How do I compare CPU speed versus number of cores for performance?

[Danny D'Amico]

How would you compare dual-core 1.5GHz versus quad-core 1.0GHz for performance?
(Windows 8, touchscreen)

My sister asked me whether it's better to get the ASUS X200CA-HCL12050 or the
Dell Inspiron i3135-3750slv inexpensive $300 laptops.

Looking at the specs, the main difference seems to be:
Asus: dual core, 1.5GHz
Dell: quad core, 1.0GHz

Which helps in performance more, additional cores? Or faster speeds?
Note: I realize both help - but which is MORE important, cores or speeds?

 

[Danny D'Amico]

Which helps in performance more, additional cores? Or faster speeds?

I googled a bunch, as my sibling was on the phone at the store
trying to make a decision.

It seems that two cores is a huge improvement over one core; but,
quad core isn't such a big improvement, mainly due to the fact
that software has to be written for the quad core.

On the other hand, it seems that CPU speed almost always helps
(assuming equal bus speeds).

So, it seems, the REAL question is whether the SOFTWARE is written
to take advantage of the quad core or just the dual core.

So, now it's a software question:

Q: How can we tell if software can take advantage of dual/quad cores?

Software is typical Windows stuff for a typical desktop user.

 

[David W. Hodgins]

Which helps in performance more, additional cores? Or faster speeds?
Note: I realize both help - but which is MORE important, cores or speeds?

Once you go beyond a single core, speed of each core will matter more
than the number of cores, for most average users.

If the system is being used as a server, with heavy use of multi-threaded
applications, then more cores would be the priority.

As most applications run on regular desktop systems are not able to use
multiple cores, the speed of each core is the priority to look at.

Regards, Dave Hodgins

 

[John Doe]

How would you compare dual-core 1.5GHz versus quad-core 1.0GHz
for performance? (Windows 8, touchscreen)

My sister asked me whether it's better to get the ASUS
X200CA-HCL12050 or the Dell Inspiron i3135-3750slv inexpensive
$300 laptops.

Looking at the specs, the main difference seems to be: Asus:
dual core, 1.5GHz Dell: quad core, 1.0GHz

Which helps in performance more, additional cores? Or faster
speeds? Note: I realize both help - but which is MORE important,
cores or speeds?

Not sure about the specifics, but...

I totally disagree with the idea that speed per core is more
important then number of cores. But it depends on the technology.
Given modern technology, multiple cores rocks, the more the
better.

Good luck and have fun.

 

[Rich]

Once you go beyond a single core, speed of each core will matter more
than the number of cores, for most average users.

If the system is being used as a server, with heavy use of multi-threaded
applications, then more cores would be the priority.

As most applications run on regular desktop systems are not able to use
multiple cores, the speed of each core is the priority to look at.

Regards, Dave Hodgins

--

What if you are using different applications concurrently on a regular
desktop? Assuming the individual applications are not optimized for multiple
cores, would the OS use an idle core or let the programs queue for the same
single core? Wouldn't that make a dual or quad core processor running at the
same speed faster on a regular desktop running a browser, e-mail, Quicken,
viewing a video on YouTube, etc.?
Also, that doesn't include the numerous background processes Windows is
constantly running. Every time I view the core usage graphs in task manager
there is constant activity in all of them. Is it that a multi-core processor
would not run multiple programs faster but just would not work as hard as a
processor with less cores?

Thanks,

Rich

 

[Paul]

How would you compare dual-core 1.5GHz versus quad-core 1.0GHz for performance?
(Windows 8, touchscreen)

My sister asked me whether it's better to get the ASUS X200CA-HCL12050 or the
Dell Inspiron i3135-3750slv inexpensive $300 laptops.

Looking at the specs, the main difference seems to be:
Asus: dual core, 1.5GHz
Dell: quad core, 1.0GHz

Which helps in performance more, additional cores? Or faster speeds?
Note: I realize both help - but which is MORE important, cores or speeds?

