This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Performance of USB 3.0 controller TUSB7340

Other Parts Discussed in Thread: TUSB7340

The USB 3.0 host controller TUSB7340 has four USB ports which can run at SuperSpeed, i.e. 5 Gbps. On the other side only one PCIe x1 interface is available, which also can run at 5 Gbps (forget the protocoll overhead for a moment).

Is it correct, that it is not possible to run more than one interface at SuperSpeed at the same time? Why didn't you implement a PCIe x4 interface, which would allow up to four ports to run at it highest speed?


  • All of the ports on a USB host controller share the bandwidth since USB is a host-centric bus, only one USB port is transferring data at one time.  On an EHCI controller, all of the high speed ports share 480 Mbps.  On an xHCI controller, all of the SuperSpeed ports share 5 Gbps.  So one lane of PCIe Gen2 is enough to guarantee 5Gbps on any active SS USB port.

    It is possible to design an xHCI controller with multiple bus instances such that not all ports share the same bandwidth, however, we did not do that for the 7340 since having multiple bus instances increases the die size of the controller.  You will find that none of the current xHCI hosts on the market are implemented in this way.

  • USB is a master-slave point-to-point bus. 

    The host controller can only be servicing one peripheral at a time and therefore only one of the four downstream ports is actually active at any instant in time and therefore the upstream PCIe bandwidth does match the active USB bandwidth at any instant in time.