Weekly Edition
Return to the Press page
| |||||||||||||
|
| |||||||||||||
Linux in Government: Optimizing Desktop Performance, Part II (Linux Journal)
Linux Journal optimizes
the Linux desktop by enabling Direct Memory Access (DMA).
"Simply turning on using_dma enables a 16-bit mode, so I switched to
32-bit mode and found that my read time improved slightly. My buffered disk
reads went from 46.50MB per second to 46.52MB, not much of an
improvement. I also looked at turning on multiple sector I/O. This is a
feature of modern IDE hard drives that permits the transfer of multiple
sectors per I/O interrupt rather than the usual one sector per
interrupt. When this feature is enabled, it typically reduces operating
system overhead for disk I/O by 30-50%. On many systems, it also provides
increased data throughput of anywhere from 5% to 50%."
(Log in to post comments)
Linux in Government: Optimizing Desktop Performance, Part II (Linux Journal) Posted May 24, 2005 4:36 UTC (Tue) by darthmdh (guest, #8032) [Link] From memory, all these kinds of settings are enabled by the kernel by default where possible; at least in current stable kernels.
Linux in Government: Optimizing Desktop Performance, Part II (Linux Journal) Posted May 24, 2005 6:15 UTC (Tue) by conman (guest, #14830) [Link] Worse than that. If you have a true UDMA mode enabled and supported in your kernel the other settings (32 bit, unmask IRQ etc) do nothing.
Linux in Government: Optimizing Desktop Performance, Part II (Linux Journal) Posted May 24, 2005 6:20 UTC (Tue) by conman (guest, #14830) [Link] Here Jens explains misunderstanding of multicount for example:http://bhhdoa.org.au/pipermail/ck/2005-February/003003.html
SATA support is worse in Linux 2.6.11 than in Linux 2.6.6 Posted May 24, 2005 19:35 UTC (Tue) by Wills (guest, #1813) [Link] Is anyone here aware of any news on when smartctl and 2.6 kernels will fully implement the same range of functions for SATA that hdparm used to provide for SATA drives under Linux 2.6.6? I'm aware of smartctl from S.M.A.R.T. Monitoring Tools, but at least for smartctl-5.33 under Linux 2.6.11 it still lacks most functions of hdparm, e.g. putting drives into standby mode.I remember using a SATA drive under Linux 2.6.6 and being able to type hdparm -y /dev/hda to put the drive into standby mode, as well as being able to use the various other hdparm functions. However, from Linux 2.6.7 onwards, SATA disks were no longer accessible via the traditional /dev/hd[a-z]due to a complete changeover to the new experimental libata driver which makes SATA disks appear like SCSI disks at /dev/sd[a-z] and which prevents you from doing any of the useful hdparm tricks anymore on SATA disks such as putting them into standby mode. No doubt the changeover to libata has good technical justifications, but for users it just looks like a big step backwards in functionality. You'd think a change of this magnitude would get at least some mention in the documentation included in kernel tarballs, but unfortunately it didn't (and still doesn't).
SATA support is worse in Linux 2.6.11 than in Linux 2.6.6 Posted May 25, 2005 11:03 UTC (Wed) by conman (guest, #14830) [Link] Look for sdparm.
SATA support is worse in Linux 2.6.11 than in Linux 2.6.6 Posted May 25, 2005 17:31 UTC (Wed) by Wills (guest, #1813) [Link] Look for sdparmThank you for that. sdparm looks very useful, but unfortunately it is not able to change the power-saving mode of SATA drives under Linux 2.6.11, unlike hdparm which was able to do this for SATA drives under Linux 2.6.6, so this useful functionality remains regressed and Linux has no method anymore of forcing a SATA drive to enter standby mode:
/dev/sda: ATA Seagate 8Z400P0 RPO7 >> Power condition mode page not supported
|
|
||||||||||||