HD cache and performance?

  • Thread starter Thread starter =?ISO-8859-1?Q?Peter_Bj=F8rn_Perls=F8?=
  • Start date Start date
De Moni said:
Rod Speed wrote
Click to expand...
So Samsung should overally be faster than Seagate in everything.
Click to expand...

Nope, most obviously with ops that involve heaps of very small files
like favourites etc. In that case the most important advantage the
Samsung has over the Seagate, the sectors per track, would be
invisible and the speed of the ops are entirely swamped by the
OS level detail and so the drives would appear much more equal.
It's strange then that I haven't noticed performance increase in any other use than copying large
files.
Click to expand...

Thats the main area where better sectors per track becomes
most noticeable because the that operation is much more
sensitive to the transfer rate from the platter surface to memory.
WinXP boot time is about the same,
Click to expand...

Because that time is completely dominated by what XP
is doing, not the transfer rate of the files from the drive.
games don't load any faster etc.
Click to expand...

Ditto.

The SR Gaming DriveMark 2002 gaming benchmark numbers in
http://www.storagereview.com/php/be...&numDrives=1&devID_0=252&devID_1=241&devCnt=2
are however something like 30% better with the Samsung.
Oh, well. The reason anyway I upgraded hard drive was that it felt like it took half a day for
Seagate to author DVD, which requires copying gigs of data. Now Samsung does it a lot faster, even
if 8Mb cache has nothing to do with it :-)
Click to expand...

Yeah, its all about sectors per track in that case too.
 
Actually huh?

Guess again, Babblebot. Ever heard of read-ahead caching?
They're both NTFS, Win XP Home, defgragmented about once a week.

Modern technology,
Click to expand...
I guess.
Click to expand...

As always. Pity you always present your drivel as fact, babblebot.
 
That is called command-queuing.
Click to expand...

No, it's not.
Thats just batch reading files from source into memory as far as memo-
ry provides and batch writing them out from memory to the destination.
And yes, modern SATA disks (and old SCSI disks) can do that.
Click to expand...

Nothing to do with it.
This is the one speed-up the OS cannot do as well,
since it does not know the real HDD geometry.
Click to expand...

Of course it does.
However at least OSes intended to be used in servers (e.g. Linux) do this
as well to some degree.

Well, for that more memory in the HDD might actually make a difference,
since it could defer more writes and optimize better.
Click to expand...

If the OS requests are of the type 'queued'.
Hmmm. OK, if we are talking about SATA drives here with native (!)
command queuing, that could give a boost in speed when writing.
Click to expand...

Nope, not with standard writes.
For reading, the native command queuing could also give a significant
speed improvement, but here the amount of available memory does not
matter so much, since the data is not stored in it, just the (very small) requests.
Click to expand...

What makes you think it cannot be done for writes too.
So probably your observed speed improvement is really a combination
of a) far faster disk b) native command queuing and c) (minor effect)
more memory in the drive.
Click to expand...

You are just guessing as you go along, aren't you babblebot.
Sleep deprevation setting in?
 
That was interesting!
Click to expand...

Yeah, but not in the way you think.

So, by "native" command queuing, you mean something the drive does unto
itself as part of its basic algorithms without explicitly having the extra
commercially named "NCQ" and an "NCQ" controller to run it?
Click to expand...

Actually, you need to use the proper read and write commands to have
the drive do reordering, other wise it just executes them as they come in,
one at a time.
As to a single command, the drives can optimize those all by themselves.
 
Ed Light said:
My new WD with 16Mb cache amazes me over my older 2Mb Samsung, in opening
the My Documents folder in file lists.
Click to expand...
The first time in a session, the Samsung takes way longer, much longer than
the 50% increase in serial reads on the WD could explain. I don't know
how much of that is the cache.
Click to expand...

The files or their info can only be in cache if they have been accessed before.
Or, if the drive does read ahead caching it reads more data then the OS
requests. So if the OS reads data and then starts to process it and then reads
some more, the drive already has the data in cache when it is finally requested.
The bigger the cache, the more data can be cached ahead, the bigger the speed-
up effect can be.
The My Documents contains thousands of cookies and favorites, all tiny.
Click to expand...

Which, IIRC, can make a difference if you are using NTFS.
The My Documents were at nearly the same position on both disks.
Click to expand...

That doesn't say much if you are listing directories, that may differ if
you are using FAT, maybe even with NTFS too. The order in which the
files were aquired can make a difference, so that directories are spread
all over the drive once the initial reserved directory space was exhausted.
A copy of such disk may result in all directory space being defragmented
and become contiguous.
 
NCQ is ''Native Command Queuing''.
Click to expand...

Which actually stands for tagged command queuing but to offset it against
the simpler command queuing of ATA/ATAPI they named it NCQ.

NCQ for SATA actually needs (for some reason
that eludes me) the support of the controller as well.
Click to expand...

Because of the different bus protocol involved, maybe?
SCSI has been doing this for a long time
Click to expand...
without any controller support whatsoever.
Click to expand...

Nope. Support of Queuing is optional, in other words:
a device (including host adapters) may not support it.
The commands themselves may not have differed but
the parameters sent with them in messages sure did.
Apart from that, it still needed OS support too.
Maybe a standard SATA controller can only send one command and
has to wait until it is completed before sending the next...
Click to expand...

Keep dreaming, Babblebot.
 
Back
Top