Due to a problem in the ModelSim*-Intel® FPGA Edition 2021.4 and Questa* Intel® FPGA Edition 2022.1, a variation in rx_phy_clk frequency leads to the signal rx_gb_underflow_err being asserted.
This problem is observed only in the following variant:
L = 16, M = 8, F = 2, DATA RATE/L = 32000.000000Mbps, FCLK_MULP = 1, WIDTH_MULP = 4
This problem impacts Intel® Quartus® Prime Software IP versions 22.2 and 22.3.
To work around this problem:
For ModelSim*, run the simulation using v2022.1 instead of v2021.4.
For Questa*, run the simulation using v2021.3 instead of v2022.1.
This problem has been fixed in the ModelSim* Intel® FPGA Edition and Questa* Intel® FPGA Edition 22.4.