--- netpipe-3.7.1.orig/debian/control +++ netpipe-3.7.1/debian/control @@ -0,0 +1,97 @@ +Source: netpipe +Section: net +Priority: optional +Maintainer: Camm Maguire +Build-Depends: lam4-dev ( >= 7.1.1-5 ), libmpich1.0-dev, pvm-dev,debhelper ( >= 4 ), debhelper ( >= 5.0 ) +Standards-Version: 3.8.3 + +Package: netpipe-tcp +Architecture: any +Depends: ${shlibs:Depends}, ${misc:Depends} +Description: A network performance tool using the TCP protocol + NetPIPE is a protocol independent performance tool that encapsulates + the best of ttcp and netperf and visually represents the network + performance under a variety of conditions. By taking the end-to-end + application view of a network, NetPIPE clearly shows the overhead + associated with different protocol layers. NetPIPE answers such + questions as: how soon will a given data block of size k arrive at its + destination? Which network and protocol will transmit size k blocks + the fastest? What is a given network's effective maximum throughput + and saturation level? Does there exist a block size k for which the + throughput is maximized? How much communication overhead is due to the + network communication protocol layer(s)? How quickly will a small (< 1 + kbyte) control message arrive, and which network and protocol are best + for this purpose? + . + This package uses a raw TCP protocol to measure network performance. + +Package: netpipe-lam +Architecture: any +Depends: lam-runtime, ${shlibs:Depends}, ${misc:Depends} +Description: A network performance tool using LAM MPI + NetPIPE is a protocol independent performance tool that encapsulates + the best of ttcp and netperf and visually represents the network + performance under a variety of conditions. By taking the end-to-end + application view of a network, NetPIPE clearly shows the overhead + associated with different protocol layers. NetPIPE answers such + questions as: how soon will a given data block of size k arrive at its + destination? Which network and protocol will transmit size k blocks + the fastest? What is a given network's effective maximum throughput + and saturation level? Does there exist a block size k for which the + throughput is maximized? How much communication overhead is due to the + network communication protocol layer(s)? How quickly will a small (< 1 + kbyte) control message arrive, and which network and protocol are best + for this purpose? + . + This package measures network performance using the MPI protocol, a + Message Passing Interface frequently used in parallel processing, and + which uses in turn TCP as its underlying transport. The implementation + of the MPI standard used by this package is that provided by the lam set + of packages. + +Package: netpipe-mpich +Architecture: any +Depends: mpich-bin, ${shlibs:Depends}, ${misc:Depends} +Description: A network performance tool using MPICH MPI + NetPIPE is a protocol independent performance tool that encapsulates + the best of ttcp and netperf and visually represents the network + performance under a variety of conditions. By taking the end-to-end + application view of a network, NetPIPE clearly shows the overhead + associated with different protocol layers. NetPIPE answers such + questions as: how soon will a given data block of size k arrive at its + destination? Which network and protocol will transmit size k blocks + the fastest? What is a given network's effective maximum throughput + and saturation level? Does there exist a block size k for which the + throughput is maximized? How much communication overhead is due to the + network communication protocol layer(s)? How quickly will a small (< 1 + kbyte) control message arrive, and which network and protocol are best + for this purpose? + . + This package measures network performance using the MPI protocol, a + Message Passing Interface frequently used in parallel processing, and + which uses in turn TCP as its underlying transport. The implementation + of the MPI standard used by this package is that provided by the mpich + package. + +Package: netpipe-pvm +Architecture: any +Depends: pvm, ${shlibs:Depends}, ${misc:Depends} +Description: A network performance tool using PVM + NetPIPE is a protocol independent performance tool that encapsulates + the best of ttcp and netperf and visually represents the network + performance under a variety of conditions. By taking the end-to-end + application view of a network, NetPIPE clearly shows the