RAID versus 10,000 rpm?

[Rod Speed]

Very generally speaking, how much of a
performance boost does that produce?

Bugger all in practice. Bet you couldnt pick it in a double blind trial.

In spades when you have enough physical ram so it doesnt get used much.

 

[Arno Wagner]

Ignorance is bliss.

More likely, use of the wheel.

Don't get too advanced, for many people reading is easier than
using a wheel.... ;-)

Arno

 

[Arno Wagner]

Very generally speaking, how much of a performance boost does that
produce?

In heavy paging situations it can be one or two orders of magnitude
at least I have seen that under Linux. But that is an extreme case.

With no paging the boost is zero.

Arno

 

[dannysdailys]

I don't think there will be much of an advantage at all either way.
Forget the 5,400 drives. I'm surprised they're still made

The main point I'd like to make, is if you use two drives in a RAID
array, you multiply your chances of a hard drive failure. If yo
lose one of the drives on a RAID 0, you lose the whole ball of wax

If you do go to a RAID 0, get a pair of 133 Maxtors. Hell, what yo
should do is get a pair of 133's and RAID 1 them. This will give yo
data security. RAID 0 isn't a true RAID at all

I'm not sure why you're concerned about it. Unless you run huge dat
bases that are constantly quieried, you won't notice any differenc
anyway. Except for boot up, your hard drive sits there basicall
idleing anyway

 

[John Doe]

I don't think there will be much of an advantage at all either way.

I'm convinced, thanks to the replies.

I'm not sure why you're concerned about it. Unless you run huge data
bases that are constantly quieried, you won't notice any difference
anyway. Except for boot up, your hard drive sits there basically
idleing anyway.

Due to some recent evaluation and testing, I have a very good feeling
that a faster disk will improve speech recognition which is a major
resource hog used for everything, including resource hungry games. (A
big game runs mostly off of the hard disk after a long load time, but
some also load scenarios/scenery during play.)

My current thinking is to go with the much less expensive 37 GB
10,000 rpm drive and use the difference for an ordinary 7200 rpm
drive as a secondary HDD.

Thanks.

 

[John Doe]

....

Are you using a noise cancelling USB microphone? If not you might
try one

I bought one, removed the analog wire, and put the microphone right
next to the little circuit board. Anyway, yup, it's excellent.
Microphones are fishy, like most analog stuff <grumble>.

I can hear the hard disk working whenever I speak. It might have
something to do with a probably necessary part of the programming.
Speech recognition is intensive stuff. The hard disk is about half
full and well defragmented. The system has 1 GB of RAM, currently
using just over 256 MB under Windows XP. I don't know why the SR
works the hard disk when trying to recognize speech, it reminds me
of hardware depraved days when pulling down a Windows menu required
hard disk access.

But what games use speech recognition?

I do it manually. It requires scripting combined with speech
recognition (like in Dragon NaturallySpeaking Professional).

For what it's worth. Real-time strategy might be the best example
(to make it more like strategy and less like an arcade game). Maybe
turn based strategy, but I don't play that. I did some research and
found that some of we flight simmers use voice scripting
(VoiceBuddy?). It is doing very well to remove the click fest in Age
of Empires Conquers. You can make time to review and think about
what is going on, to enjoy some of the untapped very deep strategy
content in RTS, instead of just pointing and clicking. It's really
amazing how much frantic clicking good players do. They can beat me
in a short game, but maybe not in a 5 1/4 hour game when the click
count reaches perhaps 20,000. No more CTS for me. Also. The
automation doesn't require painfully accurate speaking if you put
the SR on Command Mode during the game. Loud ambient sound might
strain the CPU since SR attempts to recognize it.

 

[J. Clarke]

I'm convinced, thanks to the replies.


Due to some recent evaluation and testing, I have a very good feeling
that a faster disk will improve speech recognition which is a major
resource hog used for everything, including resource hungry games. (A
big game runs mostly off of the hard disk after a long load time, but
some also load scenarios/scenery during play.)

Are you using a noise cancelling USB microphone? If not you might try one
of those before you try dinking with the disks.

But what games use speech recognition?

 

[Bob Willard]

I'm convinced, thanks to the replies.




Due to some recent evaluation and testing, I have a very good feeling
that a faster disk will improve speech recognition which is a major
resource hog used for everything, including resource hungry games. (A
big game runs mostly off of the hard disk after a long load time, but
some also load scenarios/scenery during play.)

