Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

ARITH_BITLEN rule of Lint (VLSI)

Not open for further replies.
Aug 16, 2020
Reaction score
Trophy points
Activity points
ARITH_BITLEN: This rule checks for inconsistent bitlengths of an arithmetic operator. A violation is reported when the bitlengths of arithmetic operands are not the same.
Case1: in1 is of 3-bit and in2 is of 4-bit. I understand when there'll be violation of this rule for in1+in2.
Case2: in1 and in2 are of 5-bit and sum = (in1/5'd2+in2/5'd4). Lint doesn't report with this scenario. With the denominator (5'd2 for in1 and 5'd4 for in2), the bit width will be different for 2 operands. Why is lint not able to report this? This is also a static check right?

I can't really understand your Case2, but it looks like you adding constants of fixed sizes. What were you expecting to be the error/warning here?

Not sure what ARITH_BITLEN rule specifies. Addition of operands with different bit length is generally allowed in Verilog, expression bitlength is max(length op1, length op2).

Addition of different bit length is allowed but lint throws a warning in case 1. But in case 2, when we divide in1 with 5'd2 and in2 with 5'd4 then the width of in1[4:0]/5'd2 (i.e., 4 bits are enough for resultant when we right shift the in1 by a bit right?) is different from width of in2[4:0]/5'd4 (i.e., 3 bits are enough for resultant when we right shift in2 by 2 bits). Why isn't lint giving warning when we add 4-bit operand with 3-bit operand?

Also, as @ThisIsNotSam mentioned, if we waive this rule do we've any concern? (with respect to the extended bits and any unintentional errors)?

Not open for further replies.

Part and Inventory Search

Welcome to