overhead + associated with different protocol layers. NetPIPE answers such + questions as: how soon will a given data block of size k arrive at its + destination? Which network and protocol will transmit size k blocks + the fastest? What is a given network's effective maximum throughput + and saturation level? Does there exist a block size k for which the + throughput is maximized? How much communication overhead is due to the + network communication protocol layer(s)? How quickly will a small (< 1 + kbyte) control message arrive, and which network and protocol are best + for this purpose? + . + This package measures network performance using the PVM protocol, a + Parallel Virtual Machine interface frequently used in parallel processing, + and which uses in turn TCP as its underlying transport. PVM support + is provided in its own separate pvm package on Debian systems. --- netpipe-3.7.1.orig/debian/netpipe-lam.manpages +++ netpipe-3.7.1/debian/netpipe-lam.manpages @@ -0,0 +1,2 @@ +NPlam.1 +NPlam2.1 --- netpipe-3.7.1.orig/debian/watch +++ netpipe-3.7.1/debian/watch @@ -0,0 +1,6 @@ +# Example watch control file for uscan +# Rename this file to "watch" and then you can run the "uscan" command +# to check for upstream updates and more. +# Site Directory Pattern Version Script +version=3 +http://www.scl.ameslab.gov/netpipe/code/NetPIPE[_-]([0-9\.]*)\.tar\.gz debian uupdate --- netpipe-3.7.1.orig/debian/netpipe-tcp.files +++ netpipe-3.7.1/debian/netpipe-tcp.files @@ -0,0 +1 @@ +usr/bin/NPtcp --- netpipe-3.7.1.orig/debian/netpipe-mpich.manpages +++ netpipe-3.7.1/debian/netpipe-mpich.manpages @@ -0,0 +1 @@ +NPmpich.1 --- netpipe-3.7.1.orig/debian/netpipe-lam.files +++ netpipe-3.7.1/debian/netpipe-lam.files @@ -0,0 +1,2 @@ +usr/bin/NPlam +usr/bin/NPlam2 --- netpipe-3.7.1.orig/debian/netpipe-mpich.docs +++ netpipe-3.7.1/debian/netpipe-mpich.docs @@ -0,0 +1,2 @@ +dox/README +dox/netpipe_paper.ps --- netpipe-3.7.1.orig/debian/netpipe-pvm.files +++ netpipe-3.7.1/debian/netpipe-pvm.files @@ -0,0 +1 @@ +usr/bin/NPpvm --- netpipe-3.7.1.orig/debian/compat +++ netpipe-3.7.1/debian/compat @@ -0,0 +1 @@ +5 --- netpipe-3.7.1.orig/debian/changelog +++ netpipe-3.7.1/debian/changelog @@ -0,0 +1,143 @@ +netpipe (3.7.1-1) unstable; urgency=low + + * New upstream release + * Bug fix: "debian/watch fails to report upstream's version", thanks + to Raphael Geissert (Closes: #450010). + + -- Camm Maguire Wed, 06 Jan 2010 19:07:32 +0000 + +netpipe (3.6.2-2) unstable; urgency=low + + * Fix dependencies for mpich and lam transitions, Closes: #323744. + * newer standards + + -- Camm Maguire Fri, 7 Oct 2005 14:18:20 +0000 + +netpipe (3.6.2-1) unstable; urgency=high + + * New upstream release + + -- Camm Maguire Tue, 26 Oct 2004 20:28:24 +0000 + +netpipe (3.6-4) unstable; urgency=low + + * + * Bug fix: "netpipe-tcp: invalid option -P", thanks to Kevin Turner + (Closes: #251160). New manpage from upstream elides -P option. + * Bug fix: "netpipe-tcp: option -t requires an argument", thanks to + Kevin Turner (Closes: #251162). New manpage from upstrream elides + obsolete -t option. + * Bug fix: "option -r misdocumented in manpage", thanks to Kevin Turner + (Closes: #251167). New manpage from upstream correctly documents -r + option. + * Bug fix: "netpipe-tcp: option -i misdocumented in manpage", thanks to + Kevin Turner (Closes: #251169). New manpage from upstream correctly + documents -i option. + * Bug fix: "netpipe-tcp: output contains three columns, not five", + thanks to Kevin Turner (Closes: #251619). New manpage from upstream + correctly documents np.out file format. + + -- Camm Maguire Wed, 2 Jun 2004 13:43:28 +0000 + +netpipe (3.6-3) unstable; urgency=low + + * Earlier bug close misapplied, fix here. + + -- Camm Maguire Mon, 5 Apr 2004 20:30:19 +0000 + +netpipe (3.6-2) unstable; urgency=low + + * Bug fix: "netpipe-tcp: Option -P documented but not known", thanks to + Stephane Bortzmeyer (Closes: #231882). Edited manpages and netpipe.c + + -- Camm Maguire Mon, 5 Apr 2004 20:26:06 +0000 + +netpipe (3.6-1) unstable; urgency=low + + * New upstream release + + -- Camm Maguire Fri, 23 Jan 2004 03:33:34 +0000 + +netpipe (3.5-3) unstable; urgency=low + + * Build-depend on lam4-dev + * Newer standards + * debhelper compat level 4 + + -- Camm Maguire Wed, 21 Jan 2004 21:02:27 +0000 + +netpipe (3.5-2) unstable; urgency=low + + * Applied 64bit safe patches, Closes: #214501 + + -- Camm Maguire Tue, 14 Oct 2003 17:19:19 +0000 + +netpipe (3.5-1) unstable; urgency=low + + * New upstream release, Closes: #205285 + * Added netpipe_paper.ps to docs + * Carried forward manpage from 2.4 + * Applied bounds checking patch to -o argument, Closes: #203488 + * Added NPlam2 executable + + -- Camm Maguire Wed, 20 Aug 2003 19:12:28 +0000 + +netpipe (2.4-5) unstable; urgency=low + + * Added watchfile + * Added -lutil to EXTRA_LIBS + * remove dh_suidregister + * Newer standards + * Cleaned copyright file + * Recently required link against -lpmich added + * Shared lib linking for NPmpich + + -- Camm Maguire Wed, 20 Aug 2003 17:49:06 +0000 + +netpipe (2.4-4) unstable; urgency=low + + * Varied long descriptions, Closes: #135533 + * Newer standards + + -- Camm Maguire Sun, 24 Feb 2002 19:31:29 -0500 + +netpipe (2.4-3) unstable; urgency=low + + * Fixed spelling of NetPIPE in control, Closes: #132928 + + -- Camm Maguire Mon, 11 Feb 2002 10:47:14 -0500 + +netpipe (2.4-2) unstable; urgency=low + + * Rebuild against new lam3-dev, Closes: #106949 + + -- Camm Maguire Mon, 30 Jul 2001 13:34:26 -0400 + +netpipe (2.4-1) unstable; urgency=low + + * New upstream release + * Added debhelper to Build-Depends + + -- Camm Maguire Fri, 23 Mar 2001 18:17:11 -0500 + +netpipe (2.3-3) unstable; urgency=low + + * Build depends on lam2-dev, pvm-dev and mpich + * Search for mpich includes and libs, closes: #72842 + + -- Camm Maguire Thu, 25 Jan 2001 18:16:56 -0500 + +netpipe (2.3-2) unstable; urgency=low + + * Upgraded to latest standards version + + -- Camm Maguire Sun, 2 Apr 2000 16:11:56 +0200 + +netpipe (2.3-1) unstable; urgency=low + + * Altered Makefile to allow for lam and mpich builds using runtime variables + * Initial Release. + + -- Camm Maguire Sat, 1 Jan 2000 23:01:51 -0500 + + --- netpipe-3.7.1.orig/debian/copyright +++ netpipe-3.7.1/debian/copyright @@ -0,0 +1,12 @@ +This package was debianized by Camm Maguire on +Sat, 1 Jan 2000 23:01:51 -0500. + +It was downloaded from ftp://ftp.scl.ameslab.gov/pub/netpipe/ + +Upstream Authors: Guy Helmer et.al., ghelmer@scl.ameslab.gov + Dave Turner, turner@ameslab.gov + +Copyright: + + netpipe is covered under the terms of the GPL. See the file + /usr/share/common-licenses/GPL for more information. --- netpipe-3.7.1.orig/debian/rules +++ netpipe-3.7.1/debian/rules @@ -0,0 +1,120 @@ +#!/usr/bin/make -f +#-*- makefile -*- +# Made with the aid of dh_make, by Craig Small +# Sample debian/rules that uses debhelper. GNU copyright 1997 by Joey Hess. +# Also some stuff taken from debmake scripts, by Christoph Lameter. + +# Uncomment this to turn on verbose mode. +#export DH_VERBOSE=1 + +# This is the debhelper compatability version to use. +#export DH_COMPAT=1 + +build: build-stamp +build-stamp: + dh_testdir + + $(MAKE) tcp + cp debian/netpipe.1 NPtcp.1 + $(MAKE) pvm + cp debian/netpipe.1 NPpvm.1 + $(MAKE) mpi MPICC=mpicc.lam + mv NPmpi NPlam + cp debian/netpipe.1 NPlam.1 + C_INCLUDE_PATH=/usr/lib/mpich/include $(MAKE) mpi MPICC="$$(mpicc.mpich -show | sed 's,mpich/lib ,mpich/lib/shared ,1')" + mv NPmpi NPmpich + cp debian/netpipe.1 NPmpich.1 + +# MPI2_INC not necessary, but must not be empty + $(MAKE) mpi2 MPI2CC=mpicc.lam MPI2_INC=./ + mv NPmpi2 NPlam2 + cp debian/netpipe.1 NPlam2.1 + + touch build-stamp + +clean: + dh_testdir + dh_testroot + rm -f build-stamp install-stamp + + $(MAKE) clean + rm -f NPtcp.1 NPlam.1 NPmpich.1 NPlam2.1 NPpvm.1 + rm -f NPtcp NPlam NPmpich NPlam2 NPpvm + + dh_clean + +install: install-stamp +install-stamp: build-stamp + dh_testdir + dh_testroot + dh_clean -k + dh_installdirs + + mkdir -p debian/tmp/usr/bin + cp NPtcp NPlam NPmpich NPlam2 NPpvm `pwd`/debian/tmp/usr/bin + + dh_movefiles + touch install-stamp + +# Build architecture-independent files here. +binary-indep: build install +# dh_testversion + dh_testdir -i + dh_testroot -i + dh_installdocs -i + dh_installexamples -i + dh_installmenu -i +# dh_installemacsen -i +# dh_installpam -i +# dh_installinit -i + dh_installcron -i +# dh_installmanpages -i + dh_installinfo -i +# dh_undocumented + dh_installchangelogs -i + dh_link -i + dh_compress -i + dh_fixperms -i + dh_installdeb -i +# dh_perl -i + dh_gencontrol -i + dh_md5sums -i + dh_builddeb -i + +# Build architecture-dependent files here. +binary-arch: build install +# dh_testversion + dh_testdir -a + dh_testroot -a + dh_installdocs -a + dh_installexamples -a + dh_installmenu -a +# dh_installemacsen -a +# dh_installpam -a +# dh_installinit -a + dh_installcron -a + dh_installman -a +# dh_installmanpages -pnetpipe-lam NPtcp.1 NPmpich.1 NPpvm.1 netpipe.1 +# dh_installmanpages -pnetpipe-pvm NPtcp.1 NPmpich.1 NPlam.1 NPlam2.1 netpipe.1 +# dh_installmanpages -pnetpipe-tcp NPmpich.1 NPlam.1 NPlam2.1 NPpvm.1 netpipe.1 +# dh_installmanpages -pnetpipe-mpich NPtcp.1 NPlam.1 NPlam2.1 NPpvm.1 netpipe.1 + dh_installinfo -a +# dh_undocumented + dh_installchangelogs -a + dh_strip -a + dh_link -a + dh_compress -a + dh_fixperms -a + dh_installdeb -a +# dh_makeshlibs -a +# dh_perl -a + dh_shlibdeps -a + dh_gencontrol -a + dh_md5sums -a + dh_builddeb -a + +source diff: + @echo >&2 'source and diff are obsolete - use dpkg-source -b'; false + +binary: binary-arch +.PHONY: build clean binary-indep binary-arch binary install --- netpipe-3.7.1.orig/debian/netpipe-tcp.docs +++ netpipe-3.7.1/debian/netpipe-tcp.docs @@ -0,0 +1,2 @@ +dox/README +dox/netpipe_paper.ps --- netpipe-3.7.1.orig/debian/dirs +++ netpipe-3.7.1/debian/dirs @@ -0,0 +1 @@ +usr/bin --- netpipe-3.7.1.orig/debian/netpipe-mpich.files +++ netpipe-3.7.1/debian/netpipe-mpich.files @@ -0,0 +1 @@ +usr/bin/NPmpich --- netpipe-3.7.1.orig/debian/manpage_old +++ netpipe-3.7.1/debian/manpage_old @@ -0,0 +1,396 @@ +.\" -*- nroff -*- +.\" +.\" NetPIPE -- Network Protocol Independent Performance Evaluator. +.\" Copyright 1997, 1998 Iowa State University Research Foundation, Inc. +.\" +.\" This program is free software; you can redistribute it and/or modify +.\" it under the terms of the GNU General Public License as published by +.\" the Free Software Foundation. You should have received a copy of the +.\" GNU General Public License along with this program; if not, write to the +.\" Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. +.\" +.\" netpipe.1 +.\" Created: Mon Jun 15 1998 by Guy Helmer +.\" +.\" $Id: netpipe.1,v 1.3 1998/09/24 16:23:59 ghelmer Exp $ +.\" +.TH netpipe 1 "June 15, 1998" "NetPIPE" "netpipe" + +.SH NAME +NetPIPE \- network protocol independent performance evaluator + +.SH SYNOPSIS +.B NPtcp +[\c +.BI \-A \ buffer_alignment\fR\c +] +[\c +.BR \-a \c +] +[\c +.BI \-b \ TCP_buffer_size\fR\c +] +[\c +.BI \-h \ host_name\fR\c +] +[\c +.BI \-i \ increment\fR\c +] +[\c +.BI \-l \ starting_msg_size\fR\c +] +[\c +.BI \-O \ buffer_offset\fR\c +] +[\c +.BI \-o \ output_filename\fR\c +] +[\c +.BR \-P \c +] +[\c +.BI \-p \ port\fR\c +] +[\c +.BR \-r \c +] +[\c +.BR \-s \c +] +[\c +.BR \-t \c +] +[\c +.BI \-u \ ending_msg_size\fR\c +] + +.PP + +.B NPmpi +[\c +.BI \-A \ buffer_alignment\fR\c +] +[\c +.BR \-a \c +] +[\c +.BI \-i \ increment\fR\c +] +[\c +.BI \-l \ starting_msg_size\fR\c +] +[\c +.BI \-O \ buffer_offset\fR\c +] +[\c +.BI \-o \ output_filename\fR\c +] +[\c +.BR \-P \c +] +[\c +.BR \-s \c +] +[\c +.BI \-u \ ending_msg_size\fR\c +] + +.PP + +.B NPpvm +[\c +.BI \-A \ buffer_alignment\fR\c +] +[\c +.BR \-a \c +] +[\c +.BI \-i \ increment\fR\c +] +[\c +.BI \-l \ starting_msg_size\fR\c +] +[\c +.BI \-O \ buffer_offset\fR\c +] +[\c +.BI \-o \ output_filename\fR\c +] +[\c +.BR \-P \c +] +[\c +.BR \-r \c +] +[\c +.BR \-s \c +] +[\c +.BR \-t \c +] +[\c +.BI \-u \ ending_msg_size\fR\c +] + +.SH DESCRIPTION +.PP +.B NetPIPE +is a protocol independent performance tool that encapsulates +the best of ttcp and netperf and visually represents the network +performance under a variety of conditions. By taking the end-to-end +application view of a network, +.B NetPIPE +clearly shows the overhead +associated with different protocol layers. +.B NetPIPE +answers such questions as: +.RS +How soon will a given data block of size k arrive at its destination? +.PP +Which network and protocol will transmit size k blocks the fastest? +.PP +What is a given network's effective maximum throughput and saturation +level? +.PP +Does there exist a block size k for which the throughput is maximized? +.PP +How much communication overhead is due to the network communication +protocol layer(s)? +.PP +How quickly will a small (< 1 kbyte) control message arrive, and which +network and protocol are best for this purpose? +.RE +.PP +.B NetPIPE +is provided with interfaces for TCP, MPI, and PVM, but TCP is the most +commonly used interface for general network testing purposes. It +should be easy to write new interfaces for other reliable protocols based +on the examples provided by the TCP, MPI and PVM interfaces. +.SH TESTING TCP +.PP +Typical use for TCP involves running the TCP NetPIPE receiver on one +system with the