2.5/5 GBps USB Ethernet Adapter Announced

[TheLostSwede]

Given a choice between Realtek and Marvell, I'll take Marvell any time...

The conversation here is really about the Aquantia adapter and chipset - and this is a price/performance breakthru...

Hey, you were the one saying you wanted a chipset switch, I simply replied and said it might happen soon, as I'd seen some other new stuff.
I'll think twice about posting something like that next time, as it's apparently unwanted information...

 

[System Error Message]

You seem to live in a parallel universe... There has never ever been a 2.5Gbps USB 3.0 standard. I have met all of the USB 3.0 chip makers over the years and I was the first and I think only journalist that did a comparison of all the different USB 3.0 host controller available at the point when USB 3.0 was still new-ish. Unfortunately the publication I worked for back then has managed to lose all the pictures, so not much point in sharing it. I tested the the AMD A75 chipset, the ASMedia ASM1042, the Etron EJ168A, the Fresco Logic FL1009, the NEC/Renesas µD720200, the Renesas µD720201 and the VLI VL800 in 2011 and none of them had any other rating than 5Gbps. Back then it wasn't possible to reach the speeds of today, as the device controllers weren't fast enough. Fujitsu had a solution that could do 3.2Gbps/400MB/s. However, today it's not so hard with a nice fast SSD and a moder device controller.

I also don't think that you understand the technical aspects of how a motherboard chipset works, but let me see if I can help with that.
Below is a very rough layout of the available lanes/ports inside the Intel Skylake chipset (more recent chipsets from Intel adds USB 3.1 gen 2).

So as you can see, each of the USB ports that comes from a chipset are allocated one of 10 possible lanes. These lanes have the same width as a PCIe lane. So yes, the physical USB ports might connect directly to the chipset, but in the case of Intel, a USB 3.1 gen 2 port would only have one lane of PCIe 3.0 worth of bandwidth available to it, as it wouldn't make sense for Intel to make a custom interface for USB that's faster than PCIe 3.0. This means that if you have an Intel chipset, you'll never see the maximum performance from any chipset connected USB 3.1 gen 2 ports. I would hazard a guess and say that it's likely that AMD has the same limitation, but as they haven't provided any documentation, I can't say for certain. This means that only a couple of third part PCIe 3.0 x2 host controllers can offer near full USB 3.1 gen 2 performance.
More about Intel's 300-series chipsets and lane integration here - https://videocardz.com/76275/intel-releases-full-specifications-of-z390-chipset

Oh, one last thing, the interconnect between the chipset and the CPU today is simply four PCIe 3.0 lanes with a fancy name, nothing more, nothing less. So saying this is a fast interface is simply not true any more. Btw, a Northbrdige hasn't been used in PC architecture for around 8-9 years now, so I think you need to catch up on your tech know how.

USB implementation is very different, also AMD ryzen has 2 USB3 gen 2 (10Gb/s) ports connected directly to the CPU.

https://www.dell.com/community/Desk...-USB-3-0-NEC-Renesas-Electronics/td-p/3908330
https://www.tomshardware.com/reviews/nec-controller-usb-3-pd720200,2682.html
Early controllers were limited to 2.5Gb/s (PCIe limitations) or 2.5Gb/s per direction. The USB 3 standard defines 5Gb/s total but also means 5Gb/s in a single direction as well. This will bottleneck a 5Gb/s full duplex NIC as PCIe 2.0 x1 is 2.5Gb/s (some says 2Gb/s) per direction.

 

[sfx2000]

Hey, you were the one saying you wanted a chipset switch, I simply replied and said it might happen soon, as I'd seen some other new stuff.

The info is always welcome

The realtek 10Gb switch chip you found would be a great topic in and of itself - it was a good find

 

[TheLostSwede]

USB implementation is very different, also AMD ryzen has 2 USB3 gen 2 (10Gb/s) ports connected directly to the CPU.

https://www.dell.com/community/Desk...-USB-3-0-NEC-Renesas-Electronics/td-p/3908330
https://www.tomshardware.com/reviews/nec-controller-usb-3-pd720200,2682.html
Early controllers were limited to 2.5Gb/s (PCIe limitations) or 2.5Gb/s per direction. The USB 3 standard defines 5Gb/s total but also means 5Gb/s in a single direction as well. This will bottleneck a 5Gb/s full duplex NIC as PCIe 2.0 x1 is 2.5Gb/s (some says 2Gb/s) per direction.

Did you actually test this, or are you just relying on information from sources that have done flawed testing to prove your point?
I actually did the testing as I said and much better than that half arsed article you linked to, but unfortunately I can't prove it, as the publication I wrote it for has messed it all up.

And how do you think AMD's USB 3.1 gen 2 host controller interfaces to the CPU? Via magic? I think it's via PCIe.

Not even the first NEC USB 3.0 controller used PCIe 1.x. In facto, only Fresco Logic released a PCIe 1.1 based USB 3.0 controller . You're mixing up apples with kumquats in this case and clearly have little understanding of the technology and as such, I'm going to stop this silly thread here.

 

[System Error Message]

Did you actually test this, or are you just relying on information from sources that have done flawed testing to prove your point?
I actually did the testing as I said and much better than that half arsed article you linked to, but unfortunately I can't prove it, as the publication I wrote it for has messed it all up.

And how do you think AMD's USB 3.1 gen 2 host controller interfaces to the CPU? Via magic? I think it's via PCIe.