Dell Inspiron i3135-3750slv AMD quad core
------------------------------------------

Dell Inspiron I3135-3750SLV uses A6-1450 processor. Quad core Temash, no turbo.
(If turbo is enabled on that processor, the mobile device needs better cooling.
If the advertisement makes no turbo claim, assume it's disabled for thermal reasons.
This is unlike other products with turbo, where you always get it.)

http://www.bestbuy.com/site/dell-in...lver/2843086.p?id=1219079575337&skuId=2843086
http://www.cpu-world.com/CPUs/Jaguar/AMD-A6-Series A6-1450.html

The A6-1450 Passmark is 1653. Passmark is multithreaded performance (movie transcode)
This is the kind of benchmark all commercial sites push, to make the processor
look more impressive than it really is. If all you do is transcode movies
(convert a DVD to play on your iPad screen), then this processor is the
winner. People seldom do this.

http://www.passmark.com/cpubenchmark/cpu.php?cpu=AMD+A6-1450+APU

SuperPI is for single threaded performance, and a better indicator
of typical behavior. SuperPI 1M take 46.8 seconds.

http://www.youcpu.com/en/cpu/AMD-Qu...for-Notebooks-A6-1450-APU-with-Radeon-HD-8250

The Dell processor uses half the power of the other one. No turbo...
If it had turbo, power would be close to the Celeron.

*******

ASUS X200CA-HCL1205O Intel Celeron dual core
----------------------------------------------------

The X200CA-HCL1205O actually has an "Oh" as the last character
in the part number, not a "zero".

http://www.bestbuy.com/site/asus-11...219083269681&skuId=2912671#tab=specifications

I can't find an authoritative site with the processor name.
It could be Celeron 1.5Ghz model 1007u. The Asus page for X200CA
doesn't deal in specifics, only generalities.

http://www.cpu-world.com/CPUs/Celeron_Dual-Core/Intel-Mobile Celeron 1007U.html

The 1007u has a Passmark of 1456 (movie transcode or Photoshop multithreaded filters...)

http://www.passmark.com/cpubenchmark/cpu_list.php

SuperPI is for single threaded performance, and a better indicator
of typical behavior. SuperPI 1M take 25.919 seconds. My E8400 at
3GHz takes 15.468 seconds. Some overclocked processors have done the
benchmark in 5 seconds. The 25.9 is significantly better than
the AMD 46.8 seconds. Strictly speaking, we should be using the 32M
benchmark, but nobody bothered to run it, as it takes too long. The
results of the 1M benchmark benefit from a large CPU cache. So the
1M can be "tainted", if one of the contestants has an extremely
large cache. That's not a problem with these two contestants.

http://hwbot.org/submission/2383538_costa_superpi___1m_celeron_1007u_25sec_919ms

The dual is winning on the SuperPI. But to be fair to the other processor,
it's slow, but it'll be consistently slow to the end of its life. (It's
like a diesel engine, good low end torque but not a race car.)
Whereas, as you load more and more software on the Celeron product, it's
going to slow down as it gets older. Bloat will let the air out of the
tires.

Don't forget to read the reviews for the units - things to check:

1) Flaky trackpads. It's an epidemic out there.
2) Keyboards that don't last.
3) LCD screens with poor viewing angle. TN panels are cheap and you
have to hold your head just right. IPS panels have a wider viewing
angle. You're not going to get an IPS panel on a $300 product.
4) Battery life and watt-hours. The Celeron may draw double the watts,
so check that the battery is bigger to accommodate it. If a mobile device
must have "max hours on battery), then the AMD wins in this case, as
it draws less power. And that doesn't happen very often on AMD

I have made no attempt to compare the graphics performance... (too lazy)

Summary table: Passmark SuperPI Power
(more is better) (less is better)

Dell Inspiron I3135-3750SLV A6-1450 quad 1653 46.8 seconds 8 watts
ASUS X200CA-HCL1205O Celeron dual core 1456 25.9 seconds 17 watts

HTH,
Paul

 

[Yousuf Khan]

How would you compare dual-core 1.5GHz versus quad-core 1.0GHz for performance?
(Windows 8, touchscreen)

Here's a great site for comparing the relative performances of two
processors.

http://cpuboss.com/

Just type the names of the two processors in each system into the site
above and watch the results! Very handy and quick way to compare.

