Kedar Patankar wrote:
> I have tried this thing with the special 80 core cable (that is
> needed for udma-66) and kernel 2.2.12 and 2.2.14 (with andre hendrick's
> patches for udma).
I just checked with kernel 2.3.41, there is no explicit support for the
i810 hard disk controller! What u must be using is the generic PCI
chipset with DMA enabled. Correct me if i am wrong.
> But I just don't manage to get anything *decent*, as measured
> with hdparm -t -T /dev/hd?. I keep on getting 8 MB/s as the best case.
> This is after I have turned on 32 bit i/o, dma, multisector readahead,
> with both hdparm v3.5 (that comes with RH6.1) and hdparm 3.6 - the
> latest version from mark lord's page.
AFAIK, multiword DMA requires special kernel support. Using the generic
PCI bus master drivers will be a bottle neck on the true performance of
the drive. The latest issue of Kernel Traffic (KT) has some stuff on the
fundas.
> Now the interesting part is in this month's pcq they have reviewed
> this same drive Seagate U8 and they claim that they get 24MB/s under M$.
Intel would have provided bus mastering drivers along with the
motherboard which PCQ would have used to get the very best performance.
> I would like to hear experiences if any that any people on this
> list may have to share.
1. IMHO, hdparm -t /dev/hd* alone is not last word in bench marking.
2. Hedrick's IDE patches offer only *experimental* support, they may or
may not work as good as the ones included in the stable tree.
3. Use the same machine for doing the benchmark.
4. Do u seriously think that the Seagate U8 give a throughput of 24 Mega
Bytes per sec? If it's true, then to read a file of 24 MB, it should
take only *1* sec.Try reading a file of 24 MB under windows and time it!
5. Perhaps it's the burst transfer rate they are referring to and that
too rated in Mega 'Bits' per sec.
6. Wherever possible, average transfer rates should be quoted as they
indicate real world performance.
U could create a ram disk of say 32 MB, copy a 32 MB file from the disk
to the ram disk and time it. Size of file divided by the time taken to
transfer the file, would give the avg. transfer rate. U can also do this
when u have a RAID setup where hdparm cannot be used. Correct me if i am
wrong here! (HANISH: Thanx for this tip)
Comments anyone?
--Shanu
- - - - - - - - - - - - - - - - - - - - - = = (SHANU) = = - -
St. Albert's College Debian/GNU Linux
Cochin India Potato Powered
--------------------------------------------------------------------------
LI is all for free speech, but this list was created for a purpose --
to help popularise Linux in India. If your messages are counterproductive
to that purpose, your privileges to submit messages can and will be revoked.
