What is Intel’s Thunderbolt 3 dual port controller’s total bandwidth?

https://communities.intel.com/message/560462

by roger- • 8 years ago

111 55 8 years ago

communities.intel.com

dmitrygr• 8 years ago

The first response posted is absolutely hilarious and perfectly summarizes basically every educated person's attempt to get tech support from any large company.

rayiner4• 8 years ago

I tried to get my static IPv4 and IPv6 addresses from Comcast the other day. The customer rep asked me what I meant by "static IP address." And I'm like--wait, how much do I actually explain about what a static IP is? I might as well have been talking to the cat that hangs out on my porch, but that cat was the only one who could kick me up to someone who knew what she was talking about...

pard68• 8 years ago

Did they give you static addresses? I would love one, but figure it would be a no.

userbinator• 8 years ago

It's rather sad that customer support for an ISP would not know what a static IP is. Then again, they're probably usually telling more "average" users to reset their modems and asking if things are turned on and plugged in...

pvg• 8 years ago

You probably didn't get on Comcast forums and start being a passive aggressive all-caps jerk about it, though.

grecy• 8 years ago

Oh, you have no idea.

I worked for a large telco and the new Manager of IT had to stop her first "all hands" meeting to ask what IPv6 is.

Seriously.

dctoedt• 8 years ago

kick me up to someone who knew what she was talking about...

Obligatory reference: https://xkcd.com/806/, which the other day I had occasion to pass along to a really-great [seriously] AT&T U-Verse TSR who'd never heard of it (XKCD). This one is in the running for greatest XKCD of all time.

scarface743• 8 years ago

I had an issue with AWS. First, I went to our contracted AWS support company.

<rant> They are completely clueless, you can tell that they a bunch of on prem net ops people who took a class in AWS and just translated our AWS architecture to exactly what they would have done on prem. This was before either my AWS savvy manager or I were hired. You ask them about anything that involves anything about AWS besides EC2 instances, security groups and subnets they just shrug. I’ve spent the last few months cleaning up things to try to save the company money. </rant>

Then I went to AWS directly. Thier online chat might as well been a chatbot. They just regurgitated the FAQ.

wmf4• 8 years ago

Thunderbolt is a quite under-documented protocol. I realize it's proprietary (for now) but even by those standards there is almost nothing publicly available. Maybe Intel doesn't want people to know that a dual-port "40 Gbps" Thunderbolt controller only has 32 Gbps of PCIe bandwidth total and according to this question each port only supports 22 Gbps. This may explain why real-world benchmarks don't show performance anywhere near 40 Gbps (although TB3 is generally still faster than any other consumer protocol).

kev009• 8 years ago

Without an IOMMU it's a gigantic security risk. I don't know that withholding documentation does anything to slow that down in practice but I wonder if it's somehow part of the obscurity.

daurnimator• 8 years ago

As far as I know nearly all recent Intel chips have an IOMMU: are you saying thunderbolt bypasses it?

Also, there is the thunderbolt authorization protocol (see: boltctl on Linux): is that broken in any way?

ztjio3• 8 years ago

I don't think this is right. I've come to understand the 40gbps number to be the total bidirecional bandwidth with overhead taken into account on a single port.

Which adds up, because, a single port maxes at 22gbps FULL DUPLEX shave off a 10% overhead factor, and double it = 40gbps.

vardump2• 8 years ago

If that's true, how it's possible for TB3 to drive dual 4k monitors at 24 bpp 60 Hz and have enough bandwidth on top to drive gigabit ethernet etc. peripherals?

(3840 * 2160 * 3B * 2 * 60)/s in Gbit/s = 23.887872 Gbit/s

So two 4k monitors at 60 Hz alone should require almost 24 Gbps, not counting audio streams, etc.

pkaye• 8 years ago

Thunderbolt* is customized by each system manufacturer differently on their platforms. For this reason, our recommendation is that you can contact your system manufacturer to receive accurate information about their products the Thunderbolt* 3 implementation.

This means ask your system manufacturer to give the bad news.

nottorp• 8 years ago

You were patient enough to read the article? ;) I just skipped to the comments because I assumed it's full of marketing lies.

jannes• 8 years ago

The link goes to a forum post, not an article.

fallous3• 8 years ago

Thunderbolt, and Intel's "management" of that interface spec, is an excellent contrasting example of IEEE interfaces vs closed systems. I actively worked on a project building around TB2 (and then 3) for over two years and eventually we gave up on the entire thing due to the obscurity of the protocol, the absence of details of the Intel implementations (the only game in town at the time), and a myriad other roadblocks. By contrast I have some historical familiarity with Sony's Firewire iterations and once it became IEEE-1394 (and variants) it exploded in both popularity and transparency.

If Intel treated the original USB in the same way they do ThunderBolt, we'd still be using ADB, serial, parallel, and PS/2 interfaces on all our machines.

scarface743• 8 years ago

Did FireWire actually "explode" in popularity? It was standard on Macs, Sonys, and most business Dell's at the time but it never got widespread consumer use.