My sister asked me whether it's better to get the ASUS X200CA-HCL12050 or the
Dell Inspiron i3135-3750slv inexpensive $300 laptops.

Looking at the specs, the main difference seems to be:
Asus: dual core, 1.5GHz
Dell: quad core, 1.0GHz

Which helps in performance more, additional cores? Or faster speeds?
Note: I realize both help - but which is MORE important, cores or speeds?

Well, after running the two processors through the above-mentioned site,
it comes up with this result:

http://cpuboss.com/cpus/Intel-Celeron-1007U-vs-AMD-A6-1450

And it turns out that it rates both processors exactly identically
5.9/10! So it's a very good close competition here.

Another good comparison site is cpu-world.com:

http://www.cpu-world.com/Compare/89..._vs_Intel_Mobile_Celeron_Dual-Core_1007U.html

This site doesn't do overall ratings like the other one, just shows you
graphs of individual features in comparison to each other, and let's you
decide what's a more important feature.

Yousuf Khan

 

[Yousuf Khan]

I googled a bunch, as my sibling was on the phone at the store
trying to make a decision.

It seems that two cores is a huge improvement over one core; but,
quad core isn't such a big improvement, mainly due to the fact
that software has to be written for the quad core.

That's actually a very short-sighted viewpoint. I know this viewpoint
has been around for a while, but it's totally bunk. They are mainly
looking at a number of individual programs and seeing how many of them
are written with multi-threading in mind. However, they all ignore one
extremely important piece of software that is known to be
multi-threaded: the operating system itself! Assuming your running
Windows, but this applies equally to other operating systems like Linux,
Android, Mac OS X, etc., the OS is always multi-threaded.

So even if you don't have any piece of software that you can identify as
multi-threaded, you know that it doesn't matter, because your OS is
multithreaded. Think about how you use a typical PC, even if you're just
using one software at a time, there are multiple pieces of software in
the background that you're not even aware of that are also running.
Things like virus scanners, the network is constantly sending and
receiving data which requires background processes to monitor, etc.

On the other hand, it seems that CPU speed almost always helps
(assuming equal bus speeds).

So, it seems, the REAL question is whether the SOFTWARE is written
to take advantage of the quad core or just the dual core.

So, now it's a software question:

Q: How can we tell if software can take advantage of dual/quad cores?

Software is typical Windows stuff for a typical desktop user.

Well, another thing to consider is that even if my description of the
multi-threaded nature of the Windows OS didn't convince you, there are
actually a lot of multithreaded software that runs under Windows around.
Not even specialized esoteric software, some very common pieces of
software, like your web browser. All modern web browsers are
multi-threaded, from Internet Explorer, to Firefox, to Chrome, etc.! So
you can make use of multithreading on a daily basis.

Yousuf Khan

 

[JJ]

How would you compare dual-core 1.5GHz versus quad-core 1.0GHz for performance?
(Windows 8, touchscreen)

My sister asked me whether it's better to get the ASUS X200CA-HCL12050 or the
Dell Inspiron i3135-3750slv inexpensive $300 laptops.

Looking at the specs, the main difference seems to be:
Asus: dual core, 1.5GHz
Dell: quad core, 1.0GHz

Which helps in performance more, additional cores? Or faster speeds?
Note: I realize both help - but which is MORE important, cores or speeds?

The number of cores is more important, considering that Windows 8 is used.
It'll affect then responsiveness when both the system and user applications
are busy.

Dual core isn't exactly enough since the system sometimes performs more than
one tasks simultaneously (more than 10 at startup), and more user
applications nowadays are multi-threaded. Although the dual core one is
faster, it'll be exerted to handle more than two tasks simultaneously. It
will become hot easily.

 

[Darklight]

How would you compare dual-core 1.5GHz versus quad-core 1.0GHz for
performance? (Windows 8, touchscreen)

My sister asked me whether it's better to get the ASUS X200CA-HCL12050 or
the Dell Inspiron i3135-3750slv inexpensive $300 laptops.

