diff options
author | gfavor <47677210+gfavor@users.noreply.github.com> | 2020-10-13 20:41:04 -0700 |
---|---|---|
committer | Andrew Waterman <andrew@sifive.com> | 2020-10-17 14:43:15 -0700 |
commit | 2df103e05896bd7e7377cf94e434ff2a8e92f816 (patch) | |
tree | a226ae51858878a28c709f738f502bbf065d082b | |
parent | 773a6c4cc9db7585d42ec732d5db24f930d1157a (diff) | |
download | riscv-isa-manual-2df103e05896bd7e7377cf94e434ff2a8e92f816.zip riscv-isa-manual-2df103e05896bd7e7377cf94e434ff2a8e92f816.tar.gz riscv-isa-manual-2df103e05896bd7e7377cf94e434ff2a8e92f816.tar.bz2 |
Clarify expected ballpark for Pause duration (#604)
-rw-r--r-- | src/zihintpause.tex | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/src/zihintpause.tex b/src/zihintpause.tex index c739e61..fc77b98 100644 --- a/src/zihintpause.tex +++ b/src/zihintpause.tex @@ -22,6 +22,7 @@ polling for. The duration of a PAUSE instruction's effect may vary significantly within and among implementations. +In typical implementations this duration should be much less than the time to perform a context switch, probably more on the rough order of an on-chip cache miss latency or a cacheless access to main memory. Portable software should not use more than one PAUSE instruction before re-evaluating loop conditions, else performance might substantially degrade on other implementations, or even on some executions on the same implementation. |