# NB: The timing numbers in this architecture file have been modified # to comply with our NDA with the foundry providing us with process # information. The critical path delay output by VPR WILL NOT be accurate, # as we have intentionally altered the delays to introduce inaccuracy. # The numbers are still reasonable enough to allow CAD experimentation, # though. If you want real timing numbers, you'll have to insert your own # process data for the various, R, C, and Tdel entries. # A simple architecture, for use in the FPGA CHALLENGE. # All wires span only one logic block, and all routing switches are tri-state # buffers. # The logic block is a single 4-LUT + FF. # Uniform channels, 2 pads per row or column. io_rat 2 chan_width_io 1 chan_width_x uniform 1 chan_width_y uniform 1 # 4-input LUT. LUT inputs first, then output, then clock. # Each input pin appears on one side, while the output is # on the bottom and right sides. Note that the 4 LUT inputs are # all made logically-equivalent by declaring them to be the same # class. inpin class: 0 bottom inpin class: 0 left inpin class: 0 top inpin class: 0 right outpin class: 1 bottom right inpin class: 2 global top # Clock; shouldn't matter. subblocks_per_clb 1 subblock_lut_size 4 #parameters needed only for detailed routing. switch_block_type subset Fc_type fractional Fc_output 1 Fc_input 1 Fc_pad 1 segment frequency: 1 length: 1 wire_switch: 0 opin_switch: 0 Frac_cb: 1. \ Frac_sb: 1. Rmetal: 4.16 Cmetal: 81e-15 # Switch used as a tri-state buffer within the routing, and also as the # output buffer used to drive from a CLB output pin to a routing wire. switch 0 buffered: yes R: 786.9 Cin: 7.512e-15 Cout: 10.762e-15 \ Tdel: 456e-12 # Used only by the area model. R_minW_nmos 1967 R_minW_pmos 3738 # Timing info below. C_ipin_cblock 7.512e-15 T_ipin_cblock 1.5e-9 T_ipad 478e-12 # clk_to_Q + 2:1 mux T_opad 295e-12 # Tsetup T_sblk_opin_to_sblk_ipin 0 # No local routing T_clb_ipin_to_sblk_ipin 0 # No local routing T_sblk_opin_to_clb_opin 0. # Delays through the BLE (LUT and a FF) T_subblock T_comb: 546e-12 T_seq_in: 845e-12 T_seq_out: 478e-12