II, as I have been calling it, is supposed to be either announced or released late this year (which will have to happen very quickly) or early next year (2013) rather than 2014. I certainly would be willing to wait a little while to see if MSI were going to release this with TB II (and it would give them a bit of time to work on cooling the PS).
This is genuinely interesting.
Haswell, obviously.
A second Thunderbolt port. A third Thunderbolt port if they insist on getting rid of Firewire 800. These ports are effectively like having external PCI and provide tremendous flexibility. Hook up an external PCI if you want discrete graphics, or put them whole thing in a rack mount casing for use as a server.
As many power consumption efficiencies as they can find.
One thing I haven't seen mentioned on wish lists is the drive bay height. With Western Digital pioneering a 15mm drive height with their 2TB 2.5 inch drive and others giving 12mm a shot, it's only logical that the Mac Mini should support these thicker drives. A double-2TB Mac Mini server in RAID-1 would be impressive indeed.
As more and more Mac Mini accessories come out, particularly accessories such as standard rack server cases that contain a Mac Mini, or hubs that you can put on top of a Mac Mini, it's important that Apple make a long-term public commitment to the Mini form factor. It would be a plus both for the product and the customer.
Make the server system more capable, perhaps. I get the feeling Mountain Lion server is dumbed-down too far. And we're getting to the point with miniaturization where this is the idea solution for a professional-grade server.
It adds up to the most flexible desktop computer on the market, fulfilling every role from home media center to supercomputer component, if they're able to incorporate all these changes.
Speaking of which the suggested path for Mach (Mach the CNC software) under Windows is interfacing to smooth steeper over an Ehternet connection. The problem with USB being that it is unreliable.
I'm sure TB wold be an excellent solution for this sort of thing if somebody where to build the right hardware and software drivers. Actually isn't would be a good bipusiness opportunity for someone, but honestly I have no idea if TB even works under Linux. This sort of reminds me, what the world needs is a good clearing house or promotional site for TB based hardware. Lastly as far as I know the Mac never had Parallel ports.
I hadn't thought about TB drivers for Linux - yet another snag. I should ask on the LinuxCNC forum. I like your idea of getting TB out into the wild, as it were. At this juncture, a Mac Mini with an extra chassis and a PCI pport card would be unfeasible financially, and probably wouldn't fit in the enclosure even if I removed the Mini from its case. Anyway, as I said, thanks.
Cheers,
Alex.
Quote:
Originally Posted by photoeditor
Haswell, obviously.
A second Thunderbolt port. A third Thunderbolt port if they insist on getting rid of Firewire 800. These ports are effectively like having external PCI and provide tremendous flexibility. Hook up an external PCI if you want discrete graphics, or put them whole thing in a rack mount casing for use as a server.
The current chips support a total of 2 ports, so that is the most you'll see for now. People are always mentioning the concept of eGPUs. It's not as simple as plug and play. Marvin linked a couple before that were working quite effectively, but someone has to do the driver work. A lot of cards you see run in such things are hacked or rely on ones used in the mac pro. The mac pro hasn't seen any updates, so I would worry about that potential source. It would be a better solution if companies that sell GPU cards came out with their own eGPU designs with the inclusion of appropriate drivers. Otherwise this is nothing like a plug and play kind of experience.
I think the cables have to come into it somewhere. They are $50-60 each:
http://www.elgato.com/elgato/na/mainmenu/products/storage/Thunderbolt-Cable.en.en.html
http://store.apple.com/us/product/MC913ZM/A/apple-thunderbolt-cable-2-m
If MSI sold a GUS box at $199, they'd probably have to miss out the TB cable. Each cable has two microchips inside so the cost to manufacturer them must be quite significant. A company called Gennum (now owned by Semtech) makes Apple's chips:
http://www.semtech.com/high-speed-interfaces/thunderbolt-cable-transceivers/gn2033/
Even if Redwood Ridge lowers the cost of the chips, the cables might still create a premium over passive USB 3 cables. I wish the cables had been designed as purely passive optical cables with electrical power capability on short cables and Intel could have dealt with the cost of the controller. The bandwidth should be as high as possible and it can be shared between all the ports. 50-100Gbps between 2-4 ports is fine because if you absolutey need the throughput for one application, you can give it the full bandwidth.
