On x86, glibc's sin(0x1p1023) returns:
The correct value (rounded to double precision, within 1ulp) is:
Or, rounded to the DBL_DIG digits (15):
In case there's any doubt as to the true value, it agrees with Wolfram Alpha's estimate of sin(2^1023), which is:
Apparently glibc's trig functions are failing to perform argument reduction in any sane way...
This is a duplicate of bug#13658. I'll add your test case to the testsuite.
*** This bug has been marked as a duplicate of bug 13658 ***