Looking at the specs, the main difference seems to be:
Asus: dual core, 1.5GHz
Dell: quad core, 1.0GHz

Which helps in performance more, additional cores? Or faster speeds?
Note: I realize both help - but which is MORE important, cores or speeds?

The first question is what are your sister's needs? That should be the
deciding factor!

Does she intend to run multiple apps at once or run one app at a time.

So which processor will run multiple apps better?
Which processor will run a single app better?

 

[Ed Cryer]

That's actually a very short-sighted viewpoint. I know this viewpoint
has been around for a while, but it's totally bunk. They are mainly
looking at a number of individual programs and seeing how many of them
are written with multi-threading in mind. However, they all ignore one
extremely important piece of software that is known to be
multi-threaded: the operating system itself! Assuming your running
Windows, but this applies equally to other operating systems like Linux,
Android, Mac OS X, etc., the OS is always multi-threaded.

So even if you don't have any piece of software that you can identify as
multi-threaded, you know that it doesn't matter, because your OS is
multithreaded. Think about how you use a typical PC, even if you're just
using one software at a time, there are multiple pieces of software in
the background that you're not even aware of that are also running.
Things like virus scanners, the network is constantly sending and
receiving data which requires background processes to monitor, etc.


Well, another thing to consider is that even if my description of the
multi-threaded nature of the Windows OS didn't convince you, there are
actually a lot of multithreaded software that runs under Windows around.
Not even specialized esoteric software, some very common pieces of
software, like your web browser. All modern web browsers are
multi-threaded, from Internet Explorer, to Firefox, to Chrome, etc.! So
you can make use of multithreading on a daily basis.

Yousuf Khan

You can examine how well your CPU handles stuff with two aspects of Task
Manager.
"Threads" is a column heading that can be set. I've never ever seen a 1
in that column.
The other is Resource Monitor, CPU view. That gives a real-time view of
each cpu's changing load.

Run a large video-edit, and watch.

Ed

 

[Ken Blake]

What if you are using different applications concurrently on a regular
desktop? Assuming the individual applications are not optimized for multiple
cores, would the OS use an idle core or let the programs queue for the same
single core? Wouldn't that make a dual or quad core processor running at the
same speed faster on a regular desktop running a browser, e-mail, Quicken,
viewing a video on YouTube, etc.?

Even though many of us think we are running multiple applications
concurrently, in practice, we hardly ever are.

If you are running a browser, you are normally either reading what's
on the screen or using the internet connect to load a screen. The
browser itself is seldom doing anything.

If you are doing e-mail, you are usually either reading a message or
writing a message. The e-mail program itself is seldom doing anything.

If you are running Quicken, you are normally either looking at
something in Quicken or entering something into it. Quicken itself is
seldom doing anything.

Watching a video more often uses the CPU than most of those above, but
even there, you are mostly loading the next image from the internet.

At the moment, beside background programs, I have nine major programs
loaded, but the only one doing anything is my newsreader, Agent. And
even that isn't doing very much, since all it's doing is letting me
type. So my CPU usage is very low.

So "would that make a dual or quad core processor running at the same
speed faster on a regular desktop"? Yes, but normally only very
slightly faster.

 

[Rich]

Even though many of us think we are running multiple applications
concurrently, in practice, we hardly ever are.

If you are running a browser, you are normally either reading what's
on the screen or using the internet connect to load a screen. The
browser itself is seldom doing anything.

If you are doing e-mail, you are usually either reading a message or
writing a message. The e-mail program itself is seldom doing anything.

If you are running Quicken, you are normally either looking at
something in Quicken or entering something into it. Quicken itself is
seldom doing anything.

Watching a video more often uses the CPU than most of those above, but
even there, you are mostly loading the next image from the internet.

At the moment, beside background programs, I have nine major programs
loaded, but the only one doing anything is my newsreader, Agent. And
even that isn't doing very much, since all it's doing is letting me
type. So my CPU usage is very low.

