Thanks for pointing the first one out - it should have been 'initiated in part'. As to the phrasing of the first sentence, I find it common enough where I am. Not sure if it's a UK thing or not, though US vs. UK idioms have been commented on in past reviews. As always, if anything catches your eye please feel free to email :)
Speaking of UK vs US, I had to Google your Stella Artois reference, and I actually drink the stuff occasionally. Must have been a UK specific ad campaign.
I understand that Thunderbolt is a lot faster and a different usage than USB 3. Frankly, it's not so much better than USB 3 that consumers will pay for Thunderbolt. USB 3 is already leading, and Thunderbolt will be left behind like Firewire despite the tech being superior...
It's rather meh on desktops since it's pretty easy to add new hardware internally, but it makes far more sense on laptops when you have limited number of ports. Having an external PCI-Express bus is interested, especially if external GPUs ever actually arrive at an affordable price point.
If you're going to plug in one device, yes. I think the strength of Thunderbolt is as a laptop dock - plug in one cable and you got wired network, sound, keyboard, mouse, printers external screens, any USB 1/2/3.0 device, firewire, esata, external graphics card dock, regular 3.5" HDDs and whatnot. That can have a future in many companies I think who've now chosen laptops for higher flexibility - now you can have that and dock into a full system with one cable.
It will mean a lot if they ever put thunderbolt on a mini itx board. I know a lot of SFF enthusiasts who would love to try external graphics with a sub 10 liter enclosure. But on a full size ATX board it doesn't really mean a whole lot.
It really would've been nice to see some Thunderbolt testing. I realize Anand is hogging all of the shiny TB gear, but the review didn't really test the primary draw of this MB and as such is kind of useless.
Agreed. As the summary correctly states, the raison d'etre of this board are the TB ports. Even if it's just the overpriced Promise a review should give them some stress and see how they perform. Maybe it'll get easier if QNAP ever releases their JBOD.
I have access to a TB device, but it is only the two-bay Little Big disk with a pair of Intel Drives. Can't really stress the TB implementation in terms of peak speeds, and in our copy test it can get anything from 1.1 seconds to 3.3 seconds depending on if the wind is blowing, or the tides are in (very unpredictable).
When I can get a 4-bay TB device in, I will fill it with 500MB/s+ SSDs and get down to testing. Unless there is a specific test you would like me to do (4K et al).
I am very curious about a couple points, however they are not the easiest scenarios to test.
Firstly, GIGABYTE depicts the ability to support a total of 12 connected Thunderbolt devices plus 2 displays, or 6 devices AND 1 display per port. [ http://www.gigabyte.com/microsite/306/images/thund... ] This seems to fly in the face of what we have been told by Apple and Intel about supported topologies, i.e. "up to a total of 6 devices, including up to 2 high resolution DisplayPort v1.1a displays". Can a single Cactus Ridge DSL3510L really handle that many devices? Is there some difference in implementation between Windows and Mac OS?
GIGABYTE also claims a full 10 Gbps of PCIe bandwidth from each port. Now I would also doubt that claim, and in the article you indicated this wasn't happening with a single DSL3510L. However, Anand achieved 1380 MB/s by using both Thunderbolt ports during his review of the MBPR, which also uses just one DSL3510L controller. Now ultimately this controller is bound by its PCIe 2.0 x4 back end, which should limit it to around 1600 MB/s of payload throughput, but breaking the 1000 MB/s barrier would seem to imply that there is more than one PCIe to Thunderbolt protocol adapter in these Cactus Ridge chips. This would be significant if true. Any chance you could lean on it hard enough to find out?
I agree about the need for TB testing as I have had nothing but issues with this board and a Seagate GoFlex attached on one port and a Apple TB 27" monitor on the other port. The monitor will flicker badly at times (does not happen on a competing board and a MacBook Pro) while the Seagate drive will "disappear" and requires a power on/off before being recognized again.
Ian, I think you are wrong about Gigabyte manipulating CPU voltage readings. The voltage reading 1.068 V is probably not CPU vcore, it could be VTT(VCCIO) voltage. It is known that some GB motherboards have this "feature", that they show VTT voltage in CPU-Z. For example, my GA-P67A-UD3P-B3 board shows 1.092 V in CPU-Z as vcore when I have set VTT to 1.1 V in BIOS.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
15 Comments
Back to Article
ElFenix - Monday, September 17, 2012 - link
"Despite all this, Gigabyte’s foray into the Thunderbolt world is spurned in part by the board we are reviewing today...."You probably meant 'spurred,' though that doesn't really fit either.
Also, the very first sentence should be more like "Because the exclusive license has expired...."
