What's the bottleneck when CPU and disks are not strained?

Sometimes I wonder, when none of my four CPU cores are over 50%, and
disk activity is minimal, why would a process seem slow? The system
bus?

Mainly curious. Thanks.

If the slowness might have to do with system memory (RAM), is
there some Performance Monitor counter for that? I should be able
to see something that is maxed out. Having a Performance Monitor
counter showing that bottleneck would be a big deal here.

Thanks.

John Doe writes:

Sometimes I wonder, when none of my four CPU cores are over 50%, and
disk activity is minimal, why would a process seem slow? The system
bus?

Mainly curious. Thanks.

Well, define slow. Interactive use? Another program? And compared to
what, that is, when does the system not seem slow?
--
Nature fits all her children with something to do, He who would write
and can't write, can surely review.
- James Russell Lowell

On Oct 24, 9:07*pm, John Doe wrote:
Sometimes I wonder, when none of my four CPU cores are over 50%, and
disk activity is minimal, why would a process seem slow? The system
bus?

Mainly curious. Thanks.

Probably busy waiting for orders from the botnet?

On 25 Oct 2011 03:07:46 GMT, John Doe wrote:

Sometimes I wonder, when none of my four CPU cores are over 50%, and
disk activity is minimal, why would a process seem slow? The system
bus?
1. processes than can't multithread and multithreaded stuff
that is waiting for other threads

Even things that can multithread in general can get slowed down.
Things that are piped (like Unix pipes or the stuff I do
that has virtual encrypted disks) can have lots of processes
stalled waiting for the previous stage to send it the next
buffer's worth of stuff.)

2. things doing random access or un-buffered disk activity.
The disks might not look busy but data transfer rate goes
down a great deal. The I/O count on the disks might not
look all that high, but until you look carefully you might
not see that the disk is waiting for seeks/rotation.

3. Perhaps some software resource is tied up.

4. Or, as you say, it could be a bus.

Sometimes data is moved around more times than you
might have thought. Be sure to count everything,
including graphics.

I haven't played around with this stuff for more than 20 years,
but when it then DMA was usually involved.

When you add in the graphics stuff and the fact that data movements
by DMA might use 2 to 4 times the bandwidth you though you were
using, perhaps a bus is the bottleneck.

5. Maybe the protocol overhead on USB or something is using more
bandwidth than you think. (For example, a not too busy disk
on USB might be a bottleneck.)

6. I also found on some systems that everything had to wait for
certain types of floppy disk access; perhaps there is some
modern example of this. (I never found out if the floppy issue
was due to poor software design or some hardware limitation.

Mainly curious. Thanks.

John Doe wrote:
Sometimes I wonder, when none of my four CPU cores are over 50%, and
disk activity is minimal, why would a process seem slow? The system
bus?

Mainly curious. Thanks.

I've seen cases I can't explain.

Paul

On Oct 24, 11:07 pm, John Doe wrote:
Sometimes I wonder, when none of my four CPU cores are over 50%, and
disk activity is minimal, why would a process seem slow? The system
bus?

Mainly curious. Thanks.

Coding efficiency, too. Some assembly compiled routines were
preferable (adaptable for spawning into a command interpreter process
or batched) over higher-language levels of abstractions - NET
frameworks, DLLs, or whatever else makes for accompanying arrays
graphical poison as preferably pretty to actually getting on the stick
for a rushjob. I've heard mention these new AMD processors are geared
for more efficient core interaction in terms of shared core
arbitration when dealing with programs not specifically written for a
multi-core platform;- but since they're already out and being sold,
I've as well heard a few sceptical reactions to implementing the
concept. Actually, past an unpopular conundrum for code-level
incompatibility with abandoned software, doesn't seem as there's much
choice in pragmatic terms, unless the rules of the universe were bent
past multicores contained in speeds at something higher than present
3-4Ghz processors. I mean, how many times does it take to get tired
of hearing the same 4-year-old proposal, that a multicore without
specific software lacks overall great advantage while only running
single processes over a single core without concurrence.

On Tue, 25 Oct 2011 19:15:45 -0400, Paul wrote:

John Doe wrote:
Sometimes I wonder, when none of my four CPU cores are over 50%, and
disk activity is minimal, why would a process seem slow? The system
bus?

Mainly curious. Thanks.

I've seen cases I can't explain.

It's called Windows ;-)
--
(\__/) M.
(='.'=) Due to the amount of spam posted via googlegroups and
(")_(") their inaction to the problem. I am blocking some articles
posted from there. If you wish your postings to be seen by
everyone you will need use a different method of posting.