This gen (2nd gen) is Cactus Ridge, next is Redwood Ridge and then Falcon Ridge:
http://www.techspot.com/news/49502-thunderbolt-bandwidth-to-double-in-2014-with-falcon-ridge.html
If they can get Falcon Ridge ready for Haswell, that would be great but I suspect they'll drag it out.
I was surprised they kept FW800 on the Mini instead of adding a second TB port - they dropped it from the iMac. Adding a 3rd would add 10W power allocation though, which might be a bit much for the 85W Mini. They have a 45W CPU, 10W TB, 4.5W per USB 3 port x 4 = 73W. Firewire 800 supposedly provides up to 45W but obviously the system will limit the power and FW bus-powered devices are around 7W. They need two TB though.
Speaking of the iMac, they are reported to be shipping to distribution hubs now:
http://www.macworld.co.uk/mac/news/?newsid=3411877&pagtype=allchandate
They'd have to be really, we are nearing the end of November.
Way back:
http://support.apple.com/kb/SP232
SCSI: DB-25
The Laserwriter had parallel connections:
http://support.apple.com/kb/TA35778
"There are three possible solutions for printing to an Apple LaserWriter printer using TCP/IP:
Configure the LaserWriter for TCP/IP printing (LPR)
TCP/IP - AppleTalk conversion (LPR-PAP)
Direct Serial or parallel connections"
There is a USB adaptor here but that won't work in all cases:
http://www.amazon.com/IOGEAR-GUC1284B-USB-Parallel-Adapter/dp/B00018RT1E
The easiest option is probably just to buy an old PC with a parallel port and control it over a network.
Quote:
Originally Posted by Marvin
I think the cables have to come into it somewhere. They are $50-60 each:
http://www.elgato.com/elgato/na/mainmenu/products/storage/Thunderbolt-Cable.en.en.html
http://store.apple.com/us/product/MC913ZM/A/apple-thunderbolt-cable-2-m
If MSI sold a GUS box at $199, they'd probably have to miss out the TB cable. Each cable has two microchips inside so the cost to manufacturer them must be quite significant. A company called Gennum (now owned by Semtech) makes Apple's chips:
http://www.semtech.com/high-speed-interfaces/thunderbolt-cable-transceivers/gn2033/
Even if Redwood Ridge lowers the cost of the chips, the cables might still create a premium over passive USB 3 cables. I wish the cables had been designed as purely passive optical cables with electrical power capability on short cables and Intel could have dealt with the cost of the controller. The bandwidth should be as high as possible and it can be shared between all the ports. 50-100Gbps between 2-4 ports is fine because if you absolutey need the throughput for one application, you can give it the full bandwidth.
This gen (2nd gen) is Cactus Ridge, next is Redwood Ridge and then Falcon Ridge:
http://www.techspot.com/news/49502-thunderbolt-bandwidth-to-double-in-2014-with-falcon-ridge.html
If they can get Falcon Ridge ready for Haswell, that would be great but I suspect they'll drag it out.
I was surprised they kept FW800 on the Mini instead of adding a second TB port - they dropped it from the iMac. Adding a 3rd would add 10W power allocation though, which might be a bit much for the 85W Mini. They have a 45W CPU, 10W TB, 4.5W per USB 3 port x 4 = 73W. Firewire 800 supposedly provides up to 45W but obviously the system will limit the power and FW bus-powered devices are around 7W. They need two TB though.
Speaking of the iMac, they are reported to be shipping to distribution hubs now:
http://www.macworld.co.uk/mac/news/?newsid=3411877&pagtype=allchandate
They'd have to be really, we are nearing the end of November.
Haswell is supposed to be out somewhere around March/April 2013. If Apple were to incorporate the next gen TB (I'm afraid I got lost on which gen TB chipset is the presently shipping one...I thought this was a 1st gen chipset and so I am referring to whatever is "next" as a candidate for a 2013 Mac Mini) into the next release of Macs it would both improve performance potential and possibly have a lesser cost chip. I fully agree about the cabling. I fear that, having made the mistake of putting chips in the cabling in the 1st gen TB, they will continue to do so for backwards compatibility reasons. That would be bad all the way around IMO because they simply must get the cost down if TB is to succeed at all, even if it means that the current gen TB is the red-headed step child of the TB family. I think it's that important.