IanCutress - Monday, September 17, 2012 - link
Thanks for pointing the first one out - it should have been 'initiated in part'. As to the phrasing of the first sentence, I find it common enough where I am. Not sure if it's a UK thing or not, though US vs. UK idioms have been commented on in past reviews. As always, if anything catches your eye please feel free to email :)Ian
freedom4556 - Monday, September 17, 2012 - link
Speaking of UK vs US, I had to Google your Stella Artois reference, and I actually drink the stuff occasionally. Must have been a UK specific ad campaign.lurker22 - Monday, September 17, 2012 - link
I understand that Thunderbolt is a lot faster and a different usage than USB 3. Frankly, it's not so much better than USB 3 that consumers will pay for Thunderbolt. USB 3 is already leading, and Thunderbolt will be left behind like Firewire despite the tech being superior...dagamer34 - Monday, September 17, 2012 - link
It's rather meh on desktops since it's pretty easy to add new hardware internally, but it makes far more sense on laptops when you have limited number of ports. Having an external PCI-Express bus is interested, especially if external GPUs ever actually arrive at an affordable price point.Kjella - Monday, September 17, 2012 - link
If you're going to plug in one device, yes. I think the strength of Thunderbolt is as a laptop dock - plug in one cable and you got wired network, sound, keyboard, mouse, printers external screens, any USB 1/2/3.0 device, firewire, esata, external graphics card dock, regular 3.5" HDDs and whatnot. That can have a future in many companies I think who've now chosen laptops for higher flexibility - now you can have that and dock into a full system with one cable.sean.crees - Monday, September 17, 2012 - link
It will mean a lot if they ever put thunderbolt on a mini itx board. I know a lot of SFF enthusiasts who would love to try external graphics with a sub 10 liter enclosure. But on a full size ATX board it doesn't really mean a whole lot.Skidmarks - Tuesday, September 18, 2012 - link
That's possibly true but only time will tell.GeorgeH - Monday, September 17, 2012 - link
It really would've been nice to see some Thunderbolt testing. I realize Anand is hogging all of the shiny TB gear, but the review didn't really test the primary draw of this MB and as such is kind of useless.zanon - Monday, September 17, 2012 - link
Agreed. As the summary correctly states, the raison d'etre of this board are the TB ports. Even if it's just the overpriced Promise a review should give them some stress and see how they perform. Maybe it'll get easier if QNAP ever releases their JBOD.IanCutress - Monday, September 17, 2012 - link
I have access to a TB device, but it is only the two-bay Little Big disk with a pair of Intel Drives. Can't really stress the TB implementation in terms of peak speeds, and in our copy test it can get anything from 1.1 seconds to 3.3 seconds depending on if the wind is blowing, or the tides are in (very unpredictable).When I can get a 4-bay TB device in, I will fill it with 500MB/s+ SSDs and get down to testing. Unless there is a specific test you would like me to do (4K et al).
Ian
repoman27 - Monday, September 17, 2012 - link
I am very curious about a couple points, however they are not the easiest scenarios to test.Firstly, GIGABYTE depicts the ability to support a total of 12 connected Thunderbolt devices plus 2 displays, or 6 devices AND 1 display per port. [ http://www.gigabyte.com/microsite/306/images/thund... ] This seems to fly in the face of what we have been told by Apple and Intel about supported topologies, i.e. "up to a total of 6 devices, including up to 2 high resolution DisplayPort v1.1a displays". Can a single Cactus Ridge DSL3510L really handle that many devices? Is there some difference in implementation between Windows and Mac OS?
GIGABYTE also claims a full 10 Gbps of PCIe bandwidth from each port. Now I would also doubt that claim, and in the article you indicated this wasn't happening with a single DSL3510L. However, Anand achieved 1380 MB/s by using both Thunderbolt ports during his review of the MBPR, which also uses just one DSL3510L controller. Now ultimately this controller is bound by its PCIe 2.0 x4 back end, which should limit it to around 1600 MB/s of payload throughput, but breaking the 1000 MB/s barrier would seem to imply that there is more than one PCIe to Thunderbolt protocol adapter in these Cactus Ridge chips. This would be significant if true. Any chance you could lean on it hard enough to find out?
goinginstyle - Monday, September 17, 2012 - link
I agree about the need for TB testing as I have had nothing but issues with this board and a Seagate GoFlex attached on one port and a Apple TB 27" monitor on the other port. The monitor will flicker badly at times (does not happen on a competing board and a MacBook Pro) while the Seagate drive will "disappear" and requires a power on/off before being recognized again.thewhat - Tuesday, September 18, 2012 - link
It would be great if you could test with Speedfan whether the fan speed can be controlled independently for every header.NiggaASD - Sunday, December 9, 2012 - link
Ian, I think you are wrong about Gigabyte manipulating CPU voltage readings. The voltage reading 1.068 V is probably not CPU vcore, it could be VTT(VCCIO) voltage. It is known that some GB motherboards have this "feature", that they show VTT voltage in CPU-Z. For example, my GA-P67A-UD3P-B3 board shows 1.092 V in CPU-Z as vcore when I have set VTT to 1.1 V in BIOS.