Suggestions for Core 2 Duo systems that use PCI - not PCI express?

[Paul]

Hello,

I actually do not see why a newer core would be faster. 45 nm technology,
for example, is the size of the gate or transistor, and more can be placed
on the chip but that, by itself, would not mean faster.

The only two thing I can think of, with regard to the core only, not the
caches, etc. is the fact that they have instructions for moving data 64 bits
at a time and a higher clock speed.

The clock speed limit has been reached and most/many programs are 32 bit,
so, they would not be doing 64-bit register moves.

The only other thing is possibly there are scenarios they can optimize for,
at the cpu level, but that would not make a cpu 4 or 5 times faster.

IMHO.

--g

It is the clock speed times the number of instructions processed
in parallel, that gives the total number of instructions processed.

The Core2 has slightly better parallelism, and also has
an instruction fusion feature, where two instructions
can be bundled together. It means on average, it handles
more instructions per clock cycle, than the previous
processors.

What I've tried to do, in this thread, is offer you benchmarks
run by other people. Could some of the results be faked ?
I suppose so. SuperPI was modified slightly, to try to prevent
faking the results. There is a validator, for example, which
is intended to make cheating difficult.

http://www.xtremesystems.com/pi/

At the very least, you can download SuperPI, and replicate
the 2GHz Mobile Pentium 4 - M result, and see if it agrees.

http://www.xtremesystems.com/pi/super_pi_mod-1.5.zip

Paul

 

[~misfit~]

VirtualDub is multi-core aware no? If that's right then even the
entry-level core2 Duo should do what you want.

The author's web site indicates it is NOT.

I use a benchmark designed by Australian PC User magazine (they made
it available to their users on a cover disk) called 'UserBench
Encode 2009' http://darrenyates.com.au/?p=573 that uses a mix of
single threaded and multi threaded work, to simulate real-world use.
It only benches CPU/FSB/RAM. They set the score of 10 to represent a
2GHz Pentium 4 with a 533 FSB running 1GB RAM running XP SP2.

Examples of machines I'm using here, all running XP SP3:

[IBM R51 ThinkPad] Dothan 1.7GHz/400MHz/2GB RAM. Score: 13.82
Flatmate's desktop. E4500 2.2GHz/800/2GB RAM. Score: 29.77
My desktop. E7300 2.66GHz/1066/4GB RAM. Score: 39.45
The above with the FSB raised to 1333. Score 48.81

Unfortunately, taking HD I/O out of the equation will tend to grossly
overstate relative performance gains. VirtualDub's author even warns
that on a fast CPU the app can become "disk-bound," and the faster
CPU will then work at much less than full capacity.

Thanks for that. can you tell us exactly what CPU will provide as much power
as possible but not become disk-bound using a SATA-II HDD?

No?

Didn't think so. In the meantime a fast CPU is always a good investment.

 

[JR Weiss]

VirtualDub is multi-core aware no? If that's right then even the
entry-level core2 Duo should do what you want.

The author's web site indicates it is NOT.

I use a benchmark designed by Australian PC User magazine (they made
it available to their users on a cover disk) called 'UserBench
Encode 2009' http://darrenyates.com.au/?p=573 that uses a mix of
single threaded and multi threaded work, to simulate real-world use.
It only benches CPU/FSB/RAM. They set the score of 10 to represent a
2GHz Pentium 4 with a 533 FSB running 1GB RAM running XP SP2.

Examples of machines I'm using here, all running XP SP3:

[IBM R51 ThinkPad] Dothan 1.7GHz/400MHz/2GB RAM. Score: 13.82
Flatmate's desktop. E4500 2.2GHz/800/2GB RAM. Score: 29.77
My desktop. E7300 2.66GHz/1066/4GB RAM. Score: 39.45
The above with the FSB raised to 1333. Score 48.81

Unfortunately, taking HD I/O out of the equation will tend to grossly
overstate relative performance gains. VirtualDub's author even warns
that on a fast CPU the app can become "disk-bound," and the faster
CPU will then work at much less than full capacity.

Thanks for that. can you tell us exactly what CPU will provide as much
power as possible but not become disk-bound using a SATA-II HDD?

No?

Didn't think so. In the meantime a fast CPU is always a good investment.

Gee... Nowhere did I claim that a fast CPU was NOT a good investment! In
fact, IIRC I suggested he buy as fast a C2D as possible... All I "claimed"
was that he was unlikely to see 3x system performance gains using the
constraints he placed on his system.

OTOH, in the context of the OP's question, you tend to indicate with your
numbers that upgrading the CPU/RAM/MoBo will result in overall system
performance gains of 4x or more. That is simply not the case, because total
system performance is not that simple.

