AMD QUAD 9850BE VS INTEL QUAD Q9650 FASTER? FLOPS!!

[gogothing7]

AUTHOR

Moonsub JIN KOREA OF REPUBLIC COMPUTER STORE
Kyungho LEE KOREA OF REPUBLIC LG COMPANY [/b]


IF GRAPH EXPANDED THEN MOUSE CLICK , OK???

FLOPS GRAPH




Computer performance which anyone finds out easyly to FLOPS


We will try for the world first time the measure x86 cpu performance with the flops
to be important performance index




nowadays, we are usually using CPU CLOCK, FSB, CACHE,MULTI CORE to evaluate the CPU performance.
however, it is too difficult for general people to understand the performance measurement process.




There is interesting situation to debate yet Intel Pentium4 3.0GHZ faster
than the AMD Brisbane 4200+(2.2GHZ)


As you can see upside's situation, I guess the CLOCK markting of cpu companies is one of the reason of misunderstanding.


in other words, almost everyone think CPU performance is same as CPU Clock.


The fixed idea to be the performance of the computer came into being a cpu clock at the consciousness of people due to this.




however, nowadays a multicore is going to be a general trend of computer industries.


So I think there is some problem to evaluate computer speed using past criterion.



On this point of view, I suggest CPU performance measurement method using genuine FLOPS(FPU)


this method could be evaluate CPU speed or entire Platform performance or objective performance




To evaluate to the flops for the X86 CPU to genuine is the text
to be produced to the beginning of the world


Therefore it cost the much a time and agony and effort in gathering the data
and concentrated on finding a objective performance of flops(FPU)


The writer presents to the wolrd company official
at the performance of the x86 cpu general computer let's introduce genuine flops numerical value and
this presents also we are the criteria of new world standard

IF GRAPH EXPANDED THEN MOUSE CLICK , OK???

FLOPS LIST





then How can we evaluate the FLOPS performance?


Linpack,Dhrystone,Whetstone and etc was written in fpu or alu measurement method at the past but
FPU measurement numerical value rose like the superpi to a compiler improvement at these days

As a result SPEC,AIM,Khornerstone,ETC have been written in
FPU or ALU measurement method in modern stage.




the FLOPS performance is the floating point operation speed and i think it is the fundermental parameter of computer speed.


Also, FPU or FLOPS a performance measurement method is used obviously at the reference
point to measure the performance of the million dollar supercomputer normally
and and FPU has important meaning to measure the performance of modern stage computer


if you see the table 1, someone could be confused


An intel's superpi(benchmark fpu) is far faster than AMD but
AMD's real fpu perfomance is far faster than intel


It is due to the cheating as the intel makes the core2duo conroe cpu currently
at the superpi program with the compiler.

in other words The compiler of the superpi is intentively optimized to INTEL conroe CPU




Linpack,Dhrystone,Whetstone and etc had used in FPU or ALU measurement method at the past



however, thesedays FPU measurement numerical value has been risen using a compiler
improvement like as superpi.



As a result SPEC,AIM,Khornerstone,ETC have been used in FPU or ALU measurement method in modern stage.


FPU performance important reason


FPU performance makes big influence at Online game,CD package game , rendering ,
Movie working, Encoding , Graphic works,3D rendering, 3D WORKS, UCC contents
works,Photoworks, 3D MAX, Transcoding which the generally users worked


It is also because FPU performance makes big influence at most task of Movie CG working,
Rendering Works,Climate Change , Encoding , 3D Simulation ,Sea Research,Earth Science,
Astrophysics ,Particle Physics ,Plasma Physics ,Nano Science ,Materia Technology ,
Biotechnology ,Protein kinetics ,Molecule dynamics ,Thermodynamics ,
Energy control ,Medical science,Architecture Simulation ,Numerical Analysis ,
Mathematics works , Data Analysis industry whole

 

[Dopekitten]

I actually wouldn't be surprised if the 9850BE outperformed the QX9650 in FLOPS, because the phenoms are optimized for servers, i.e. they are very good at quad-core optimized operations where FLOPS is what matters.

However, i think is this idea is crap because it doesn't matter to me how much FLOPS it has, it matters how well it actually performs. the Phenom could kick the shit outa my Q9450, but i still wouldn't buy it, because in the real world it doesn't perform.

Thus this test is fud.

Perhaps its AMD trying to get an advantage over intel because bigger numbers = better real world performance right?

 

[myocardia]

Originally posted by: Dopekitten
Thus this test is fud.

You took the words right out of my mouth, dopekitten.

 

[zach0624]

Dopekitten seems to have summed it up(I don't know how flops relate to servers so I can't comment on that) The 9850BE is not faster than the Q9650 in most apps and phenoms only perform well in server situations, also are these numbers legit(black man avatar, one post)? I normally don't see FLOPS benches so I don't know.

 

[OCGuy]

Wheres the free iPod?

Anyway, no.

 

[jones377]

Since when is a 9850BE+ an 8-core CPU ?? You can't use 2 of them either, you'd need an Opteron 2xxx. Anyway great first post, I wonder if you'll grace us with a second or is your trolling here complete?

 

[Keysplayr]

OP lays everything out like bullet points. Looks corporate to me.

But besides that, it's good to see we have a bunch of smart folks in here. Kudos.

 

[Idontcare]

See this is the kind of crap that AMD would be served well by having an appointed rep roaming the forums making sure it gets taken down as soon as it is discovered.

Does not help their cause to have Korean spam/viral-campaign plastered across the forums.

The gating issue with any FLOP benchmark (or any benchmark, when it comes down to it) is the specific mix of instructions that are exercised in the benchmark as not all instructions have the same latency to completion.