I don't know how Intel would get the Falcon Ridge (Gen 4) chips ready next year unless they simply chose to skip the intervening generations of chips which would assume that they have had remarkable success implementing both the design and fabrication of that chip set. I would certainly welcome a deviation from the usual incremental upgrades. I really don't think the cost of optical fiber, surrounded by some sort of protective casing to keep it from getting kinked and broken, would cost that much if there were not those chips... which also contribute to overhead in the transmission of data.
If Apple needed to go to a 95 or 100 W power supply in order to support 2 TB ports I don't see that as a particular problem. There probably would need to be a few tweaks in the system cooling, maybe a second fan or a revised air inlet/outlet system ("in with the good air, out with the bad air"), but it shouldn't be that big a deal. Even if the form factor of the Mini needed to be revised to a slightly larger package, so what. It's still smaller than most anything other than a Raspberry Pi.
On the other hand, Haswell is supposed to have a lower power draw than the current Ivy Bridge CPUs so it might make up the difference and allow the use of the same rated power supply and cooling system.
Cheers
In any event the chips in those cables perform very important functions and allow an easy transition to fiber optics for those that need it. In effect Apple/Intel have decoupled the physical media from the current port standard. This gives them real flexibility in the future. TB is already a huge success. I suspect it is being used exactly as Apple has imagined, that is as a single cable docking solution for laptops. Intel may not like that, you may not like that and frankly I might not like it but for Apple they have solved one very significant issue for themselves. What chips! Data cables, that is fiber based data cables have been terminated for years with the electronic interface in the cable. The other option is an optical connector built into your Mac which isn't exactly a user friendly termination technique. Remember we are talking high performance data transmissions here, this is not the same class of performance as seen in your stereo system. Well this is one thing I can agree with Apple needs to have a power supply option in the Mini that allows a true distinction between the base Mini and an uprated model.
Ideally any power savings in Haswell will be reallocated to higher performance graphics. Thus I'm really hoping for GT3 class GPUs in the next upper end Mini. Again the base model would likely be more run of the mill but that is really the point. If I buy a $799 Mini I really want to get very good GPU performance for that extra Money.
Quote:
Originally Posted by wizard69
I'm not sure where this idea comes from but putting electronics in the cable is not a mistake at all. Please stop all comparisons with USB 3 as the two ports aren't even remotely similar nor are the targeted at similar usages.
In any event the chips in those cables perform very important functions and allow an easy transition to fiber optics for those that need it. In effect Apple/Intel have decoupled the physical media from the current port standard. This gives them real flexibility in the future.
TB is already a huge success. I suspect it is being used exactly as Apple has imagined, that is as a single cable docking solution for laptops. Intel may not like that, you may not like that and frankly I might not like it but for Apple they have solved one very significant issue for themselves.
What chips! Data cables, that is fiber based data cables have been terminated for years with the electronic interface in the cable. The other option is an optical connector built into your Mac which isn't exactly a user friendly termination technique. Remember we are talking high performance data transmissions here, this is not the same class of performance as seen in your stereo system.
Well this is one thing I can agree with Apple needs to have a power supply option in the Mini that allows a true distinction between the base Mini and an uprated model.
Ideally any power savings in Haswell will be reallocated to higher performance graphics. Thus I'm really hoping for GT3 class GPUs in the next upper end Mini. Again the base model would likely be more run of the mill but that is really the point. If I buy a $799 Mini I really want to get very good GPU performance for that extra Money.
I disagree in part. The idea of "not putting chips in cables" is, quite simply, a cost issue which goes to the very heart of TB. Whether TB is a "huge success" or not must certainly depend upon one's perspective and what one considers "success". In my view TB is not yet a failure and neither is it a success at this time. Unless something is done about the cost issues, I question whether it will be broadly adopted. It could easily become a niche product.
As far as putting decent graphics in the Mini, I could not agree more. The power supply is only a part of the problem. The other problem is thermal design/management and the question of the size of the box. Experience has shown that Ive places a greater premium on appearance and size than functionality. I would gleefully take a Mini that was a bit bigger if I could get decent graphics in it and improve the thermal design a bit while they were at it.
