Merge branch 'tensor_core' into rtl

hw/rtl/core/VX_operands.sv

@@ -294,6 +294,27 @@ module VX_operands import VX_gpu_pkg::*; #(
                 .raddr (gpr_rd_addr),
                 .rdata (gpr_rd_data[j])
             );
+
+            // blast read register file because printf is slowge
+            logic [31:0] cycle, cycle_n;
+            assign cycle_n = cycle + 32'd1;
+            always @(posedge clk) begin
+                if (reset) begin
+                    cycle <= '0;
+                end
+                else begin
+                    cycle <= cycle_n;
+                end
+
+                if (cycle == 32'd25000) begin
+                    for (integer k = 0; k < `NUM_REGS * ISSUE_RATIO; ++k) begin
+                        integer warp = i * ISSUE_RATIO + (k / `NUM_REGS);
+                        integer thread = j;
+                        integer register = k % `NUM_REGS;
+                        $display("warp %0d, thread %0d, register %0d: %0x", warp, thread, register, gpr_ram.ram[k]);
+                    end
+                end
+            end
         end
     end