diff options
author | Krste Asanovic <krste@eecs.berkeley.edu> | 2017-05-20 05:15:29 -0700 |
---|---|---|
committer | Krste Asanovic <krste@eecs.berkeley.edu> | 2017-05-20 05:15:29 -0700 |
commit | e7a897f26417387bb1f7829e3618786ead6d8175 (patch) | |
tree | d0c87feb5c6adcb77101ceb05aaf499bcf4eff4d /src/d.tex | |
parent | 926f81f7f3b085b0f7a47dbe64928d79a38057c7 (diff) | |
download | riscv-isa-manual-e7a897f26417387bb1f7829e3618786ead6d8175.zip riscv-isa-manual-e7a897f26417387bb1f7829e3618786ead6d8175.tar.gz riscv-isa-manual-e7a897f26417387bb1f7829e3618786ead6d8175.tar.bz2 |
Simplified and made more comprehensive the description of handling narrower n-bit floating-point source operands.
Diffstat (limited to 'src/d.tex')
-rw-r--r-- | src/d.tex | 20 |
1 files changed, 6 insertions, 14 deletions
@@ -54,20 +54,12 @@ valid NaN-boxed value. A narrower $n$-bit transfer out of the floating-point registers will transfer the lower $n$ bits of the register ignoring the upper FLEN$-n$ bits. -Floating-point compute and sign-injection operations calculate results -based on the FLEN-bit values held in the {\tt f} registers. A narrow -$n$-bit operation, where \mbox{$n<$ FLEN}, checks that input operands -are correctly NaN-boxed, i.e., all upper FLEN$-n$ bits are 1. If so, -the $n$ least-significant bits of the input are used as the input -value, otherwise the input value is treated as an $n$-bit canonical -NaN. An $n$-bit floating-point result is written as a NaN-boxed value -to the destination {\tt f} register. - -Conversions from integer to floating-point (e.g., FCVT.S.X), will -NaN-box any results narrower than FLEN. Conversions from narrower -$n$-bit floating-point values to integer (e.g., FCVT.X.S) will check -for legal NaN-boxing and treat the input as the $n$-bit canonical NaN -if not a legal $n$-bit value. +Apart from transfer operations described in the previous paragraph, +all other floating-point operations on narrower $n$-bit operations, +\mbox{$n<$ FLEN}, check if the input operands are correctly NaN-boxed, +i.e., all upper FLEN$-n$ bits are 1. If so, the $n$ least-significant +bits of the input are used as the input value, otherwise the input +value is treated as an $n$-bit canonical NaN. \begin{commentary} Earlier versions of this document did not define the behavior of |