• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Lol. Brand-new G4s beaten by ancient PIII

M

Martin

Lifer
According to the Reg. They say the latest SPEC numbers show the G4s have dismall FP performance and decent integer performance.


I guess the truth comes out if you use something other than a few PS filters as the benchmark....
 
I dont think the G4 did all that bad in those benchmarks. Ill definitely be buying another Apple when I get the money. And hell, Im not buying a machine based on some artificial benchmarks.
 
Didn't the so called ancient p3 also wax some low end P4 chips?

Some people live by benchmarks and girly mags. Other people prefer the real thing.
 
SPEC, btw, is one of the highest regarded benchmark suites in the business. Every reputable manufacturer posts SPEC values for its systems.

But, to each their own.
 
SPEC CPU became bad in the eyes of the hardware community at around the time the P3 suddenly beat the Athlon. Some of us may remember the SPEC numbers AMD touted at the launch back in 99 when compared with the Xeon (yes, Xeon). A few compiler releases later the P3 caught up and actually surpassed the Athlon. And SPEC hasn't really regained it's reputation since. Not that professionals (engineers, architects etc) care though, it's still the best cross platform benchmark suite there is.
 


<< SPEC, btw, is one of the highest regarded benchmark suites in the business. Every reputable manufacturer posts SPEC values for its systems.

But, to each their own. >>



But how much do benchmarks really mean? If I am running that benchmark suite it would be great, but if Im doing actual work the benchmarks wont help all that much. The only benchmarks I think are worthwhile are benchmarks using actual situations I may see in the real world.
 
If you read about SPEC, they try to model certain applications as well as possible:

SpecInt is supposed to emulate work on the desktop (which is mostly integer based, except for some image filters etc.)
SpecFP is supposed to show how well scientific, floating point heavy, software runs.

It's as realistic as a benchmark can get, but yes, it is still a benchmark, so somewhat theoretical.

Andreasl:
Intel has the money (which it drags out of the nose of the consumers) to create very efficient compilers. This is why the P3 was able to surpass the Athlon, especially in the SpecFP, where some SSE commands were used for optimization. I think the latest SPECs for the AthlonXP restore the balance once more - especially and even when compiled with the Intel compiler. For your claim that hardware people do not like SPEC anymore I would like to see a link.

What SPEC does is benchmark the combo: Compiler+CPU+memory+motherboard, with the emphasis on the first 3.
It's really interesting if you read into it.
 


<< Lol. Brand-new G4s beaten by ancient PIII >>



not surprising... just witty..
we have G4's here.. they are most certainly overrated and overpriced.. I think they suck.
 
For example, from:
http://www.spec.org/osg/cpu2000/results/cint2000.html

An XP 1800+ gets 648 / 671 (base / peak value)
An XP 2000+ gets 697 / 724 (base / peak value)
A P4 1.7 GHz gets 574 / 591 (base /peak value)
A P4 2.0 GHz gets 648 / 664 (STILL below the Athlon XP with lower MHz)
A P4A 2.0 GHz finally gets 735 / 753


from:
http://www.spec.org/osg/cpu2000/results/cfp2000.html

An XP 2000+ gets 596 / 642
A P4 2Ghz gets 649 / 669

This is where I think the SSE2 optimization for floating point operation kicks in, and the advantage of using an Intel compiler.
Considering the MHz difference, the XP 2000+ still holds up.

 
Ill have to re-read it, but did the compilers for the G4 actually use the velocity engine thingy ptimizations?

EDIT: too many different things going around in my greyware 😛

EDIT: It looks like they didnt do any SSE/SSE2 opts either.
 
"The SPEC benchmarks disable a second CPU if present, and don't necessarily take advantage of the Altivec and Screaming Sindy dedicated FP instructions - that's down to the compiler."

"Considering the MHz difference, the XP 2000+ still holds up."

This has been beaten to death, clock speed means jack. If you want the fastest CPU today, Intel is it, regardless of how many additional GHz it takes to get there.
 
Nah the athlon will always be faster than the p4 because it has better IPC and thus efficiancy. When it racks up in clockspeed it will rape that p4 anyways so AMD will rule once again.
 


<< This has been beaten to death, clock speed means jack. If you want the fastest CPU today, Intel is it, regardless of how many additional GHz it takes to get there. >>



Yes, of course it has. And I disagree with you - if you want fastest CPU, it's neither Athlon nor Pentium, but an Alpha or a SunBlade...

They have similar integer power, but higher FPU power.

 


<<

<< This has been beaten to death, clock speed means jack. If you want the fastest CPU today, Intel is it, regardless of how many additional GHz it takes to get there. >>



Yes, of course it has. And I disagree with you - if you want fastest CPU, it's neither Athlon nor Pentium, but an Alpha or a SunBlade...

They have similar integer power, but higher FPU power. >>



You arent just talking CPU there, those are basically much better architectures all around.
 
