Athlon 64 vs Pentium 4

[Moderately Confused]

Hi,

I had a look at those benchmarks and it seems as soon as you put the
resolution up the Athlon 64s drop nearly 20fps while the Intel ones seem
to drop a much smaller amount.

This seems to indicate that the Athlon 64s don't perform very well when
you put them under any real pressure.

-Steve

Bingo, I said that in another post. He ended up disputing something else I
said...

MC

 

[Chuck]

Not the facts, more of his biased marketing spiel towards

consumers sending AMD a boatload of cash for their high-end part
at the moment. Perhaps you have not witnessed or realized his
promotion of AMD, and only AMD, hundreds of times if not more
only considering the "Jeffrey Karp" handle alone?

JK ~ Jeffrey Karp

http://groups.google.com/groups?q=Jeffrey+Karp+buy+AMD

Hehe. I was going to say something about it. I see JK post in several
groups which I read and post in. I don't think I've seen a post yet where
he fails to place a link to a CPU benchmark site and tout AMD.

 

[JK]

Stop comparing apples to oranges. You can't compare a 64 bit processor to a
32 bit processor.

Of course you can, provided you are running the same 32 bit OS and 32 bit
software in each. Of course the Athlon 64 will probably be much faster
in its 64 bit mode with 64 bit software and a 64 bit OS than with 32 bit
software and a 32 bit OS. It is very reasonable to compare a P4 3ghz
Prescott to an Athlon 64 3200+, since they are very close in price.

 

[John R Weiss]

Stop comparing apples to oranges. You can't compare a 64 bit processor to a
32 bit processor. It's like comparing the gas mileage in an electric hybrid
car and a regular combustion engine.

Actually, it is fair to compare the Athlon 64 to the Pentium 4 when both are
marketed to the Win32 market as the 'latest' in high-performance processors
for home use.

Also, the OP cited 2 similarly priced variants.

Also, why can't you compare gas mileage in the Civic gas to the Civic
hybrid, or any other comparable car?

 

[~misfit~]

Overclocking is not recommended if you want a stable system.
Overclocking also tends to reduce the life of the processor, and
might require expensive water cooling to overclock by large margin.

You mean I might have to throw these two Celeron 600's that have been
faithfully running at 900Mhz for over four years away soon?

Even if you're right, it don't bother me none. If I was running them at 600
I would have got rid of them ages ago anyway as they'd be too slow for what
I use them for now.

BTW, I'm using the original Intel HSF's on them.

 

[~misfit~]

It is an argument for not driving a car above the speed limit.

Your analogy is flawed. OCing a CPU, if being compared to a car, isn't like
breaking the speed limit, it's like hotting it up. You know, big bore
exhaust, high compression heads, Nox (NO2) kit, increase bore/stroke,
turbocharging, balance the crankshaft/pistons/con rods. That sort of thing.
Getting more performance out of it.

The analogy of breaking the speed limit is just plain wrong, and intended to
support your luddite views. Create a climate of fear. Next you're going to
say that people who overclock their CPUs are known to have a tendancy to
write virus' and trojans and look at p0rn.

 

[Fitz]

My CPU is only mildly o/c'ed, I only look at soft porn, and I don't have any
trojans.

Fitz

sorry- couldn't resist

 

[kony]

Bingo, I said that in another post. He ended up disputing something else I
said...

AS much as I hate to agree with JK, it is true that this is an
indication of video card bottlenecks, not CPU performance.

 

[kony]

Stop comparing apples to oranges. You can't compare a 64 bit processor to a
32 bit processor. It's like comparing the gas mileage in an electric hybrid
car and a regular combustion engine. When Intel comes out with their own 64
bit processor, than you can start with the whole benchmark thing.

Reread the linked article, 64 bit is irrelevant as it wasn't
doing anything 64 bit. It is true that eventually Intel will
also have higher performance CPUs, but then so will AMD... world
keeps spinning...