The "problem" with Thunderbolt is the same as with Firewire, it's overkill for most of the consumer market. I don't think that's a deal breaker, people who are willing and able to pay for computers with Thunderbolt and devices have plenty of options.

wmf4• 8 years ago

At least with Thunderbolt the port can be used for other purposes since it's backwards compatible.

fallous3• 8 years ago

Firewire certainly exploded in popularity post-IEEE 1394 compared to when it was a Sony controlled interface with special dispensations for Apple. The number of vendors offering Firewire peripherals as well as the innovative software built around it made it a mainstay in market segments such as video editing and external storage.

Hey, I'm willing to pay for Thunderbolt on an AMD Threadripper (because PCIe lanes matter) and I'm happy to build devices based on it (as noted, that was what I working on for a couple of years). Feel free to show me the options available to achieve that goal absent being a special preferred Intel vendor.

Amusingly, I have some history with Intel... I worked there. I've also seen how they mismanage or abandon anything that isn't core x86.

dryst• 8 years ago

The answer is, it depends. PCIe is not a streaming interface, so the effective bandwidth depends on transaction type, latency, packet size, address depth, etc. etc. The only certainty is that it will not be anywhere close to 32gbps, that is line rate.

kbumsik2• 8 years ago

AFAIR Intel alway is like that. Intel hardly open their documentations to the public because they only communicate with big PC/motherboard manufacturers, not individuals.

Intel did the same thing to hobbyist communities when Intel tried to enter consumer IoT market with Arduino101/Edison/Curie. They did open sourced the code (more than a year late since released the hardware though) but never provided good enough public documentations about the chips. Intel eventually gave up the market.

solarkraft5• 8 years ago

It's always amazing to me when a company has garbage developer support and then wonders why nobody wants to use their stuff or does anything good with it.

wmf4• 8 years ago

I think Intel has good NDA support for paying customers. It's just the public they don't care about.

kbumsik2• 8 years ago

Intel actually is a chip manufacturer that has the best dev support for organizations. e.g. Intel is the biggest contributor for the Linux ecosystem. It is just their business model didn't work well for open-source consumer IoT device market which is driven by hobbyist and relatively small companies.

thought_alarm• 8 years ago

Meanwhile, the bandwidth of a single-port 9-pin Atari controller is almost well understood.

raverbashing• 8 years ago

But does it do 22Gbps?

solarkraft5• 8 years ago

You're giving me an idea.

djsumdog• 8 years ago

Level 1 Techs does a great video on getting Thunderbolt via USB-C working on Ryzen Threadripper. You can run Thunderbolt on Threadripper, but not Ryzen 7.

https://www.youtube.com/watch?v=uOlQbP63lDQ

I think the reason this question so so hard to answer for Intel is because Thunderbold on Intel goes through the chipset, after which it's on a shared PCIx4 lane to the processor. The main reason for this is hotplugging PCI-E is difficult.

On Threadripper, the Leve1Tech use a lot of hacks to get an Intel Thurnderbolt controller connected directly to the PCIE lanes that go straight to the processor.

incadenza• 8 years ago

What if they told you their products are so fast they don’t need a total bandwidth?

Quequau• 8 years ago

I'm still pretty disappointed that AMD's alternative to Intel's Thunderbolt, which I think they called something like "Lightning Bolt", never went anywhere. It struck me then as a cheaper open alternative which could do most of the things people would really want to use Thunderbolt for but without the costs.

izacus• 8 years ago

What would the benefit be though? It seems that while not as fast, USB 3.0 pretty much covers all the usecases and is actually widely deployed due to its lower cost. Would AMD's port fare any better than TB?

nirv• 8 years ago

What would the benefit be though? It seems that while not as fast, USB 3.0 pretty much covers all the usecases

Foremost, eGPU and external RAID storage units. However, only native Thunderbolt 3 implementation makes sense in order to meet compatibility standards.

285tech2• 8 years ago

Same here dude.

285tech2• 8 years ago

https://www.youtube.com/watch?v=hjOg0z2XzPE This was it but never took off. If they re launched it next year or 2020 it could be a game changer as an AMD only protocol. DP, PCIe, USB-C and Power Delivery. Intel is never going to let AMD in on Thuderbolt. Not Officially.

l1k• 8 years ago

A Thunderbolt controller contains a Converged I/O switch (CIO) plus a PCIe switch. The Downstream Ports on the PCIe switch are attached to the CIO switch and represent so-called PCIe adapters which encapsulate PCIe datagrams into Thunderbolt datagrams.

The PCIe throughput is thus the minimum of the CIO switch's capacity, the PCIe switch's capacity and the PCIe adapter's capacity. The PCIe 4 lane interface is only how the Upstream Port on the PCIe switch is attached to the CPU or PCH.

It is very unlikely that Intel will answer this question satisfactorily because they'd reveal details of the controller's internal architecture.

Note that a PCIe tunnel may pass through a Thunderbolt controller without being terminated on it. The CIO switch must then handle that tunnel's traffic in addition to any tunnels terminated on it.

(Disclosure, I've authored a couple of commits in the Linux kernel's Thunderbolt driver.)

solarkraft5• 8 years ago

It is very unlikely that Intel will answer this question satisfactorily because they'd reveal details of the controller's internal architecture.

Amazingly dense. It's like they don't want anyone to use Thunderbolt.

wmf4• 8 years ago

They want you to use it as a black box. It works, just don't ask how it works.

What is Intel’s Thunderbolt 3 dual port controller’s total bandwidth?

Related Stories