I'm referring to consumer level CPU's that run software that a typical user would actually care about.

"Nah the athlon will always be faster than the p4 because it has better IPC and thus efficiancy."

Always in this case not including now, apparently. AMD will not capture the speed crown again until Clawhammer at the earliest. The phenomenal OC'ing results of the new Northwoods exhibits that Intel can raise clock speeds at will in response to any Athlon AMD can come up with, which hasn't been much in the past few months.
 


<< AMD will not capture the speed crown again until Clawhammer at the earliest. >>



I completely agree with that, especially as the P4 is going to ramp increasingly higher in clockspeed while AMD is having difficulties with their .13u process. Plus the P4 should see an IPC boost from the 533MHz FSB, whereas the Athlon isnt going to see any changed between now and Thoroughtbred and possibly not even then.

I strongly disagree with your comments that Intel has clearly gained the performance crown right at this moment however. Overclockability really isnt a factor, as 99.9% of the global market doesnt overclock... and at least right now the AthlonXP 2000+ is maintaining parity with the Northwood 2.2GHz according to the vast majority of the benchmarks.
And similarly then Xeon does not have a clear lead over the AthlonMP in the areas relative for workstation/server performance.


As for the G4's poor Spec performance, SPEC is a highly regarded benchmark yes. BUT it's as much if not more-so a compiler benchmark then it is a processor benchmark.
And in this case that point is especially valid... look at the compiler used for the G4 in the Spec2000 benchmarks. The compiler they utilized is often cited among the PPC community as producing extremely poorly performing binaries.

It wouldnt take much effort for me to find Spec2000 benchmarks that position the P4 as scoring even lower then that depending upon the compiler used.
The most powerful rpocessor in the world can look sickening without a decent compiler in Spec, this has often been AMD's biggest problem with their relatively lackluster Spec scores relative to their excellent real world performance.
 
You arent just talking CPU there, those are basically much better architectures all around.

Of course I am. But the same holds true if you compare a P4 with SDRAM or a P4 with DDR or RAMBUS...
A CPU without a platform is nothing. And actually, they do just have much more raw FPU power.
 


<< Andreasl:
Intel has the money (which it drags out of the nose of the consumers) to create very efficient compilers. This is why the P3 was able to surpass the Athlon, especially in the SpecFP, where some SSE commands were used for optimization. I think the latest SPECs for the AthlonXP restore the balance once more - especially and even when compiled with the Intel compiler. For your claim that hardware people do not like SPEC anymore I would like to see a link.

What SPEC does is benchmark the combo: Compiler+CPU+memory+motherboard, with the emphasis on the first 3.
It's really interesting if you read into it. >>



I never claimed otherwise. The Athlon also got a speed boost from the updated compilers. I was just trying to describe the reaction of the community at the time, which I found rather amusing. Instead of realising how much effect a compiler can have on performance the entire suite was discredited instead.

In a few weeks AMD will release the XP2100. It represents less than a 4% clock speed increase and even less real performance increase in benchmarks. Some people tend to think that is very important anyway. Yet a new compiler could increase the performance of the same benchmark (when recompiled) with 15% easily and sometimes much more (Sun Forte, hehe). Of course few apps have the luxury of being recompiled every so often.
 
Some people live by benchmarks and girly mags. Other people prefer the real thing.

HeeHee! I definitely like the real thing, oh, you mean the benchmarks.
 
compiler optimizations are huge when you're dealing with calculations that take a supercomputer days or weeks... thats why SPEC is important.
 
Regardless, I don't see how someone can justify spending the extra money on an apple of any sort. Whoopie, its got a dvd burner! And and LCD monitor! Too bad you can get more performace at a lower price if you go non-apple. I just dont see the point of buying one anymore.
 


<< Regardless, I don't see how someone can justify spending the extra money on an apple of any sort. Whoopie, its got a dvd burner! And and LCD monitor! Too bad you can get more performace at a lower price if you go non-apple. I just dont see the point of buying one anymore. >>



To get away from the legacy x86 crap. To use an OS that is not as bad as others. Variety. All of the wonderful applications that come with it.
 


<< . Plus the P4 should see an IPC boost from the 533MHz FSB >>



How do you figure that? All that will do is increase the memory bandwidth.

Example: We have a hypothetical 2Ghz CPU. It is unlocked.

If we run it @ 2x 1000FSB, or 1x 2000FSB, we are going to get the same IPC, but the 2000FSB will have double the memory bandwidth.

Please correct me if im wrong, but i dont see how more memory bandwidth will increase the IPC.
 


<< Regardless, I don't see how someone can justify spending the extra money on an apple of any sort. Whoopie, its got a dvd burner! And and LCD monitor! Too bad you can get more performace at a lower price if you go non-apple. I just dont see the point of buying one anymore. >>




Editing.


 
You must log in or register to reply here.

TRENDING THREADS

Back
Top