Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 25 Apr 1998 15:52:17 -0500 (EST)
From:      "John S. Dyson" <toor@dyson.iquest.net>
To:        jak@cetlink.net (John Kelly)
Cc:        toor@dyson.iquest.net, dg@root.com, chat@FreeBSD.ORG
Subject:   Re: Andreessen: Linux use growing
Message-ID:  <199804252052.PAA12151@dyson.iquest.net>
In-Reply-To: <35432f5d.444084@mail.cetlink.net> from John Kelly at "Apr 25, 98 07:05:10 pm"

next in thread | previous in thread | raw e-mail | index | archive | help
> 
> Besides Slackware sales at Walnut Creek, I have not heard any evidence
> to support this.
> 
> While I have your attention, I found a BSDI source license at yard
> sale prices and decided to performance test it (2.1 and 3.0) against
> FreeBSD -current.  I was surprised that my disk benchmark was about 5%
> faster with BSDI, even compared against the latest FreeBSD -current.  
>
Raw disk performance differences (like IOZONE) shouldn't be all that
important.  There are other differences, like you might find that we
*have* to do *less* disk I/O than they do, so the effect for end user
performance,  under load, that FreeBSD is faster.

Note that it is possible for FreeBSD to be faster (larger MAXPHYS values),
with the new SCSI code...  That hasn't come to pass yet.  Our
latest/greatest IDE code, does allow for full (255*512) transfers now.

> 
> And their NFS leaves FreeBSD in the dust.  It's nearly as fast as a
> disk to disk copy on a single machine, and CPU consumption is lower.  
> 
Doesn't surprise me.  IMO, our (*BSD's) NFS needs work.  There is a
productization movement afoot to fix NFS.  Below is the output of my
iozone (NFS, between two -current machines), so I don't see how we
are all that terrible in performance, but I don't trust our NFS
very much...

>
> When all is quiet, BSDI running TOP shows 100% idle on a 486, and on
> that same machine, I've never seen better than 99.6% idle on FreeBSD.
> I wonder what they did differently to achieve that.
> 
Our LL code is probably totally different, also we have a daemon that
runs all of the time, sucking about that much CPU.

John



	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 1 Megabyte sequential file consisting of
	256 records which are each 4096 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 1 Megabyte file, 'iozone.tmp'...0.218750 seconds
Reading the file...0.015625 seconds

IOZONE performance measurements:
	4793490 bytes/second for writing the file
	67108864 bytes/second for reading the file

The test completed too quickly to give a good result
You will get a more precise measure of this machine's
performance by re-running IOZONE using the command:

	iozone 85 	(i.e., file size = 85 megabytes)

	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 1 Megabyte sequential file consisting of
	128 records which are each 8192 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 1 Megabyte file, 'iozone.tmp'...0.210938 seconds
Reading the file...0.007812 seconds

IOZONE performance measurements:
	4971026 bytes/second for writing the file
	134217728 bytes/second for reading the file

The test completed too quickly to give a good result
You will get a more precise measure of this machine's
performance by re-running IOZONE using the command:

	iozone 91 	(i.e., file size = 91 megabytes)

	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 1 Megabyte sequential file consisting of
	64 records which are each 16384 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 1 Megabyte file, 'iozone.tmp'...0.250000 seconds
Reading the file...0.007812 seconds

IOZONE performance measurements:
	4194304 bytes/second for writing the file
	134217728 bytes/second for reading the file

The test completed too quickly to give a good result
You will get a more precise measure of this machine's
performance by re-running IOZONE using the command:

	iozone 77 	(i.e., file size = 77 megabytes)

	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 1 Megabyte sequential file consisting of
	16 records which are each 65536 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 1 Megabyte file, 'iozone.tmp'...0.226562 seconds
Reading the file...0.015625 seconds

IOZONE performance measurements:
	4628197 bytes/second for writing the file
	67108864 bytes/second for reading the file

The test completed too quickly to give a good result
You will get a more precise measure of this machine's
performance by re-running IOZONE using the command:

	iozone 82 	(i.e., file size = 82 megabytes)

	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 32 Megabyte sequential file consisting of
	4096 records which are each 8192 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 32 Megabyte file, 'iozone.tmp'...10.664062 seconds
Reading the file...0.609375 seconds

IOZONE performance measurements:
	3146496 bytes/second for writing the file
	55063683 bytes/second for reading the file

	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 128 Megabyte sequential file consisting of
	16384 records which are each 8192 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 128 Megabyte file, 'iozone.tmp'...30.359375 seconds
Reading the file...2.382812 seconds

IOZONE performance measurements:
	4420964 bytes/second for writing the file
	56327439 bytes/second for reading the file

	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 128 Megabyte sequential file consisting of
	16384 records which are each 8192 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 128 Megabyte file, 'iozone.tmp'...32.117188 seconds
Reading the file...2.351562 seconds

IOZONE performance measurements:
	4179000 bytes/second for writing the file
	57075977 bytes/second for reading the file

	IOZONE: Performance Test of Sequential File I/O  --  V2.01 (10/21/94)
		By Bill Norcott

	Operating System: FreeBSD 2.x -- using fsync()

	Send comments to:	b_norcott@xway.com

	IOZONE writes a 256 Megabyte sequential file consisting of
	32768 records which are each 8192 bytes in length.
	It then reads the file.  It prints the bytes-per-second
	rate at which the computer can read and write files.


Writing the 256 Megabyte file, 'iozone.tmp'...79.187500 seconds
Reading the file...55.210938 seconds

IOZONE performance measurements:
	3389871 bytes/second for writing the file
	4861997 bytes/second for reading the file



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199804252052.PAA12151>