I do grasp that the entire reason for being of TB is high speed data transfer. Quite frankly, that has been one of the big limitations of external storage. Apple, for reasons best known to them, entirely skipped eSATA. In my experience, it was a welcome option for data backup as well as external data storage.
Out of curiosity, not to challenge you, but just how much utilization of TB as a single cable docking solution for laptops have you seen? I have seen very little utilization of TB for anything, but you probably are exposed to a wider group of users than I.
As an aside, I am just waiting for OWC to come out with some TB storage products. Even if the current TB "isn't [edit] all that fast", it is a great deal faster than any current alternative and would do nicely with the Mini I just pulled the trigger on because I just could not wait any longer.
Regards
Quote:
Originally Posted by wizard69
TB is already a huge success. I suspect it is being used exactly as Apple has imagined, that is as a single cable docking solution for laptops. Intel may not like that, you may not like that and frankly I might not like it but for Apple they have solved one very significant issue for themselves.
I've said it before. Apple was able to implement a display/docking station and storage solution with the launch of thunderbolt. Prior to that they had mini displayport there anyway. It's not like anything was really compromised. The disconnect was in the lord of the rings complex projected on it. I remain curious where Apple is going with their gpu solutions. Memory access seems to be the main limiting factor at the moment.
Unfortunately, I don't have the link, nor do I remember exactly where I read that Article.
[URL=http://arstechnica.com/apple/2012/07/why-thunderbolt-cables-will-be-expensive-until-2013/]http://arstechnica.com/apple/2012/07/why-thunderbolt-cables-will-be-expensive-until-2013/[/URL]
Quote:
Originally Posted by RBR
I do grasp that the entire reason for being of TB is high speed data transfer.
Except that "high speed data transfer" is not at all Thunderbolt's reason for being.
Its reason for being is a universal, simultaneously multi-protocol expansion port for the current and future generations of computers. As an embodiment of PCI-E, it can carry any current or future device that would be feasible with a PCI-based architecture, including multiple devices at the same time.
High-speed data transfer is simply a capability that is needed to meet this goal.
Quote:
Apple, for reasons best known to them, entirely skipped eSATA.
Apple skipped eSATA because it is a non-universal, single-protocol port that is likely only desired by a tiny percentace of their customers. I've never seen an eSATA drive in use, despite plenty of laptops with ports. USB 2/3 drives are commonplace and cheap, and serve most people's needs. For special storage solutions, Thunderbolt-connected drives exist. If you really *must* have that eSATA plug, Thunderbolt can do that: http://www.lacie.com/products/product.htm?id=10574 (I won't call it cheap, though...)
Originally Posted by mbmcavoy
…despite plenty of laptops with ports.
Knew a guy with one. It was an eSATA+USB port, which I thought was nifty, if nothing else. Also probably the only way they could get anyone to actually use the port, since he never used it for anything but USB devices. Claimed it was "faster than the others".
Quote:
Originally Posted by Tallest Skil
Knew a guy with one. It was an eSATA+USB port, which I thought was nifty, if nothing else. Also probably the only way they could get anyone to actually use the port, since he never used it for anything but USB devices. Claimed it was "faster than the others".
The eSATA standard can be faster than usb. It depends on the implementation, but you can get SATA bandwidth with minimal overhead. Some cards are much faster than others. I've never seen one with a usb connector as opposed to an eSATA connector though.
Quote:
Originally Posted by hmm
The eSATA standard can be faster than usb. It depends on the implementation, but you can get SATA bandwidth with minimal overhead. Some cards are much faster than others. I've never seen one with a usb connector as opposed to an eSATA connector though.
Precisely! The iMacs had an extra internal SATA controller port. You could run some wiring from it to an eSATA port (OWC did this) and get the benefit of an external drive at essentially the same speed an an internal SATA drive. It was faster than FW800 in my experience.
There seems to be some confusion about it not being a standard. SATA is a standard. It is just that there are three generations of it and a SATA 3 drive on a SATA 1 controller is only going to transfer data at SATA 1 rates.
There was a combination port I have sen on a few PCs which could physically accept either a USB plug or an eSATA plug. I have not first hand experience with them however.