I am currently working on some code (written in C) where sin() and cos() may be fed with large arguments. The development environment is Microsoft Visual Studio 2026 under Windows 11, the target platform is Windows x86 (yes, 32 bit).

I'd like to know whether I can expect accurate results from trigonometric functions that are passed large arguments on the target platform if I install the current Microsoft C(++) x86 runtime on the target machines.

For example, in the .NET runtime, this problem has been solved not before 2017, according to this report. Unfortunately, I was not able to find relevant information for the normal Microsoft C(++) runtime. From the linked report, we can eventually conclude that the normal runtime got this problem fixed earlier, but that's not completely clear to me.