Linked by Eugenia Loli on Sat 12th Aug 2006 19:06 UTC, submitted by Rahul
Fedora Core "Overall, FC6 looks and feels like a more solid, polished version of FC5 rather than a wholesale revamp. I look forward to the finished product this fall." Check the short commentary here.
Thread beginning with comment 151882
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE: I/O Wait
by shotsman on Sun 13th Aug 2006 06:02 UTC in reply to "I/O Wait"
shotsman
Member since:
2005-07-22

Have you collected all the data and filed a bugzilla report yet?
IMHO, the SATA interface on many motheboards is a bit of a kludge and could do with a complete overhaul.
However, it could be a driver issue. Have you done an 'hdparm' to see how your drivers are configured?

I get some bad performance from SATA on Windows server 2003 as well as lower than expected throughput on Linux.

Reply Parent Score: 1

RE[2]: I/O Wait
by hraq on Sun 13th Aug 2006 06:23 in reply to "RE: I/O Wait"
hraq Member since:
2005-07-06

"Have you collected all the data and filed a bugzilla report yet?"

Actually I don't know whome to send this info to, which Unix distro and which linux distro

"I get some bad performance from SATA on Windows server 2003 as well as lower than expected throughput on Linux."

Actually I was working most of the time with PATA dirves (ie the classical ATA drives which are brand new from Seagate and Western Digital. And still the same issue appeared with me, all the times and in all occasions of heavy HDD access.

But performance wise the SATA drives made linux boot and behave faster all over; but felt like dominating and veto the CPU if other processes inturrupt for CPU time, I guess but I am not sure it is related to HDD scheduling kernel code in Unix and Linux, maybe, as MS pay more attention to these things especially with vista where they imply a very good scheduling for HDD time, and they have the so called no new process if the HDD activity is saturated, whereas the running processes that depend on HDD will get 100% guaranttee of uninturruptability.

On windows server 2003 with SP1 the problems happen most of the time from udf burned DVD+R disks, which cannot read from it at all, whereas it can burn it; the same issue doesn't exist at all in any Unix and Linux especially with fedora which is so reliable in this regard.

so, I/O access of HDDs are better in windows, whereas I/O access of Optical Drives are better in any Unix/Linux, all obtained from my experience.

Reply Parent Score: 1

RE[2]: I/O Wait
by netpython on Sun 13th Aug 2006 07:44 in reply to "RE: I/O Wait"
netpython Member since:
2005-07-06

I think it depends as usual on the chipset.
Using nforce3 doesn't give me any problems on whatever linux.In fact the performance of my raid0 configuration is overall better than that of XP.

Reply Parent Score: 1

RE[3]: I/O Wait
by hraq on Sun 13th Aug 2006 08:11 in reply to "RE[2]: I/O Wait"
hraq Member since:
2005-07-06

Do you mean that intel chipsets 865PE and 875 are not that good for linux; I think that intel publish more than nvidia to open source community. But if this is the case I will investigate this claim soon. Thanks anyway.

Reply Parent Score: 1