Processes/programs that want to receive network TCP/IP traffic will tell the operating system one of the following: "listen" on one IP+specific port, "listen" on more than one IP+specific port, or "listen" on any IP+specific port (the IP address 0.0.0.0 is usually used to tell the OS to receive traffic from any IP).
And that's the whole point of ports: so that the TCP/IP stack on the receiving system can decide which process/program gets the traffic. Without ports you'd have to have one process listening to the NIC, receiving all traffic and then distributing it to programs, which is slower than if simply done by the OS.
Normally speed and throughput are limited by hardware and the capability/reliability of any network medium between source and destination. TCP adapts to unknown/unpredictable conditions by use of the sliding window algorithm. UDP (a thin layer on top of plain IP) offloads that responsibility to the processes/programs if they want anything like that.
Anyway, modern OSes (especially Linux) and things like firewalls/management devices have ways to limit traffic speed on specific criteria, and ports can be one of those criteria.
So if there is something limiting traffic per port, such as some type of traffic control policy or software, then using more than one port can break through that limit, assuming nothing else is interfering.
Most people smart enough to set up a speed limit per port will also do it per IP address so you're unlikely to bypass anything there.