My current thinking is to go with the much less expensive 37 GB
10,000 rpm drive and use the difference for an ordinary 7200 rpm
drive as a secondary HDD.

Thanks.

The 74GB 10K RPM HD was a redesign of the earlier 37GB 10K RPM HD, and
is much faster; check with www.StorageReview.com for details. This PC
has a 74GB 10K RPM HD as the system HD, and a 250GB 7.2K RPM HD as the
data HD; seems to be a great combo.

 

[John Doe]

....

The 74GB 10K RPM HD was a redesign of the earlier 37GB 10K RPM HD,
and is much faster; check with www.StorageReview.com for details.
This PC has a 74GB 10K RPM HD as the system HD, and a 250GB 7.2K
RPM HD as the data HD; seems to be a great combo.

Seems to me that the current 37 GB 10000 rpm drive is a near
duplicate of the 74 GB drive. I find nothing on Western Digital's
web site to suggest the two drives are different except for size.

 

[Bob Willard]

...



Seems to me that the current 37 GB 10000 rpm drive is a near
duplicate of the 74 GB drive. I find nothing on Western Digital's
web site to suggest the two drives are different except for size.

Looks like you are right -- it seems that WDC now has a 37GB HD
which is a 1-platter version of its 74GB 2-platter HD and would
have the same performance. I was comparing the WDC 74GB Raptor
with the earlier 37GB Raptor, and there was a significant gain
in performance (65 MB/s v. 50 MB/s in STR); my apologies.

And to think that WDC didn't personally notify me. ;-)

 

[Gerhard Fiedler]

I can hear the hard disk working whenever I speak. It might have
something to do with a probably necessary part of the programming.
Speech recognition is intensive stuff. The hard disk is about half
full and well defragmented. The system has 1 GB of RAM, currently
using just over 256 MB under Windows XP. I don't know why the SR
works the hard disk when trying to recognize speech, it reminds me
of hardware depraved days when pulling down a Windows menu required
hard disk access.

I'm having a problem that looks similar, and I almost think this is a
Windows thing.

I often have quite a number of programs loaded. One of them that is up most
of the time is Mozilla. It stays in the background, and once in a while I
switch to it.

In a typical situation, Task Manager's process list would tell me that
Mozilla occupies 50 MB of physical memory, and 100 MB of virtual memory.
Switching to it causes quite some disk activity and because of this takes a
while. But on the Performance tab, the Task Manager says that there is 250
MB of physical memory /available/. I wonder why Windows pages Mozilla out
if it still has 250 MB of physical memory available. Seems to me that it
could leave it in physical memory until that memory is actually needed.

Maybe your problem has also to do with paging behavior, and maybe there's a
solution in configuring this rather than only looking at disk speed.

Does anybody know how to make sense of the memory figures in Task Manager,
and how to configure the paging behavior (WinXP)?

Gerhard

 

[Peter]

I'm having a problem that looks similar, and I almost think this is a

Windows thing.

I often have quite a number of programs loaded. One of them that is up most
of the time is Mozilla. It stays in the background, and once in a while I
switch to it.

In a typical situation, Task Manager's process list would tell me that
Mozilla occupies 50 MB of physical memory, and 100 MB of virtual memory.
Switching to it causes quite some disk activity and because of this takes a
while. But on the Performance tab, the Task Manager says that there is 250
MB of physical memory /available/. I wonder why Windows pages Mozilla out
if it still has 250 MB of physical memory available. Seems to me that it
could leave it in physical memory until that memory is actually needed.

Maybe your problem has also to do with paging behavior, and maybe there's a
solution in configuring this rather than only looking at disk speed.

Does anybody know how to make sense of the memory figures in Task Manager,
and how to configure the paging behavior (WinXP)?

So Mozilla has a lot used memory swapped out to pagefile (disk). 50 MB, that
is alot, no wonder it takes a moment to gat it back when switching tasks.
You can set a very small page file. Then not much memory will be swapped out
to disk. But there might be some problems with that approach.
There is not much you can do to optimize paging behavior though:
http://www.petri.co.il/pagefile_optimization.htm
http://www.tweakhound.com/xp/virtualmemory.htm

 

[John Doe]

A big game runs mostly off of the hard disk after a long load

Sorry, I meant "runs off of RAM" of course.