cancel
Showing results for 
Search instead for 
Did you mean: 

Testing USB Ports

smithkid
Level 7
I have always had intermittent problems with USB on this MB since the beginning. October 2017.
So I bought a USB3 test plug from Passmark.
I have 2 USB ports on the front of the case connected to the USB 3.1 Gen1 connector. There are 2 of these on this board and I have tried both.

Results of TEST:
36 MB/s on Loopback
46 MB/s on Benchmark

There are 4 ports vertcally next to the WIFI ports which are ASmedia I presume. I have my Logitech unifying receiver in one of these (for my MX Master 2S).
When I plug the Passmark tester into one of these it is not recognised and my Logitech Mouse stops working.

There are next 2 then 2 USB ports that I presume are the Intel ones.
When I plug the Passmark Tester into these the mouse now works but again the tester is not recognised.

Finally we have the ASMedia 3.1 Gen 2 ports. One is a type C and the other type A.
When I plug the Passmark Tester into the type A it is recognised.
Results of TEST:
156 MB/s on Loopback
416 MB/s on Benchmark

Is the Passmark Tester not a good product (not cheap)?

I don't know what I am doing (probably)!

I have the wrong USB drivers!!?? - Using the latest drivers on ASUS ROG site.

Have used the TESTER on my ASUS X99 rig and get similar readings on the front ports.
126 MB/s on Loopback
408 MB/s on Benchmark

:confused:
233 Views
4 REPLIES 4

Korth
Level 14
Are you comparing USB2 ports vs USB3 ports? The numbers are off by nearly a magnitude.

Hardware is the first potential bottleneck. USB ports integrated within the PCH or main chipset are always going to be the best choices because USB ports serviced by chipset add-on parts (like ASMedia hubs) are going to link through the main chipset anyhow. I'm sure there's some variations in USB controller part-to-part implementations, but I don't think it would be much because USB is strictly standardized and is pretty much a collection of old problems solved long ago in a variety of clever ways.

But if you're already saturating the USB controller busses or the PCH or the CPU or the PCH-to-CPU (DMI) bus between them with other activity then the USB data has to queue for priority along with everything else. Ideally, you'd want minimum USB devices connected and minimum load on every processing component in the system. In reality, you're gonna have some things running in the CPU and some data moving through the PCH and drive activity and some other USB devices connected simultaneously. Enough of these things together can make meaningful USB hardware benchmarks difficult to isolate.

Software is the other potential bottleneck. USB standards aren't as strictly enforced on USB drivers, they can pretty much implement whatever they want in any way they want and each USB manufacturer is primarily interested only in assuring their own USB product works properly ... they don't get together and test "conflicting" hardware combinations like they used to ... there's so many layers of USB code revisions (and API revisions and OS revisions) that things are sloppy and good code practices fall to the wayside behind whatever code practices produce "working" results with minimum time and effort. And all it takes it one "bad" USB device to enumerate improperly or hog all the bandwidth or power on the bus to diminish performance of all downstream USB devices. And in Windows, which even Microsoft can barely keep working, let alone countless USB manufacturers, things are really messy indeed.

So ideally you'd want to use the exact same operating system, drivers, and software on both machines to make a meaningful comparison between their USB performances.

Just saying that plugging your tester into two different machines and getting two different results is not informative. You can only get "accurate" measures by isolating and controlling all of the other hardware and software parameters which might affect USB performance.
"All opinions are not equal. Some are a very great deal more robust, sophisticated and well supported in logic and argument than others." - Douglas Adams

[/Korth]

Korth wrote:
Are you comparing USB2 ports vs USB3 ports? The numbers are off by nearly a magnitude.

Hardware is the first potential bottleneck. USB ports integrated within the PCH or main chipset are always going to be the best choices because USB ports serviced by chipset add-on parts (like ASMedia hubs) are going to link through the main chipset anyhow. I'm sure there's some variations in USB controller part-to-part implementations, but I don't think it would be much because USB is strictly standardized and is pretty much a collection of old problems solved long ago in a variety of clever ways.

But if you're already saturating the USB controller busses or the PCH or the CPU or the PCH-to-CPU (DMI) bus between them with other activity then the USB data has to queue for priority along with everything else. Ideally, you'd want minimum USB devices connected and minimum load on every processing component in the system. In reality, you're gonna have some things running in the CPU and some data moving through the PCH and drive activity and some other USB devices connected simultaneously. Enough of these things together can make meaningful USB hardware benchmarks difficult to isolate.

Software is the other potential bottleneck. USB standards aren't as strictly enforced on USB drivers, they can pretty much implement whatever they want in any way they want and each USB manufacturer is primarily interested only in assuring their own USB product works properly ... they don't get together and test "conflicting" hardware combinations like they used to ... there's so many layers of USB code revisions (and API revisions and OS revisions) that things are sloppy and good code practices fall to the wayside behind whatever code practices produce "working" results with minimum time and effort. And all it takes it one "bad" USB device to enumerate improperly or hog all the bandwidth or power on the bus to diminish performance of all downstream USB devices. And in Windows, which even Microsoft can barely keep working, let alone countless USB manufacturers, things are really messy indeed.

So ideally you'd want to use the exact same operating system, drivers, and software on both machines to make a meaningful comparison between their USB performances.

Just saying that plugging your tester into two different machines and getting two different results is not informative. You can only get "accurate" measures by isolating and controlling all of the other hardware and software parameters which might affect USB performance.


You make some very valid points and thank you for your reply. I will retest with all ports "empty" and see if there is a difference.

Sprayingmango
Level 10
I have noticed many strange USB behaviors on the Rampage V Extreme Edition 10 that I had before this board, and currently on the Rampage VI Extreme. Asus did strange things with shared USB ports on the RVE10, there were always issues using an Oculus Rift / over saturating the USB hub.

The RVIE sometimes acts strange when certain devices are plugged in to the ASmedia ports vs the Intel ports.

Sprayingmango wrote:
Asus did strange things with shared USB ports on the RVE10, there were always issues using an Oculus Rift / over saturating the USB hub.


Now you know why Oculus strongly suggested using a particular add-on USB card for the Rift. I never plugged the Rift into the RV10E, only the card and don't have any problems with it. I have also tossed any poorly implemented USB card in the waste bin instead of messing with it trying to get my devices working on it.