command +.PP +.Ex +NPtcp \-r +.Ee +.PP +and running the TCP NetPIPE transmitter on another system with the +command +.PP +.Ex +NPtcp \-h receiver_hostname \-o output_filename \-P \-t +.Ee +.PP +If any options are used that modify the test protocol, including \-i, +\-l, \-p, \-s, and \-u, those parameters +.B must +be used on both the transmitter and the receiver, or the test +will not run properly. +.SH TESTING PVM +.PP +Typical use for PVM first requires starting PVM with the command +.PP +.Ex +pvm +.Ee +.PP +and adding a second machine with the PVM command +.PP +.Ex +add othermachine +.Ee +.PP +(then exit the PVM command line interface). Then run the PVM NetPIPE +receiver on one system with the command +.PP +.Ex +NPpvm \-r +.Ee +.PP +and run the TCP NetPIPE transmitter on the other system with the +command +.PP +.Ex +NPpvm \-t \-o output_filename \-P +.Ee +.PP +If any options are used that modify the test protocol, including \-i, +\-l, \-p, \-s, and \-u, those parameters +.B must +be used on both the transmitter and the receiver, or the test +will not run properly. +.SH TESTING MPI +.PP +Use of the MPI interface for NetPIPE depends on the MPI implementation +used. For the Argonne MPICH implementation using the p4 device (for a +cluster of individual systems interconnected using TCP/IP), create a +file that contains the hostnames of the two systems you want to +include in the test, with one hostname on each line of the file +(assume the file is named "machines.p4"). Then, use the command +.PP +.Ex +mpirun \-machinefile machines.p4 \-np 2 NPmpi \-o output_filename \-P +.Ee +.PP +to start the test. MPICH will start an NPmpi process on each of the +two selected machines and the test will begin. +.SH TESTING METHODOLOGY +.PP +.B NetPIPE +tests network performance by sending a number of messages at each +block size, starting from the lower bound on message size. +.B NetPIPE +increments the message size until the upper bound on message size is +reached or the time to transmit a block exceeds one second, which ever +occurs first. +.PP +.B NetPIPE\c +\'s output file may be graphed with a program such as +.B gnuplot(1) +to view the results of the test. +.B NetPIPE\c +\'s +output file contains five columns: time to transfer the block, bits +per second, bits in block, bytes in block, and variance. These +columns may be graphed to represent and compare the network's +performance. For example, the +.B network signature graph +can be created by graphing time versus bits per second. Sample +.B gnuplot(1) +commands for such a graph would be +.PP +.Ex +set logscale x +.Ee +.PP +.Ex +plot "NetPIPE.out" using 1:2 +.Ee +.PP +The more traditional +.B throughput versus block size +graph can be created by graphing bytes versus bits per second. +Sample +.B gnuplot(1) +commands for such a graph would be +.PP +.Ex +set logscale x +.Ee +.PP +.Ex +plot "NetPIPE.out" using 4:2 +.Ee + +.ne 5 +.SH OPTIONS +.TP +.B \-A \ \fIalignment\fR +Align buffers to the given boundary. For example, a value of 4 would +align buffers to 4-byte (word) boundaries. +.ne 3 +.TP +.B \-a +Specify asynchronous receive (a.k.a. preposted receive), if the +underlying protocol supports it. +.ne 3 +.TP +.BI \-b \ \fIbuffer_size\fR +[TCP only] Set send and receive TCP buffer sizes. +.ne 3 +.TP +.BI \-h \ \fIhostname\fR +[TCP transmitter only] Specify name of host to which to connect. +.ne 3 +.TP +.BI \-i \ \fIincrement\fR +Specify increment step size (default is an exponentially increasing +increment). +.ne 3 +.TP +.BI \-l \ \fIstart_msg_size\fR +Specify the starting message size. The test will start with messages +of this size and increment, either exponentially or with an increment +specified by the +.B \-i +flag, until a block requires more than one second to transmit or the +ending message size specified by the +.B \-u +flag is reached, which ever occurs first. +.ne 3 +.TP +.BI \-O \ \fIbuffer_offset\fR +Specify offset of buffers from alignment. For example, specifying an +alignment of 4 (with \-A) and an offset of 1 would align buffers to +the first byte after a word boundary. +.ne 3 +.TP +.BI \-o \ \fIoutput_filename\fR +Specify output filename. By default, the output filename is +.IR NetPIPE.out . +.ne 3 +.\".TP +.\".B \-P +.\"Print results on screen during execution of the test. By default, +.\"NetPIPE is silent during execution of the test. +.