AMD downplays Intel's Thunderbolt due to low bandwidth

By on March 3, 2011, 10:23 AM
AMD is not impressed with Intel's Thunderbolt since its bandwidth can be easily outstripped using a combination of existing standards. As a result, the company believes that the technology will become another proprietary standard and may not be widely adopted.

"Existing standards offer remarkable connectivity and together far exceed the 10Gb/s peak bandwidth of Thunderbolt. These solutions meet and exceed the bandwidth utilization of many peripherals," an AMD spokesperson told Xbit Labs. "The DisplayPort1.2 standard offers up to 17Gb/s of peak bandwidth for displays. [...] Many AMD-based platforms support USB 3.0 which offers 4.8Gb/s of peak bandwidth, AMD natively supports SATA 6Gb/s with our 8-series chipsets. [Meanwhile], the total bandwidth stated for a Thunderbolt channel is only 20% higher than one PCI Express 3.0 lane and about 52% higher than a single USB 3.0 port. Employing Thunderbolt in the DisplayPort connector implementation decreases the bandwidth available for DisplayPort reducing the bandwidth available for various multi-display configurations. Consumers generally benefit by having standard, high-speed ports available on their mobile devices. Proprietary ports, or the requirement of a dongle to employ those industry-standard ports may be an obstacle to consumers having the full computing experience at home or on the road."

Currently there is an overall lack of devices that can take advantage of Thunderbolt, though Intel is working hard to change this. Does AMD have a point, or is it just worried?





User Comments: 16

Got something to say? Post a comment
Raswan Raswan said:

How about both?

mario mario, Ex-TS Developer, said:

AMD does have a point, but every technology will have it's detractors with their counterpoints specially if they are pushing their own technologies.

I think Thunderbolt will be a success and AMD is really worried, as they should be Intel has never been a fair player.

Kenrick said:

I guess AMD is worried that thunderbolt would leave usb 3.0 behind. If that will happen, AMD will pay royalty fees to intel just to see thunderbolt in amd systems. well, just my opinion folks.

And also, from the article

"Employing Thunderbolt in the DisplayPort connector implementation decreases the bandwidth available for DisplayPort reducing the bandwidth available for various multi-display configurations. "

Partly true, In desktops you can have both displayport and thunderbolt so it won't affect multi-display configurations But In laptops, due to design limitation, thunderbolt's multi-port will have more benefits.

BrianUMR said:

kenrick said:

Partly true, In desktops you can have both displayport and thunderbolt so it won't affect multi-display configurations But In laptops, due to design limitation, thunderbolt's multi-port will have more benefits.

I think this is the important part, that for laptops you need far less range in ports. Look at the impact USB had on laptops where it allowed you to connect pretty much everything you had to your laptop and it didn't have to be a huge laptop. Thunderbolt is aiming to take it one step further making the variety in ports even less.

princeton princeton said:

Downplaying because they know they cant design anything superior. Keep it classy AMD.

Vrmithrax Vrmithrax, TechSpot Paladin, said:

kenrick said:

Partly true, In desktops you can have both displayport and thunderbolt so it won't affect multi-display configurations But In laptops, due to design limitation, thunderbolt's multi-port will have more benefits.

I agree with that, in principle, and even expanding the concept could mean that just having a few Thunderbolt ports on mobile devices would simplify designs and ultimately cut costs. Personally, the flexibility and simplicity of Intel's technology on this interface has always impressed me.

However, just take a look at how much whining and moaning is thrown into the air when devices "don't have enough ports" or, like the iPad, require a dongle for any kind of connection. It becomes a big detriment, and is used as a fighting point in sales against that product. Intel may have a bit of an uphill PR climb to overcome the dongle stigma with Thunderbolt.

Relic Relic, TechSpot Chancellor, said:

I find it a mix of both, since they do have valid points.

Vrmithrax said:

However, just take a look at how much whining and moaning is thrown into the air when devices "don't have enough ports" or, like the iPad, require a dongle for any kind of connection. It becomes a big detriment, and is used as a fighting point in sales against that product. Intel may have a bit of an uphill PR climb to overcome the dongle stigma with Thunderbolt.

Could be that proprietary dongles are sold separately and at a higher cost which is irritating.

On a side note, every time I read the name Thunderbolt I think of that Lightningbolt dude lol =/.

Jibberish18 said:

Meanwhile], the total bandwidth stated for a Thunderbolt channel is only 20% higher than one PCI Express 3.0 lane and about 52% higher than a single USB 3.0 port.

I don't understand, he uses 2 existing ports for an example yet both are slower than Thunderbolt. Is he saying that Thunderbolt should be faster than it is now or that existing standards are fast enough? I'm sort of confused by the remarks. Furthermore, when he talks about the use of a dongle to operate standard ports....well at least you CAN still operate those ports. As long as the dongle isn't big and stupid. Either way eventually people will have to update to a higher standard at some point and honestly Thunderbolt looks to surpass USB 3.0 in every way. Besides that USB is always theoretical speeds, never reality. Tell me how fast the fastest USB 2.0 is you've ever seen?

wizardB wizardB said:

Intel only created thunderbolt because it didn't own USB3 it isn't really the best thing since sliced bread and then some

dividebyzero dividebyzero, trainee n00b, said:

@Jibberish18

AMD-speak from another PR flak...Did you expect it make sense?

DisplayPort 1.2 = 17.28Gb/sec -all good but video only and it's download only (assuming we're talking about four PCIE lanes -if not then it's 5.18Gb/sec (1 lane), 8.64Gb/sec (2 lanes)

Thunderbolt = 10Gb/sec upload AND download (20Gb/sec total bandwidth).

USB 3.0 = (max theoretical) 4.8Gb/sec. 10Gb/sec is not "52% higher" than 4.8Gb/sec (try 108%).