 

[Moderately Confused]

to

Actually, it is fair to compare the Athlon 64 to the Pentium 4 when both are
marketed to the Win32 market as the 'latest' in high-performance processors
for home use.

Also, the OP cited 2 similarly priced variants.

Also, why can't you compare gas mileage in the Civic gas to the Civic
hybrid, or any other comparable car?

Ok, maybe that was a bad example, but why compare something that shouldn't
be compared that way? Of course 64 bit is going to be better than 32 bit.
Maybe it's like comparing a screwdriver to a cordless drill?

MC

 

[Moderately Confused]

Reread the linked article, 64 bit is irrelevant as it wasn't
doing anything 64 bit. It is true that eventually Intel will
also have higher performance CPUs, but then so will AMD... world
keeps spinning...

Still, the 64 bit processor will out perform the 32 bit one, although it
will be a minor performance increase. My problem isn't with AMD, it's the
"proof" of JK's claims. Sure, AMD 64 might be better in Business Winstone,
but it's only one piece of software. Whoop-de-doo, it out performs Intel in
Doom3, which I heard sucks anyway. All of his "arguments" are based on two
links.

MC

 

[JK]

Ok, maybe that was a bad example, but why compare something that shouldn't
be compared that way? Of course 64 bit is going to be better than 32 bit.

The benchmarks in the article are done with 32 bit software and a 32 bit OS.
It isn't necessarily the case that a cpu that has a 64 bit mode will outperform
comparably priced 32 bit processors running 32 bit software. For the Athlon 64
it happens to be true though.

Maybe it's like comparing a screwdriver to a cordless drill?

Not quite. Think of the Athlon 64 chips as being like a car that is
a convertible that can be used with the top up or the top down.
The Athlon 64 has two modes, 32 bit or 64 bit. The 64 bit mode
is with the use of a 64 bit OS, while the 32 bit mode is with a
32 bit OS. In the 32 bit mode, only 32 bit software can be run.
In the 64 bit mode, 64 bit or 32 bit software or both side
by side can be run.

 

[Stephen Gordon]

AS much as I hate to agree with JK, it is true that this is an
indication of video card bottlenecks, not CPU performance.

As I;ve already pointed out elsewhere in the thread in that case what's
the point of spending so much on a fast CPU if you can't possibly get a
GFX card to match it?

-Steve

 

[David Maynard]

As I;ve already pointed out elsewhere in the thread in that case what's
the point of spending so much on a fast CPU if you can't possibly get a
GFX card to match it?

-Steve

I suspect there may be uses for PCs other than playing games.

 

[JK]

Still, the 64 bit processor will out perform the 32 bit one, although it
will be a minor performance increase. My problem isn't with AMD, it's the
"proof" of JK's claims. Sure, AMD 64 might be better in Business Winstone,
but it's only one piece of software.

Business Winstone is not one piece of software. It is composed of several
applications.

http://www.veritest.com/benchmarks/bwinstone/s1wsapps.asp

Whoop-de-doo, it out performs Intel in
Doom3, which I heard sucks anyway.

Many other games as well.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2065&p=10

Take a look at this.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2065&p=12

Keep in mind that these 32 bit applications will probably run faster
using a 64 bit OS, and that 64 bit versions of software will probably
run much faster than that.

All of his "arguments" are based on two
links.

The is also Advanced Virus Protection (AVP) when an Athlon 64 chip
is used With Windows XP with Service Pack 2(SP2) installed.

http://www.technewsworld.com/story/athlon_64_amd_xp_sp2/35675.html

Another review.
http://www.aceshardware.com/read.jsp?id=65000304


More links.

http://www.xbitlabs.com/articles/cpu/display/athlon64-3800.html

 

[Dave C.]

be compared that way? Of course 64 bit is going to be better than 32 bit.

For what? -Dave

 

[Dave C.]

