diff options
author | Eric S. Raymond <esr@thyrsus.com> | 2005-02-22 03:55:09 +0000 |
---|---|---|
committer | Eric S. Raymond <esr@thyrsus.com> | 2005-02-22 03:55:09 +0000 |
commit | 73fa34c4c341fcac5a6645f64dd1d0252e258ce8 (patch) | |
tree | 5bf21f414ed1e1a4f14bcf489634d3a78a5cb4c9 | |
parent | d0c47fc5e51cbbe976305b9b838a91ca1d25f7b5 (diff) | |
download | gpsd-73fa34c4c341fcac5a6645f64dd1d0252e258ce8.tar.gz |
Documentation fixes.
-rw-r--r-- | AUTHORS | 2 | ||||
-rw-r--r-- | README | 2 | ||||
-rw-r--r-- | www/performance/performance.xml | 14 |
3 files changed, 10 insertions, 8 deletions
@@ -2,4 +2,4 @@ Remco Treffkorn <remco@rvt.com> Derrick J. Brashear <shadow@dementia.org> Russ Nelson <nelson@crynwyr.com> Eric S. Raymond <esr@thyrsus.com> -Arnim Laeuger <arnim.laeuger@gmx.net> + @@ -49,7 +49,7 @@ Brook Milligan <brook@trillium.NMSU.Edu> combined gpsd and gpsclient into one package and autoconfiscated it. Derrick J. Brashear <shadow@dementia.org> (KB3EGH) added code for the -EarthMate DeLame. He also added "incredibly gross code to output +EarthMate DeLorme. He also added "incredibly gross code to output NMEA sentences" (his own words :-) He also did the first cut at DGPS support (see http://www.wsrcc.com/wolfgang/gps/dgps-ip.html), for the Earthmate. diff --git a/www/performance/performance.xml b/www/performance/performance.xml index 75e4d153..9d7d6af1 100644 --- a/www/performance/performance.xml +++ b/www/performance/performance.xml @@ -247,8 +247,9 @@ moderately fast machine in early 2005, but average processor utilization remained low throughout the tests.</para> <para>The version of the GPSD software I used for the test was -released as 2.13. It was configured with --enable-profiling. All -graphs and figures were generated with +released as 2.13. It was configured with +——enable-profiling. All graphs and figures were generated +with <citerefentry><refentrytitle>gpsprof</refentrytitle><manvolnum>1</manvolnum></citerefentry>, a tool built for this purpose and included in the distribution.</para> @@ -299,10 +300,11 @@ that the profiling reports are not introducing any measurable latency. But what is more interesting is to notice that D1 + W + E2 + T2 + D2 vanishes — at this timescale, all we can see is E1 and T1.</para> -<para>The <ulink url='profile2.txt'>raw data</ulink> bears this out. All -times besides E1 and T1 are so small that they are comparable to the -noise level of the measurements. This may be a bit surprising unless -one remembers that W=0 is expected in this setup, and that a modern +<para>The <ulink url='profile2.txt'>raw data</ulink> bears this out. +All times besides E1 and T1 are so small that they are comparable to +the noise level of the measurements. This may be a bit surprising +unless one knows that a W near 0 is expected in this setup; +<application>gpsprof</application> sets watcher mode. Also, a modern zero-copy TCP/IP stack like Linux's implements local sockets with very low overhead. It is also a little surprising that E1 is so large relative to E1+T1. Recall, however, that this may be measurement |