PCI-E 3.0 is 8Gb/sec. 10Gb/sec is 25% higher than 8Gb/sec -not 20%.

While Thunderbolt isn't the be all and end all -the bidirectional 10Gb/sec is cumulative of every device plugged into the daisy-chain for instance, having some talking head spouting random numbers might not exactly be helping AMD's cause. I would also note that there seems nothing to stop a vendor having more than one Thunderbolt controller if the chipset/CPU has the available PCIE lanes. Given that one TB contoller can handle six connected devices simultaneously (bandwidth permitting) I'd say that giving up some I/O interfaces that use the same resources (USB 2.0, PCI, PCIex1/x4, Firewire, eSATA) shouldn't be a big deal.

Intel only created thunderbolt because it didn't own USB3 it isn't really the best thing since sliced bread and then some

Considering Intel brought the world USB that sounds like an incredibly stupid statement, given that neither is Intel charging a royalty/icensing fee for using the interface:

[link]

princeton princeton said:

dividebyzero said:

@Jibberish18

AMD-speak from another PR flak...Did you expect it make sense?

DisplayPort 1.2 = 17.28Gb/sec -all good but video only and it's download only (assuming we're talking about four PCIE lanes -if not then it's 5.18Gb/sec (1 lane), 8.64Gb/sec (2 lanes)

Thunderbolt = 10Gb/sec upload AND download (20Gb/sec total bandwidth).

USB 3.0 = (max theoretical) 4.8Gb/sec. 10Gb/sec is not "52% higher" than 4.8Gb/sec (try 108%).

PCI-E 3.0 is 8Gb/sec. 10Gb/sec is 25% higher than 8Gb/sec -not 20%.

While Thunderbolt isn't the be all and end all -the bidirectional 10Gb/sec is cumulative of every device plugged into the daisy-chain for instance, having some talking head spouting random numbers might not exactly be helping AMD's cause. I would also note that there seems nothing to stop a vendor having more than one Thunderbolt controller if the chipset/CPU has the available PCIE lanes. Given that one TB contoller can handle six connected devices simultaneously (bandwidth permitting) I'd say that giving up some I/O interfaces that use the same resources (USB 2.0, PCI, PCIex1/x4, Firewire, eSATA) shouldn't be a big deal.

Intel only created thunderbolt because it didn't own USB3 it isn't really the best thing since sliced bread and then some

Considering Intel brought the world USB that sounds like an incredibly stupid statement, given that neither is Intel charging a royalty/icensing fee for using the interface:

[link]

Leave it to DBZ to debunk every ***** in the thread and educate us at the same time.

Kenrick said:

AMD should hire dividebyzero as their PR. dividebyzero's statements makes sense than AMD.

Lokalaskurar Lokalaskurar said:

dividebyzero said:

Intel only created thunderbolt because it didn't own USB3 it isn't really the best thing since sliced bread and then some

Considering Intel brought the world USB that sounds like an incredibly stupid statement, given that neither is Intel charging a royalty/icensing fee for using the interface:

Well... Intel has done some pretty stupid things in the past. Not trying to argue against your comment or anything (it's probably the best post I've ever seen on TechSpot ), but perhaps you are familiar with the P5=P1? Intel has its special drawbacks on some pretty simple things really.

dividebyzero dividebyzero, trainee n00b, said:

I don't believe anyone- most especially myself- is arguing that Intel has had flawless execution throughout it's existance. Quite what the FDIV bug (which is what I assume you're driving at) has to do with USB and/or Thunderbolt escapes me.

BTW: You do know that FDIV was basically a hypothetical bug. I don't think I have ever heard of a documented case shown by a user of a P5. Absolutely amazing that the number of people that use FDIV as a whipping boy for Intel when the number of affected users was effectively nil, and Intel have some quite sizeable questionable decisions in their history that the same people seem completely ignorant of.

Feel free to start a thread calling out Intel's missteps. I'll start you off:

Larrabee

Cougar Point B2

Persevering with Netburst past it's use by date

Itanium

non-politically correct advertising

Antitrust suit with AMD

Antitrust suit with FTC

Locking Nvidia out of the chipset business

Shall I start preparing lists for AMD, IBM and the rest of the processor makers? We can turn the forum into a massive fanboy clusterf*ck.

princeton princeton said:

dividebyzero said:

I don't believe anyone- most especially myself- is arguing that Intel has had flawless execution throughout it's existance. Quite what the FDIV bug (which is what I assume you're driving at) has to do with USB and/or Thunderbolt escapes me.

BTW: You do know that FDIV was basically a hypothetical bug. I don't think I have ever heard of a documented case shown by a user of a P5. Absolutely amazing that the number of people that use FDIV as a whipping boy for Intel when the number of affected users was effectively nil, and Intel have some quite sizeable questionable decisions in their history that the same people seem completely ignorant of.

Feel free to start a thread calling out Intel's missteps. I'll start you off:

Larrabee

Cougar Point B2

Persevering with Netburst past it's use by date

Itanium

non-politically correct advertising

Antitrust suit with AMD

Antitrust suit with FTC

Locking Nvidia out of the chipset business

Shall I start preparing lists for AMD, IBM and the rest of the processor makers? We can turn the forum into a massive fanboy clusterf*ck.

LOL I've never seen that intel ad before. Not very P.C guys.

Also DBZ for techspot supreme ruler 2011.

pixelstuff pixelstuff said:

I think the HD BaseT connector specification is the way to go. http://www.hdbaset.org

Essentially replacing HDMI or DisplayPort with Cat5/RJ45 and allowing 100M cable lengths.

Hehehe. Can't wait to see Monster Cable's marketing attempt with that.

Load all comments...

Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...
Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.