\".ne 3 +.TP +.BI \-p \ \fIport_number\fR +[TCP only] Specify TCP port number to which to connect (for the +transmitter) or the port on which to listen for connections (for the +receiver). +.ne 3 +.TP +.B \-r +[TCP only] This process is a TCP receiver. +.ne 3 +.TP +.B \-s +Set streaming mode: data is only transmitted in one direction. By +default, the transmitter measures the time taken as each data block is +sent from the transmitter to the receiver and back, then divides the +round-trip time by two to obtain the time taken by the message to +travel in each direction. In streaming mode, the receiver measures +the time required to receive the message and sends the measured time +back to the transmitter for posting to the output file. +.ne 3 +.TP +.B \-t +[TCP only] This process is a TCP transmitter. +.ne 3 +.TP +.BI \-u \ \fIending_msg_size\fR +Specify the ending message size. By default, the test will end when +the time to transmit a block exceeds one second. If +.B \-u +is specified, the test will end when either the test time exceeds one +second or the ending message size is reached, which ever occurs first. + +.ne 3 +.SH FILES +.TP +.I NetPIPE.out +Default output file for +.BR NetPIPE . +Overridden by the +.B \-o +option. + +.SH AUTHOR +.PP +Quinn Snell , Guy Helmer , +and others. +.PP +Clark Dorman contributed the PVM interface. +.PP +Information about +.B NetPIPE +can be found on the World Wide Web at +http://www.scl.ameslab.gov/netpipe/. + +.SH BUGS +By nature, +.B NetPIPE +will use as much of the network bandwidth as possible. Other users of +the network may notice the effect. --- netpipe-3.7.1.orig/debian/netpipe-tcp.manpages +++ netpipe-3.7.1/debian/netpipe-tcp.manpages @@ -0,0 +1 @@ +NPtcp.1 --- netpipe-3.7.1.orig/debian/netpipe-pvm.docs +++ netpipe-3.7.1/debian/netpipe-pvm.docs @@ -0,0 +1,2 @@ +dox/README +dox/netpipe_paper.ps --- netpipe-3.7.1.orig/debian/docs +++ netpipe-3.7.1/debian/docs @@ -0,0 +1,2 @@ +dox/README +dox/netpipe_paper.ps --- netpipe-3.7.1.orig/debian/netpipe-pvm.manpages +++ netpipe-3.7.1/debian/netpipe-pvm.manpages @@ -0,0 +1 @@ +NPpvm.1 --- netpipe-3.7.1.orig/debian/netpipe-lam.docs +++ netpipe-3.7.1/debian/netpipe-lam.docs @@ -0,0 +1,2 @@ +dox/README +dox/netpipe_paper.ps --- netpipe-3.7.1.orig/debian/netpipe.1 +++ netpipe-3.7.1/debian/netpipe.1 @@ -0,0 +1,418 @@ +.\" -*- nroff -*- +.\" +.\" NetPIPE -- Network Protocol Independent Performance Evaluator. +.\" Copyright 1997, 1998 Iowa State University Research Foundation, Inc. +.\" +.\" This program is free software; you can redistribute it and/or modify +.\" it under the terms of the GNU General Public License as published by +.\" the Free Software Foundation. You should have received a copy of the +.\" GNU General Public License along with this program; if not, write to the +.\" Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. +.\" +.\" netpipe.1 +.\" Created: Mon Jun 15 1998 by Guy Helmer +.\" Rewritten: Jun 1 2004 by Dave Turner +.\" +.\" $Id: netpipe.1,v 1.3 1998/09/24 16:23:59 ghelmer Exp $ +.\" +.TH netpipe 1 "June 1, 2004" "NetPIPE" "netpipe" + +.SH NAME +NetPIPE \- +.IB Net work +.IB P rotocol +.IB I ndependent +.IB P erformance +.IB E valuator + +.SH SYNOPSIS +.B NPtcp +[\c +.BI \-h \ receiver_hostname\fR\c +] +[\c +.BI \-b \ TCP_buffer_sizes\fR\c +] +[options] + +.PP + +mpirun +[\c +.BI \-machinefile \ hostlist\fR\c +] +-np 2 +.B NPmpi +[-a] [-S] [-z] [options] + +.PP + +mpirun +[\c +.BI \-machinefile \ hostlist\fR\c +] +-np 2 +.B NPmpi2 +[-f] [-g] [options] + + +.PP + +.B NPpvm +[options] + +See the TESTING sections below for a more complete description of +how to run NetPIPE in each environment. +The OPTIONS section describes the general options available for +all modules. +See the README file from the tar-ball at +http://www.scl.ameslab.gov/Projects/NetPIPE/ for documentation on +the InfiniBand, GM, SHMEM, LAPI, and memcpy modules. + +.SH DESCRIPTION +.PP + +.B NetPIPE +uses a simple series of ping-pong tests over a range of message +sizes to provide a complete measure of the performance of a network. +It bounces messages of increasing size between two processes, whether across a +network or within an SMP system. +Message sizes are chosen at regular intervals, and with slight perturbations, +to provide a complete evaluation of the communication system. +Each data point involves many ping-pong tests to provide an accurate timing. +Latencies are calculated by dividing the round trip time in half for small +messages ( less than 64 Bytes ). +.PP +The