The OP gave an indication of the primary apps he uses, and it is unlikely it
can take advantage of more than [parts of] 2 cores. Also, HD I/O will
likely limit the effect of a very fast CPU. Therefore all the
multithreadded performance indicated by your benchmark would not be achieved
for the app he cited.

 

[~misfit~]

VirtualDub is multi-core aware no? If that's right then even the
entry-level core2 Duo should do what you want.

The author's web site indicates it is NOT.


I use a benchmark designed by Australian PC User magazine (they
made it available to their users on a cover disk) called 'UserBench
Encode 2009' http://darrenyates.com.au/?p=573 that uses a mix of
single threaded and multi threaded work, to simulate real-world
use. It only benches CPU/FSB/RAM. They set the score of 10 to
represent a 2GHz Pentium 4 with a 533 FSB running 1GB RAM running
XP SP2. Examples of machines I'm using here, all running XP SP3:

[IBM R51 ThinkPad] Dothan 1.7GHz/400MHz/2GB RAM. Score: 13.82
Flatmate's desktop. E4500 2.2GHz/800/2GB RAM. Score: 29.77
My desktop. E7300 2.66GHz/1066/4GB RAM. Score: 39.45
The above with the FSB raised to 1333. Score 48.81

Unfortunately, taking HD I/O out of the equation will tend to
grossly overstate relative performance gains. VirtualDub's author
even warns that on a fast CPU the app can become "disk-bound," and
the faster CPU will then work at much less than full capacity.

Thanks for that. can you tell us exactly what CPU will provide as
much power as possible but not become disk-bound using a SATA-II HDD?

No?

Didn't think so. In the meantime a fast CPU is always a good
investment.

Gee... Nowhere did I claim that a fast CPU was NOT a good
investment! In fact, IIRC I suggested he buy as fast a C2D as
possible... All I "claimed" was that he was unlikely to see 3x
system performance gains using the constraints he placed on his
system.
OTOH, in the context of the OP's question, you tend to indicate with
your numbers that upgrading the CPU/RAM/MoBo will result in overall
system performance gains of 4x or more. That is simply not the case,
because total system performance is not that simple.

The OP gave an indication of the primary apps he uses, and it is
unlikely it can take advantage of more than [parts of] 2 cores. Also, HD
I/O will likely limit the effect of a very fast CPU. Therefore all the
multithreadded performance indicated by your
benchmark would not be achieved for the app he cited.

On top of the data from benches, I know that when I went from an unlocked
Barton XP2500+ running at XP3200+ speed (2.2GHz) to my first C2D (E4500)
converting DVDs to xvid (AGK, utilising VirtualDub) and converting xvids to
DVDs (using ConvertXtoDVD, a reasonably up-to-date version, the old version
wasn't multi-core aware) was easilly three times faster and the CPU speed in
MHz was exactly the same.

So that's personal experience as well as theoretical.

 

[DevilsPGD]

In message <[email protected]> "~misfit~"

Thanks for that. can you tell us exactly what CPU will provide as much power
as possible but not become disk-bound using a SATA-II HDD?

No?

If you limit yourself to a single SATA-II drive, you'll almost
definitely be disk-bound. Try again reading from one RAID-0 array and
writing to another and you'll have a different game entirely.

I've maxed out my Q6600 trying to tweak for slightly better video
quality in a similar configuration, but with a maximum bitrate cap since
the goal was to make a video playable from a USB flash drive of somewhat
limited specifications.

Didn't think so. In the meantime a fast CPU is always a good investment.

Not if you're more interested in using less energy.

 

[~misfit~]

In message <[email protected]> "~misfit~"


If you limit yourself to a single SATA-II drive, you'll almost
definitely be disk-bound. Try again reading from one RAID-0 array and
writing to another and you'll have a different game entirely.

My machine has five SATA-II drives and I always make sure source and
destination are different. This is basic stuff no?

I've maxed out my Q6600 trying to tweak for slightly better video
quality in a similar configuration, but with a maximum bitrate cap
since the goal was to make a video playable from a USB flash drive of
somewhat limited specifications.


Not if you're more interested in using less energy.

Huh? My E7300 (2.66GHz) uses less energy than my E4500 (2.2GHz) did. The
change from 65nm to 45nm probably made a big difference as well as CPU
optimisation. Also, Core2Duos drop their speed and vcore when not needed so
are very economical indeed to run.

Cheers,

 

[DevilsPGD]

In message <[email protected]> "~misfit~"

My machine has five SATA-II drives and I always make sure source and
destination are different. This is basic stuff no?

Maybe, but your message said "not become disk-bound using a SATA-II
HDD", not "not become disk-bound using SATA-II HDDs", The difference in
performance between a couple otherwise unoccupied pair of
source/destination drives vs a single congested
OS+applications+data+source+destination drive can be more then double
due to the need for seek times.

