If the NFS client has more than one network interface, it's not possible to specify which interface is used for the mount or data transfer unless different IP's have different routes. Tests

bigfile is 5629710336 bytes or 5.24 GB. Local interface eth0 has IP 10.111.63.201 with 1Gb speed. Interface bondeth0 has IP 10.111.92.101 with 10Gb speed.

$ mount | egrep 'fast|slow'
mynfsserver:/ifs/is/dmp1_exp on /fast type nfs (rw,rsize=32768,wsize=32768,clientaddr=10.111.92.101,addr=10.113.19.47)
mynfsserver:/ifs/is/dmp1_exp on /slow type nfs (rw,rsize=32768,wsize=32768,clientaddr=10.111.63.201,addr=10.113.19.58)

$ ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'; time cp bigfile /fast; ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'
          RX bytes:102662765939 (95.6 GiB)  TX bytes:16062673800 (14.9 GiB)        <-- for eth0
          RX bytes:29788698069110 (27.0 TiB)  TX bytes:121164401087449 (110.1 TiB) <-- for bondeth0

real    0m40.321s
user    0m0.042s
sys     0m8.135s
          RX bytes:102662789095 (95.6 GiB)  TX bytes:16062679667 (14.9 GiB)
          RX bytes:29788799390543 (27.0 TiB)  TX bytes:121170341939240 (110.2 TiB)
$ ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'; time cp bigfile /slow; ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'
          RX bytes:102662813123 (95.6 GiB)  TX bytes:16062695063 (14.9 GiB)
          RX bytes:29788810951542 (27.0 TiB)  TX bytes:121170369710628 (110.2 TiB)

real    0m46.436s
user    0m0.021s
sys     0m7.101s
          RX bytes:102662894244 (95.6 GiB)  TX bytes:16062719069 (14.9 GiB)
          RX bytes:29788910889321 (27.0 TiB)  TX bytes:121176303875504 (110.2 TiB)

In both fast and slow cases, of all the 4 numbers, only TX (transfer) for bondeth0 has gone up by a little more than 5 GB.

$ ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'; time cp /fast/bigfile .; ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'
          RX bytes:102663833940 (95.6 GiB)  TX bytes:16062890727 (14.9 GiB)
          RX bytes:29789178594110 (27.0 TiB)  TX bytes:121176826792260 (110.2 TiB)

real    0m34.779s
user    0m0.065s
sys     0m22.260s
          RX bytes:102663856218 (95.6 GiB)  TX bytes:16062892207 (14.9 GiB)
          RX bytes:29794916570270 (27.0 TiB)  TX bytes:121176991315154 (110.2 TiB)
$ ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'; time cp /slow/bigfile .; ifconfig eth0 | grep 'RX bytes'; ifconfig bondeth0 | grep 'RX bytes'
          RX bytes:102663950902 (95.6 GiB)  TX bytes:16062925881 (14.9 GiB)
          RX bytes:29794991190449 (27.0 TiB)  TX bytes:121177168197974 (110.2 TiB)

real    0m36.907s
user    0m0.068s
sys     0m21.588s
          RX bytes:102663975015 (95.6 GiB)  TX bytes:16062932014 (14.9 GiB)
          RX bytes:29800731391185 (27.1 TiB)  TX bytes:121177335571929 (110.2 TiB)

Regardless using fast or slow interface, of all 4 numbers, only RX (receive) for bondeth0 has gone up by a little more than 5 GB.

The fast interface does seem to be about 10% faster than the slow one in copying this file. But it may be because the clientaddr NFS mount option still does some security or metadata check.