With regards to the query “128-Bit Precision with GCC” in the Best of Tech column of the December 2001 issue, the reply was neither correct nor helpful—GMP is not the equivalent and using GMP means rewriting code. Some compilers, AIX amongst them, will carry out calculations using 64 bits (or 128 bits, depending on the processor) with the appropriate option. (Using this, means that the product of two 32-bit numbers will always fit within an int. No code modification is needed.) GMP, however, is a multi-precision package that defines certain data structures wherein the multi-precision numbers are put. To use it, you must extensively rewrite your code. Some compilers, including gcc, accept the “long long” extension and use 64 bits for calculation but that still requires modifying your code and raises portability questions. My answer would be that, sadly, there is no equivalent without some sort of code modification. Such an option to gcc would be nice, though.
Upon rereading my letter [above], I realized that it reads far harsher than I intended. Both the column and the individual answering the query have done your readers good service in the past. I did not mean to slight either, and I apologise if anyone took amiss. I merely wished to indicate that the solution is by no means simple.