Adding a quality RAID-0 controller and making sure source and
destination are on different arrays backed by different physical disks
can nearly double performance of disk-bound linear reads vs an optimal
dedicated-source and dedicated-destination solution.

Who here compresses enough videos and cares about performance enough to
justify drives entirely for that purpose?

Huh? My E7300 (2.66GHz) uses less energy than my E4500 (2.2GHz) did. The
change from 65nm to 45nm probably made a big difference as well as CPU
optimisation. Also, Core2Duos drop their speed and vcore when not needed so
are very economical indeed to run.

Sure, but the Atom I've got running in my firewall uses far less power,
your CPU may well use more power at load then that entire box does
running at load.

The 65nm to 45nm makes a difference, but all other things being equal, a
slower CPU still users less power.

 

[~misfit~]

In message <[email protected]> "~misfit~"


Maybe, but your message said "not become disk-bound using a SATA-II
HDD", not "not become disk-bound using SATA-II HDDs",

My bad. I just figured that anyone looking for performance would be using
more than one disk. I meantioned 'disk' simply because it was the interface
I was talking about, SATA-II. I added an 'a' and missed an 's', shoot me.

The difference
in performance between a couple otherwise unoccupied pair of
source/destination drives vs a single congested
OS+applications+data+source+destination drive can be more then double
due to the need for seek times.

Adding a quality RAID-0 controller and making sure source and
destination are on different arrays backed by different physical disks
can nearly double performance of disk-bound linear reads vs an optimal
dedicated-source and dedicated-destination solution.

Who here compresses enough videos and cares about performance enough
to justify drives entirely for that purpose?

Perhaps the OP seemingly?

Sure, but the Atom I've got running in my firewall uses far less
power,

Yeah, and it's well powerful enough to run a firewall. I wouldn't want to
run much else on an Atom though, at least not a single core one. They bench
at about the same as my old Celeron 1GHz, excellent for Windows 98, a bit
crap for XP. (Ok, the Atom uses around a 10th of the power but that's
irrelevant if it's not doing what you want. Horses for courses.)

your CPU may well use more power at load then that entire box
does running at load.

Not with all those HDDs humming along. ;-)

I've run it through a meter and the difference in consumption between idle
and full load (when overclocked to ~3.6GHz) is around 35W, taking the
machine from 180W to 215W. Run a 3D benchmark to work the GPU and that'll
take the thing over the 250W mark.

The 65nm to 45nm makes a difference, but all other things being
equal, a slower CPU still users less power.

Yes, but all things are rarely equal. A single core Pentium 4 Prescott
running at less than 75% the speed of my 45nm C2D uses far more juice.....
They didn't call them PrescHOT for nothing. What about those old Athlon
Thunderbirds? They got pretty damn hot and heat-pipe coolers hadn't the
mainstream by then. For contrast, the Intel Tualatins (the grandfather of
the 'Core' architecture) were very cool running and efficient.

 

[DevilsPGD]

In message <[email protected]> kony

If you have the videos, you already have a need for that
much storage space.

Why?

I do a fair amount of video processing, but rarely have more then
20-40GB worth of video on my system at once. As soon as the videos are
processed they're offloaded to a long term storage device (in my case a
set of file servers, for most of my friends this mean DVDs)

I can't justify the space of more then a single 400GB drive or so on my
desktop, it isn't until you consider performance that several large
drives make sense for my usage pattern.

Not everyone is me, but having a need to process videos doesn't
automatically imply a need to have multiple large capacity storage
devices.

Beyond a certain capacity it is even
less expensive to buy two hard drives at half the capacity
of one larger one, plus these two drives do not have to be
dedicated solely to video processing, they can also speed up
the entire system by dividing which holds OS, applications,
pagefile, user data, etc.

At what point would that be, exactly? I just compared every Seagate
7200.10 and .11 drive on my price lists, the $/GB is cheapest across the
board on the largest drives, at no point is it ever more economical to
buy two smaller drives.

 

[DevilsPGD]

In message <[email protected]> "~misfit~"

Yeah, and it's well powerful enough to run a firewall. I wouldn't want to
run much else on an Atom though, at least not a single core one. They bench
at about the same as my old Celeron 1GHz, excellent for Windows 98, a bit
crap for XP. (Ok, the Atom uses around a 10th of the power but that's
irrelevant if it's not doing what you want. Horses for courses.)

Mine is actually a dual core...

I've run it through a meter and the difference in consumption between idle
and full load (when overclocked to ~3.6GHz) is around 35W, taking the
machine from 180W to 215W. Run a 3D benchmark to work the GPU and that'll
take the thing over the 250W mark.

IIRC I've only got a 180W PSU and I'm not running near capacity. The
next time I need to reboot it I'll stick in a power meter (It's FreeBSD,
so give me about 4-5 years unless I decide to upgrade the kernel at some
point) and find out what sort of power it's actually consuming.