mismatch results with disk performance

scuba at centroin.com.br scuba at centroin.com.br
Thu Jun 9 18:31:14 GMT 2005


Hi,

	No one else has seem this?
	I upgraded to 5.4 and the problem was kept.
	Should it be a problem with ICH5 driver?


On Tue, 7 Jun 2005 scuba at centroin.com.br wrote:

|Hi ALL,
|
|	I hope it´s not a known issue.
|	I´ve installed (express) a FreeBSD 5.3 (I know it´s not the last
|stable) on a new machine with the following hardware:
|
|	Asus P4P800 SE  (BIOS v. 1008)
|	2GB RAM ( 4x 512 DDR400 )
|	2 HDD Samsung SP0802N (80GB 7200rpm ata-100) 80 pins cable.
|
|	The HD were formated with newfs defaults, and the following
|results were the same using both as master (primary e secondary) or with a
|master / slave (same interface).
|
|	With diskinfo both performance are the same, but with "dd", the
|second disc (the slave or the secondary master), is always worst as if it
|were working in DMA2.
|
|	what should be the right results?
|
|	Here are the results:
|
|	DD: dd if=/dev/zero of=TEST count=1000 bs=64k
|
|	master:  60 Mb/s
|	slave:   16 Mb/s
|
|	The results with diskinfo (almost the same):
|
|/dev/ad0
|	512         	# sectorsize
|	80060424192 	# mediasize in bytes (75G)
|	156368016   	# mediasize in sectors
|	155127      	# Cylinders according to firmware.
|	16          	# Heads according to firmware.
|	63          	# Sectors according to firmware.
|
|Seek times:
|	Full stroke:	  250 iter in   5.477568 sec =   21.910 msec
|	Half stroke:	  250 iter in   4.140590 sec =   16.562 msec
|	Quarter stroke:	  500 iter in   6.093340 sec =   12.187 msec
|	Short forward:	  400 iter in   2.088111 sec =    5.220 msec
|	Short backward:	  400 iter in   2.532713 sec =    6.332 msec
|	Seq outer:	 2048 iter in   0.261596 sec =    0.128 msec
|	Seq inner:	 2048 iter in   0.259723 sec =    0.127 msec
|Transfer rates:
|	outside:       102400 kbytes in   1.743910 sec =    58719 kbytes/sec
|	middle:        102400 kbytes in   2.031982 sec =    50394 kbytes/sec
|	inside:        102400 kbytes in   3.354742 sec =    30524 kbytes/sec
|
|=========
|
|/dev/ad1
|	512         	# sectorsize
|	80060424192 	# mediasize in bytes (75G)
|	156368016   	# mediasize in sectors
|	155127      	# Cylinders according to firmware.
|	16          	# Heads according to firmware.
|	63          	# Sectors according to firmware.
|
|Seek times:
|	Full stroke:	  250 iter in   5.431290 sec =   21.725 msec
|	Half stroke:	  250 iter in   4.111275 sec =   16.445 msec
|	Quarter stroke:	  500 iter in   6.282551 sec =   12.565 msec
|	Short forward:	  400 iter in   1.741538 sec =    4.354 msec
|	Short backward:	  400 iter in   3.285028 sec =    8.213 msec
|	Seq outer:	 2048 iter in   0.259503 sec =    0.127 msec
|	Seq inner:	 2048 iter in   0.258212 sec =    0.126 msec
|Transfer rates:
|	outside:       102400 kbytes in   1.720602 sec =    59514 kbytes/sec
|	middle:        102400 kbytes in   1.998063 sec =    51250 kbytes/sec
|	inside:        102400 kbytes in   3.235904 sec =    31645 kbytes/sec
|
|
|- Marcelo
|



More information about the freebsd-questions mailing list