Not even the first NEC USB 3.0 controller used PCIe 1.x. In facto, only Fresco Logic released a PCIe 1.1 based USB 3.0 controller . You're mixing up apples with kumquats in this case and clearly have little understanding of the technology and as such, I'm going to stop this silly thread here.

i had one of those dell XPS with said NEC usb 3 controller, and 2.5Gb/s was actually the spec listed on those ports with 4.8Gb/s being the theoratical max if both directions were used.
Telling you from experience.

Also https://www.asus.com/my/Motherboards/ROG-STRIX-B350-F-GAMING/specifications/ I have this motherboard and an AMD ryzen 7.
https://www.bit-tech.net/reviews/tech/amd-ryzen-1800x-and-am4-platform-review/2/

There you go, USB ports directly connected to CPU with the controller on the CPU itself. You are wrong, those ports on the CPU arent using PCIe, for instance look at many mobile or not SoCs that have everything integrated. They communicate via some bus that usually either gets assigned or talks PCI but unlike PCI or PCIe the interconnects are really fast, just PCI is in the OS as a protocol and way of communication (easier on the driver side). Physically, everything on the SoC or CPU gets connected via bus and not PCI or PCIe. Even for PCIe devices the drivers and OS will say PCI as thats the protocol used.

I study architecture for fun, i like looking at block diagrams. Even if PCIe is used on the inside, just look at the silicon level diagram where it will highlight different areas, and if you look at the gaps you'd see something that looks like a communications chip or registers or in some cases, power regulators but those never get highlighted, you'd need to recognise the pattern to know what you're looking at.

 

[sfx2000]

There you go, USB ports directly connected to CPU with the controller on the CPU itself.

Not so fast - Ryzen has USB 3.0 ports, but the AM4 platform can offer the USB3.1Gen2 via the Asmedia chipset on the B350 - that is a PCIe 3.0 4-lane interface. Really depends on whether the vendor wants to populate the ports or not - and Ryzen can support TB3 if they integrate Intel's Alpine Ridge USB3.1g2/TB3, but that starts to get odd if one is running an APU and using onboard rather than discrete graphics (but again, Alpine Ridge is going to be over PCIe)

In any event - the B350 can support this 10Gbe dongle... maybe not at full speed, depending on the mainboard and choices there, I'm thinking this is more for the 2.5/5.0 crowd, and there it's a good thing...

@System Error Message - @TheLostSwede knows what he's talking about, better to listen and consider, rather than argue what color a shed should be painted. You are both valued forum members here...

 

[System Error Message]

Not so fast - Ryzen has USB 3.0 ports, but the AM4 platform can offer the USB3.1Gen2 via the Asmedia chipset on the B350 - that is a PCIe 3.0 4-lane interface. Really depends on whether the vendor wants to populate the ports or not - and Ryzen can support TB3 if they integrate Intel's Alpine Ridge USB3.1g2/TB3, but that starts to get odd if one is running an APU and using onboard rather than discrete graphics (but again, Alpine Ridge is going to be over PCIe)

In any event - the B350 can support this 10Gbe dongle... maybe not at full speed, depending on the mainboard and choices there, I'm thinking this is more for the 2.5/5.0 crowd, and there it's a good thing...

@System Error Message - @TheLostSwede knows what he's talking about, better to listen and consider, rather than argue what color a shed should be painted. You are both valued forum members here...

What im arguing about is that there are USB3 controllers limited to 2.5Gb/s. Earlier laptops used PCIe x1 interfaces and he also talked about controller limitations, and some usb3 cards also use the same PCIe interface as well limiting those speed. There may be nothing wrong in the controller itself syncing at 5Gb/s but could be limited in some ways. Im not talking about the usb3 spec which initially started with 4.8Gb/s or 5Gb/s however one sees it but the varied implementation which can cause issues if one tries to use a 5Gb/s NIC and pass 5Gb/s in one direction or both ways at the same time. Its the same issue with USB-C, its where a lot of manufacturers cheap out on where they say its USB-C when its just a regular 5Gb/s usb3 port in usb-c form.

With regards to the B350 and x370 boards, some ports on the board itself will only work with certain CPUs. For instance on my motherboard there are 2 usb3 gen 2 10Gb/s ports connected to the CPU, if the CPU doesnt have those ports they wont work. I am fairly certain that most CPUs and SoC integrated ports will run at their advertised speed as there is literally almost no bottleneck inside but i've not seen a CPU or SoC use PCIe to connect everything inside because PCIe footprint is fairly big that CPU manufacturers on mainstream dont give more than 20 lanes (or 24 lanes for ryzen). PCIe is interesting in that it is very much like a network switch where the lanes are all links. In the case of ryzen and intel iseries they use PCIe to connect to other parts because it is cheaper for instance 4 lanes for m.2 and 4 lanes for the chipset as there already are PCIe lanes put there for the GPU, so 8 extra lanes for other external devices rather than a DMI or some other high speed bus reduces cost. On the inside or anything integrated however, PCIe footprint is just too big.

Each PCIe 3.0 lane supports 5Gb/s per direction, PCie 2.0 supports 2.5Gb/s per direction. with an x4 link at PCIe 3.0 it will easily support 10Gb/s as you only need 2 PCIe 3.0 lanes for 10Gb/s. So thanks to PCIe 3 there will be no issues with the new adapters, only with older PCIe 2.0 and some controllers. I remember a few years ago VIA PCMCIA sata would run at a few Mb/s if you used their drivers but standard windows drivers/ software was far faster because the controller itself was super slow. So what im arguing here is just the manufacturer implementation that can cause issues with all the new high speed devices.

 

[sfx2000]

@System Error Message - let it go man