So "would that make a dual or quad core processor running at the same
speed faster on a regular desktop"? Yes, but normally only very
slightly faster.

I see what you mean, makes sense. I guess if you're playing a game and
occasionally performing another task, then you may see a slightly better
improvement because the game can constantly use its own core exclusively
leaving other cores for intermittent mundane tasks.
Thanks for the input.

Rich

 

[David W. Hodgins]

What if you are using different applications concurrently on a regular
desktop? Assuming the individual applications are not optimized for multiple
cores, would the OS use an idle core or let the programs queue for the same
single core? Wouldn't that make a dual or quad core processor running at the
same speed faster on a regular desktop running a browser, e-mail, Quicken,
viewing a video on YouTube, etc.?
Also, that doesn't include the numerous background processes Windows is
constantly running. Every time I view the core usage graphs in task manager
there is constant activity in all of them. Is it that a multi-core processor
would not run multiple programs faster but just would not work as hard as a
processor with less cores?

If the applications are all cpu intensive, then more cores would matter
more than individual core speed. As most applications for regular
desktop users spend most of their time waiting for disk i/o, having
fewer, but faster cores, for those few apps that are cpu intensive,
is more important. I'm basing this on having seen the difference,
first hand.

You'll find switching from a regular hard drive, to an ssd drive, will
have a much greater impact on performance, since disk seek time no
longer applies.

Regards, Dave Hodgins.

 

[Ken Blake]

I see what you mean, makes sense. I guess if you're playing a game and
occasionally performing another task, then you may see a slightly better
improvement because the game can constantly use its own core exclusively
leaving other cores for intermittent mundane tasks.
Thanks for the input.

You're welcome. Glad to help.

 

[Paul]

If the applications are all cpu intensive, then more cores would matter
more than individual core speed. As most applications for regular
desktop users spend most of their time waiting for disk i/o, having
fewer, but faster cores, for those few apps that are cpu intensive,
is more important. I'm basing this on having seen the difference,
first hand.

You'll find switching from a regular hard drive, to an ssd drive, will
have a much greater impact on performance, since disk seek time no
longer applies.

Regards, Dave Hodgins.

The Passmark numbers aren't all that different for the two processors.
If you had multiple applications all running at the same time,
and they're processor intensive, then the Passmark is the better
performance estimator. And the quad core in this case, is 1.135x.
So it's barely faster in aggregate.

Summary table: Passmark SuperPI Power
(more is better) (less is better)

Dell Inspiron I3135-3750SLV A6-1450 quad 1653 46.8 seconds 8 watts
ASUS X200CA-HCL1205O Celeron dual core 1456 25.9 seconds 17 watts

If the A6-1450 had Turbo enabled, then it would be a better all-round
CPU, because the single threaded would be getting closer to the Celeron.
But apparently that doesn't happen in tablet designs, for
thermal reasons.

http://forums.anandtech.com/showthread.php?t=2319740

The setup here, turbo is apparently working. And in pictures
I can find of this 11.6 notebook, on the bottom there are no vents
or evidence of a cooling fan.

http://www.notebookcheck.net/Review-AMD-A6-1450-APU-Temash.92264.0.html

I don't really think you'll be running multiple applications
on that quad core, as it'll be too damn slow

Paul

 

[Danny D'Amico]

http://cpuboss.com/cpus/Intel-Celeron-1007U-vs-AMD-A6-1450
... it rates both processors exactly identically

That's amazing!

 

[Danny D'Amico]

Even though many of us think we are running multiple applications
concurrently, in practice, we hardly ever are.

I guess that's why the Windows task manager shows dozens of processes,
but, only a handful using the CPU at any one time.

Is that correct?

 

[Ken Blake]

I guess that's why the Windows task manager shows dozens of processes,
but, only a handful using the CPU at any one time.

Is that correct?

Yes.

 

[Paul]

That's amazing!

And not actually true.

If you go to the store and test both units, the
one with the Celeron in this case, will "feel faster".

This is why you do in-store testing, to see which
one of us is right

Paul