SCCS-vsn: old/pcc/cc/cc.1 6.2
.\" All rights reserved. The Berkeley software License Agreement
.\" specifies the terms and conditions for redistribution.
.\"
.\" All rights reserved. The Berkeley software License Agreement
.\" specifies the terms and conditions for redistribution.
.\"
-.\" @(#)cc.1 6.1 (Berkeley) %G%
+.\" @(#)cc.1 6.2 (Berkeley) %G%
only floats to double. This does not conform to the standard which states
that all intermediate results should be converted to double but does
provide a speed improvement for programs which don't require full double
only floats to double. This does not conform to the standard which states
that all intermediate results should be converted to double but does
provide a speed improvement for programs which don't require full double
+precision. This option also makes
+.B "register float"
+variables work appropriately.
.TP
.SM
.BI \-B \*Sstring
.TP
.SM
.BI \-B \*Sstring
.SH BUGS
The compiler currently ignores advice to put
\fBchar\fR, \fBunsigned char\fR,
.SH BUGS
The compiler currently ignores advice to put
\fBchar\fR, \fBunsigned char\fR,
-\fBshort\fR or \fBunsigned short\fR variables in registers. It previously
+\fBshort\fR, \fBunsigned short\fR,
+\fBfloat\fR, or \fBdouble\fR
+variables in registers, except as noted above. It previously
produced poor, and in some cases incorrect, code for such declarations.
produced poor, and in some cases incorrect, code for such declarations.