By: 309 (309.delete@this.example.com), July 27, 2012 10:34 pm
Room: Moderated Discussions
Why only look at double-precision operations?
DP is only used in some scientific/HPC applications, while more common applications like games use single-precision and integer, while the rest tends to be mostly or exclusively integer (and most of integer is 32-bit, except for bignum computations and some special cases like chess).
How about looking at the maximum between 32-bit float and 32-bit integer instead?
DP is only used in some scientific/HPC applications, while more common applications like games use single-precision and integer, while the rest tends to be mostly or exclusively integer (and most of integer is 32-bit, except for bignum computations and some special cases like chess).
How about looking at the maximum between 32-bit float and 32-bit integer instead?