Your analogy is flawed. OCing a CPU, if being compared to a car, isn't like
breaking the speed limit, it's like hotting it up. You know, big bore
exhaust, high compression heads, Nox (NO2) kit, increase bore/stroke,
turbocharging, balance the crankshaft/pistons/con rods. That sort of

thing.

No. OCing a CPU is like taking a STOCK car and never running the engine
BELOW redline. You are doing something with the car that it was not
designed to do. And yes, it will be fast, until the engine and all other
mechanical components give out on you. THAT is what OCing a CPU is,
exactly. -Dave

 

[kony]

thing.

No. OCing a CPU is like taking a STOCK car and never running the engine
BELOW redline.

Who said anything about running it THAT far at the borderline?
Sure if you're a masochist you can destroy anything, but it would
have to either be a goal or be done quite recklessly, just like
anything else.

It is more similar to a german shipping over a Corvette for
driving on the autobahn, then finding there is a governor
restricting it to 80MPH, so they alter the artifical limiter to
reach it's full potential.

You are doing something with the car that it was not
designed to do.

And yes, it will be fast, until the engine and all other
mechanical components give out on you. THAT is what OCing a CPU is,
exactly. -Dave

Grand theory, but where are those stacks of dead CPUs?
There MUST be stacks and stacks of 'em, because quite a few
people o'c and have CPU that've ran that way for years. How many
years should we wait to see if the CPU died? In a previous post
I mentioned an example of Celeron 300 o'c to 450... those are
about 8 years old now, do we need to get 10-20 years out of a
Celeron 300? Possibly on a space station that would be
important, but back on the mother planet that Celeron 300 is not
going to die before the motherboard, power supply, video card,
etc, to the extent that odds are very high the rest of the system
will be dead before CPU died, so it was abandoned due to no
platform to run it. Running a celeron @ 450 can't be argued as a
significant cause of motherboard or power supply failure since it
wasn't as much of power or heat problem as it's predecessors
running at stock speed on same platform(s).

 

[David Maynard]

Who said anything about running it THAT far at the borderline?
Sure if you're a masochist you can destroy anything, but it would
have to either be a goal or be done quite recklessly, just like
anything else.

It is more similar to a german shipping over a Corvette for
driving on the autobahn, then finding there is a governor
restricting it to 80MPH, so they alter the artifical limiter to
reach it's full potential.




Grand theory, but where are those stacks of dead CPUs?
There MUST be stacks and stacks of 'em, because quite a few
people o'c and have CPU that've ran that way for years. How many
years should we wait to see if the CPU died? In a previous post
I mentioned an example of Celeron 300 o'c to 450... those are
about 8 years old now, do we need to get 10-20 years out of a
Celeron 300? Possibly on a space station that would be
important, but back on the mother planet that Celeron 300 is not
going to die before the motherboard, power supply, video card,
etc, to the extent that odds are very high the rest of the system
will be dead before CPU died, so it was abandoned due to no
platform to run it. Running a celeron @ 450 can't be argued as a
significant cause of motherboard or power supply failure since it
wasn't as much of power or heat problem as it's predecessors
running at stock speed on same platform(s).

Yes. My BP6, running dual 300As overclocked to 495, began life with NT4
server but is still going strong today as my 24/7 internet/LAN server
running Win2000.

 

[Noozer]

That is not true. The Athlon 64 3200+ will beat the P4 3ghz Prescott

Stop comparing apples to oranges. You can't compare a 64 bit processor to a
32 bit processor. It's like comparing the gas mileage in an electric hybrid
car and a regular combustion engine. When Intel comes out with their own 64
bit processor, than you can start with the whole benchmark thing.

A processor is a processor... If a 64bit CPU can beat a 32bit CPU and cost
the same as the 32bit then it's obviously better and the one to choose (as
long as the rest of the platform doesn't drive up the price - final cost is
the main issue)

When comparing anything, it's price to performance that counts, regardless
of technology. (Performance here also includes durability/quality as well).
The biggest/fastest/etc isn't always the winner if a cheaper solution will
still get the job done.