communication time for small messages is dominated by the +overhead in the communication layers, meaning that the transmission +is latency bound. +For larger messages, the communication rate becomes bandwidth limited by +some component in +the communication subsystem (PCI bus, network card link, network switch). +.PP +These measurements can be done at the message-passing layer +(MPI, MPI-2, and PVM) or at the native communications layers +that that run upon (TCP/IP, GM for Myrinet cards, InfiniBand, +SHMEM for the Cray T3E systems, and LAPI for IBM SP systems). +Recent work is being aimed at measuring some internal system properties +such as the memcpy module that measures the internal memory copy rates, +or a disk module under development that measures the performance +to various I/O devices. +.PP + +Some uses for NetPIPE include: +.RS +.PP +Comparing the latency and maximum throughput of various network cards. +.PP +Comparing the performance between different types of networks. +.PP +Looking for inefficiencies in the message-passing layer by comparing it +to the native communication layer. +.PP +Optimizing the message-passing layer and tune OS and driver parameters +for optimal performance of the communication subsystem. + +.RE +.PP + +.B NetPIPE +is provided with many modules allowing it to interface with a wide +variety of communication layers. +It is fairly easy to write new interfaces for other reliable protocols +by using the existing modules as examples. + + + +.SH TESTING TCP +.PP +NPtcp can now be launched in two ways, by manually starting NPtcp on +both systems or by using a nplaunch script. To manually start NPtcp, +the NetPIPE receiver must be +started first on the remote system using the command: +.PP +NPtcp [options] +.PP +then the primary transmitter is started on the local system with the +command +.PP +NPtcp \-h +.I receiver_hostname +[options] +.PP +Any options used must be the same on both sides. + +The nplaunch script uses ssh to launch the remote receiver +before starting the local transmitter. To use rsh, simply change +the nplaunch script. +.PP +nplaunch NPtcp -h +.I receiver_hostname +[options] +.PP +The +.BI \-b \ TCP_buffer_sizes\fR\c + option sets the TCP socket buffer size, which can greatly influence +the maximum throughput on some systems. A throughput graph that +flattens out suddenly may be a sign of the performance being limited +by the socket buffer sizes. + + +.SH TESTING MPI and MPI-2 +.PP +Use of the MPI interface for NetPIPE depends on the MPI implementation +being used. +All will require the number of processes to be specified, usually +with a +.I -np 2 +argument. Clusters environments may require a list of the +hosts being used, either during initialization of MPI (during lamboot +for LAM-MPI) or when each job is run (using a -machinefile argument +for MPICH). +For LAM-MPI, for example, put the list of hosts in hostlist then boot LAM +and run NetPIPE using: +.PP +lamboot -v -b +.I hostlist +.PP +mpirun \-np 2 NPmpi [NetPIPE options] +.PP + +For MPICH use a command like: +.PP +mpirun \-machinefile +.I hostlist +\-np 2 NPmpi [NetPIPE options] +.PP + +To test the 1-sided communications of the MPI-2 standard, compile +using: +.PP +.B make mpi2 +.PP +Running as described above and MPI will use 1-sided MPI_Put() +calls in both directions, with each receiver blocking until the +last byte has been overwritten before bouncing the message back. +Use the +.I -f +option to force usage of a fence to block rather than an overwrite +of the last byte. +The +.I -g +option will use MP_Get() functions to transfer the data rather than +MP_Put(). + + +.SH TESTING PVM +.PP +Start the pvm system using: +.PP +pvm +.PP +and adding a second machine with the PVM command +.PP +add +.I receiver_hostname +.PP +Exit the PVM command line interface using quit, then run the PVM NetPIPE +receiver on one system with the command: +.PP +NPpvm [options] +.PP +and run the TCP NetPIPE transmitter on the other system with the +command: +.PP +NPpvm -h +.I receiver hostname +[options] +.PP +Any options used must be the same on both sides. +The nplaunch script may also be used with NPpvm as described above +for NPtcp. + +.SH TESTING METHODOLOGY +.PP +.B NetPIPE +tests network performance by sending a number of messages at each +block size, starting from the lower bound on the message sizes. + +The message size is incremented until the upper bound on the message size is +reached or the