Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 5 Dec 1996 23:14:58 +0100
From:      se@FreeBSD.ORG (Stefan Esser)
To:        mark@quickweb.com (Mark Mayo)
Cc:        msmith@atrad.adelaide.edu.au (Michael Smith), hackers@FreeBSD.ORG
Subject:   Re: New benchmarks to design / lmbench results
Message-ID:  <Mutt.19961205231458.se@x14.mi.uni-koeln.de>
In-Reply-To: <Pine.BSF.3.94.961205133119.7282A-100000@vinyl.quickweb.com>; from Mark Mayo on Dec 5, 1996 13:48:41 -0500
References:  <199612050218.MAA19340@genesis.atrad.adelaide.edu.au> <Pine.BSF.3.94.961205133119.7282A-100000@vinyl.quickweb.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Dec 5, mark@quickweb.com (Mark Mayo) wrote:
> Also, I ran lmbench out of curiosity! Wow, does FreeBSD rock on Pipe
> bandwidth!! I get 147.20 MB/s pipe bandwidth. Also, the context switches
> were blazing (166667 - whatever that number is, or 6 usecs) on 2 and 8
> "small processes". Pipe transactions were at 33 usec. The fork/exec
> times were also very impressive, at 1203 usec (831/s). Basically,
> everything involving TCP/UDP/socket/Pipe had amazing results - the highest
> of the EXAMPLE group by factors. I'll have to do searching and see what
> newer machines are doing. The results are quite impressive, and fun to
> look at :-) My machines seems to have very high memory read times
> (according to lmbench anyways) with the memory sum and read bandwidth
> hovering around 160 MB/s and mmap reads at around 108 MB/s. The writes
> weere quite slow (~50MB/s) and bcopy results slow as well.

Bcopy does depend on tghe size of available RAM. Your system
will page heavily, if not all of the test arrays can be held
at a time. Remember, lmbench was originally written to test 
workstation class performance, with 32MB being the LOW end.
 
> It was the first time I've ever ran lmbench, and it was quite fun. Maybe
> I'll go and get 1.1 and tweak it to pieces and see what numbers I can
> squeeze out to poke fun at my Linux friends  ;-)

Please apply the following patch to the lmbench port directory
(/usr/ports/benchmarks/lmbench). This is what I prepared to 
update lmbench to 1.1, but I never got around to complete it.

There are LOTS of new example results, and I bet somebody will 
be faster, now :)

Regards, STefan

Index: Makefile
===================================================================
RCS file: /usr/cvs/ports/benchmarks/lmbench/Makefile,v
retrieving revision 1.6
diff -C2 -r1.6 Makefile
*** Makefile	1996/11/18 11:21:57	1.6
--- Makefile	1996/11/22 21:35:20
***************
*** 7,11 ****
  #
  
! DISTNAME=	lmbench-1.0
  CATEGORIES=	benchmarks
  MASTER_SITES=	ftp://forte.mathematik.uni-bremen.de/pub/unix/benchmarks/
--- 7,11 ----
  #
  
! DISTNAME=	lmbench-1.1
  CATEGORIES=	benchmarks
  MASTER_SITES=	ftp://forte.mathematik.uni-bremen.de/pub/unix/benchmarks/
Index: files/md5
===================================================================
RCS file: /usr/cvs/ports/benchmarks/lmbench/files/md5,v
retrieving revision 1.1.1.1
diff -C2 -r1.1.1.1 md5
*** md5	1995/05/06 11:49:56	1.1.1.1
--- md5	1996/09/24 18:18:52
***************
*** 1 ****
! MD5 (lmbench-1.0.tar.gz) = 22651dd664376a13279f02af4eaacf84
--- 1 ----
! MD5 (lmbench-1.1.tar.gz) = 8dad46afd0c7fbaf8967d22b9d157da8
Index: patches/patch-aa
===================================================================
RCS file: /usr/cvs/ports/benchmarks/lmbench/patches/patch-aa,v
retrieving revision 1.1.1.1
diff -C2 -r1.1.1.1 patch-aa
*** patch-aa	1995/05/06 11:49:56	1.1.1.1
--- patch-aa	1996/09/24 18:37:24
***************
*** 1,11 ****
! --- ./src/Makefile.org	Fri Nov 25 08:53:00 1994
! +++ ./src/Makefile	Sat May  6 03:38:01 1995
! @@ -91,7 +91,7 @@
!  	$(MAKE) O=$O CC=cc CFLAGS="$(CFLAGS) -Dvfork=fork" all
!  
!  bsd:
! -	$(MAKE) O=$O CC=$(CC) CFLAGS="$CFLAGS -Duint='unsigned int'" all
! +	$(MAKE) O=$O CC=$(CC) CFLAGS="$(CFLAGS)" all
!  
!  Wall:
!  	@$(MAKE) clean
--- 1,28 ----
! *** src/Makefile.orig	Tue Sep 24 20:28:43 1996
! --- src/Makefile	Tue Sep 24 20:31:16 1996
! ***************
! *** 94,98 ****
!   
!   bsd:
! ! 	$(MAKE) O=$O CC=$(CC) CFLAGS="$(CFLAGS) -Duint='unsigned int'" all
!   
!   Wall:
! --- 94,98 ----
!   
!   bsd:
! ! 	$(MAKE) O=$O CC=$(CC) CFLAGS="$(CFLAGS)" all
!   
!   Wall:
! ***************
! *** 132,136 ****
!   $O/lat_pagefault:  lat_pagefault.c timing.c bench.h
!   	if [ $O = ../bin/linux -o $O = ../bin/bsd ];                               \
! ! 	then    cp /bin/true $O/lat_pagefault;                  \
!   	else    $(COMPILE) -o $O/lat_pagefault lat_pagefault.c $(LDLIBS); \
!   	fi
! --- 132,136 ----
!   $O/lat_pagefault:  lat_pagefault.c timing.c bench.h
!   	if [ $O = ../bin/linux -o $O = ../bin/bsd ];                               \
! ! 	then    cp /usr/bin/true $O/lat_pagefault;                  \
!   	else    $(COMPILE) -o $O/lat_pagefault lat_pagefault.c $(LDLIBS); \
!   	fi
Index: patches/patch-ab
===================================================================
RCS file: /usr/cvs/ports/benchmarks/lmbench/patches/patch-ab,v
retrieving revision 1.1
diff -C2 -r1.1 patch-ab
*** patch-ab	1996/02/01 07:55:41	1.1
--- patch-ab	1996/09/24 18:38:16
***************
*** 1,14 ****
- --- ./src/bw_tcp.c.org	Fri Nov 25 08:53:00 1994
- +++ ./src/bw_tcp.c	Wed Jan 31 23:25:22 1996
- @@ -128,9 +128,11 @@
-  	char	buf[SOCKBUF];
-  	int	sockbuf = SOCKBUF;
-  
- +#ifndef __FreeBSD__
-  	while (setsockopt(data, SOL_SOCKET, SO_RCVBUF, &sockbuf, sizeof(int))) {
-  		sockbuf -= 128;
-  	}
- +#endif
-  #endif
-  	bzero(buf, SOCKBUF);
-  	if (read(control, buf, SOCKBUF) <= 0) {
--- 0 ----



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Mutt.19961205231458.se>