time to transmit a block exceeds one second, which ever +occurs first. Message sizes are chosen at regular intervals, and for +slight perturbations from them to provide a more complete evaluation +of the communication subsystem. +.PP +The +.B NetPIPE\c + output file may be graphed using a program such as +.B gnuplot(1)\. +The output file contains three columns: the number of bytes in the block, +the transfer rate in bits per second, and +the time to transfer the block (half the round-trip time). +The first two columns are normally used to graph the throughput +vs block size, while the third column provides the latency. +For example, the +.B throughput versus block size +graph can be created by graphing bytes versus bits per second. +Sample +.B gnuplot(1) +commands for such a graph would be +.PP +set logscale x +.PP +plot "np.out" + + +.ne 5 +.SH OPTIONS +.TP +.B \-a +asynchronous mode: prepost receives (MPI, IB modules) +.ne 3 +.TP +.BI \-b \ \fITCP_buffer_sizes\fR +Set the send and receive TCP buffer sizes (TCP module only). +.ne 3 + +.TP +.B \-B +Burst mode where all receives are preposted at once (MPI, IB modules). +.ne 3 + +.TP +.B \-f +Use a fence to block for completion (MPI2 module only). +.ne 3 + +.TP +.B \-g +Use MPI_Get() instead of MPI_Put() (MPI2 module only). +.ne 3 + +.TP +.BI \-h \ \fIhostname\fR +Specify the name of the receiver host to connect to (TCP, PVM, IB, GM). +.ne 3 + +.TP +.B \-I +Invalidate cache to measure performance without cache effects (mostly affects +IB and memcpy modules). +.ne 3 + +.TP +.B \-i +Do an integrity check instead of a performance evaluation. +.ne 3 + +.TP +.BI \-l \ \fIstarting_msg_size\fR +Specify the lower bound for the size of messages to be tested. +.ne 3 +.TP + +.TP +.BI \-n \ \fInrepeats\fR +Set the number of repeats for each test to a constant. +Otherwise, the number of repeats is chosen to provide an accurate +timing for each test. Be very careful if specifying a low number +so that the time for the ping-pong test exceeds the timer accuracy. +.ne 3 + +.TP +.BI \-O \ \fIsource_offset,dest_offset\fR +Specify the source and destination offsets of the buffers from perfect +page alignment. +.ne 3 +.TP + +.BI \-o \ \fIoutput_filename\fR +Specify the output filename (default is np.out). +.ne 3 + +.TP +.BI \-p \ \fIperturbation_size\fR +NetPIPE chooses the message sizes at regular intervals, increasing them +exponentially from the lower boundary to the upper boundary. +At each point, it also tests perturbations of 3 bytes above and 3 bytes +below each test point to find idiosyncrasies in the system. +This perturbation value can be changed using the +.I -p +option, or turned +off using +.I -p +.I 0 +.B . +.ne 3 + +.TP +.B \-r +This option resets the TCP sockets after every test (TCP module only). +It is necessary for some streaming tests to get good measurements +since the socket window size may otherwise collapse. +.ne 3 + +.TP +.B \-s +Set streaming mode where data is only transmitted in one direction. +.ne 3 + +.TP +.B \-S +Use synchronous sends (MPI module only). +.ne 3 + +.TP +.BI \-u \ \fIupper_bound\fR +Specify the upper boundary to the size of message being tested. +By default, NetPIPE will stop when +the time to transmit a block exceeds one second. + +.TP +.B \-z +Receive messages using MPI_ANY_SOURCE (MPI module only) +.ne 3 + +.TP +.B \-2 +Set bi-directional mode where both sides send and receive at the +same time (supported by most modules). +You may need to use +.I -a +to choose asynchronous communications for MPI to avoid freeze-ups. +For TCP, the maximum test size will be limited by the TCP +buffer sizes. +.ne 3 + +.ne 3 +.SH FILES +.TP +.I np.out +Default output file for +.BR NetPIPE . +Overridden by the +.B \-o +option. + +.SH AUTHOR +.PP +The original NetPIPE core plus TCP and MPI modules were written by +Quinn Snell, Armin Mikler, Guy Helmer, and John Gustafson. +NetPIPE is currently being developed and maintained by Dave Turner +with contributions from many students (Bogdan Vasiliu, Adam Oline, +Xuehua Chen, and Brian Smith). + +.PP +Send comments/bug-reports to: +.I +. +.PP +Additional information about +.B NetPIPE +can be found on the World Wide Web at +.I http://www.scl.ameslab.gov/Projects/NetPIPE/ + +.SH BUGS +As of version 3.6.1, there is a bug that causes NetPIPE to segfault on +RedHat Enterprise systems. I will debug this as soon as I get access to a +few such systems. -Dave Turner (turner@ameslab.gov)