Here is an example of a real-world attempt to dissect the performance aspects of Phenom's architecture: http://www.digit-life.com/arti...3/cpu/rmma-phenom.html

 

[CTho9305]

FLOPS are irrelevant if you can't take advantage of the throughput with real-world code. A modern GPU is >500 GFLOPS - according to your testing methodology, we should start building 386-based systems with PCI-E slots and just run everything on the GPU. For highly-optimized kernels used in HPC, you may be able to get fairly close to the theoretical FLOPS throughput, but the vast majority of software doesn't look anything like that (partly because some tasks just don't work that way, and partly because almost nobody has the time to optimize their code enough to truly maximize performance on each different microarchitecture).

Some of the numbers are definitely wrong too - a 1.2GHz Athlon is capable of substantially more than 0.7 GFLOPS - the Athlon could do 1 FMUL and 1 FADD per cycle, so you should be able to get 2.4 GFLOPS out of it. All P4 processors had SSE, so they should also have been able to effectively do multiple FP ops per cycle. Something is seriously wrong with your benchmark code.

Personally I don't care how many GFLOPS my PC can do so long as the things I use are fast (and much of what I do is integer code, e.g. file compression & decompression, compiling code, etc).

I'm not speaking for any companies.

edit:

We will try for the world first time the measure x86 cpu performance with the flops
to be important performance index




nowadays, we are usually using CPU CLOCK, FSB, CACHE,MULTI CORE to evaluate the CPU performance.
however, it is too difficult for general people to understand the performance measurement process.

Maybe we should switch to using the number of pins on the CPU. It'd be really easy for the general public to understand that one.

 

[AmberClad]

OP has been spamming the same stuff over at the German THG forums, and another website...

 

[jones377]

Originally posted by: AmberClad
OP has been spamming the same stuff over at the German THG forums, and another website...

this thread should be locked then...

 

[Denithor]

I just wasted five minutes of my life on this thread?

I want a refund!

:beer:

 

[nyker96]

i don't mind him presenting some benchmarks here but it ain't enough for me to buy amd right now since i don't run server load, maybe for companies it would matter.

 

[aigomorla]

lets all go out and buy phenoms.

oh wait i did that... and it sucked hardcore!

 

[BLaber]

Run from here lol , I see all intel zombies trying to eat gogothing7 alive

 

[myocardia]

Originally posted by: Idontcare
See this is the kind of crap that AMD would be served well by having an appointed rep roaming the forums making sure it gets taken down as soon as it is discovered.

I thought this was their rep. Like keysplayr said, "Looks corporate to me." Hey, maybe Henry Ruin just doesn't speak very good Engrish?

 

[bradley]

I rarely see even enthusiasts using FLOPs to measure computer performance. That's language mostly reserved for marketers and promoters. Although, almost every hardware site has its own cadre of participants buying based on SuperPi. And that's an even less relevant and meaningful barometric. Unless you feel determining the amount of a CPU's cache without looking at specs is important.

 

[aigomorla]

uhh i7

Thats all i need to say.

 

[OCGuy]

If AMD is going to do some sort of viral/spam PR campaign, I would hope they could do better than "IF GRAPH EXPANDED THEN MOUSE CLICK , OK???" *sigh*

 

[Idontcare]

Originally posted by: Ocguy31
If AMD is going to do some sort of viral/spam PR campaign, I would hope they could do better than "IF GRAPH EXPANDED THEN MOUSE CLICK , OK???" *sigh*

Well I 100% doubt this is an AMD affiliated campaign...but if it were then you would certainly hope it was being done in a fiscally restrictive manner and in such a case one might conclude they got what they paid for.

 

[magreen]

LOL. Great thread, OP. Thanks for the post! Good for quite a few laughs.

Reminds me of these cheap little Chinese toys I got my 3-year-old son. They were plastic lookalikes of food -- little carrots, vegetables, the like.

After I bought it, I read the box. One of the slogans was "Great food. Pleasure every orifice." :roll:

 

[Idontcare]

Originally posted by: magreen
LOL. Great thread, OP. Thanks for the post! Good for quite a few laughs.

Reminds me of these cheap little Chinese toys I got my 3-year-old son. They were plastic lookalikes of food -- little carrots, vegetables, the like.

After I bought it, I read the box. One of the slogans was "Great food. Pleasure every orifice." :roll:

Which begs the question...did you follow the directions on the box?

And what was your consumer review - satisfied or dissatisfied with the product's performance?

 

[piesquared]

Really, I don't think there is any greater testimony to this having real merit in the real world than the quick response from the viral marketers and the "Intel retail edge" crowd, attempting to make these numbers insignificant. The numbers speak for themselves, nothing more, nothing less. It is the Intel camp turning this into something it isn't. To those I say relax, go take your SuperPi for a few laps around the track to make yourselves feel better. Those that are interested in FLOPS, which by the way DOES have real world implications, may find this interesting.

On a side note, I wonder where companies like SiSoft earn revenue? From all the benchmarkers on the web that run these programs who pony up barrels full of cash? Or maybe from the ten or so website's that use them in their reviews? Maybe it's part of the professional's criteria in determining which processor best suits their needs? I don't think it's open source? Obviously the insignificant amount of revenue generated from such benchmarks as Everest, SiSandra, and the like are not enough to warrant development of complex benchmarking programs. It doesn't take a genius to figure that out. Hint: The PCMark controversy a month or so back should be a pretty good clue.
So, the OP's compiler statement holds a fair bit of merit. When programs that do the same task vary so wildly in performance between Intel and AMD, it is irresponsible not to take notice.

 

[OCGuy]

Originally posted by: piesquared
Really, I don't think there is any greater testimony to this having real merit in the real world than the quick response from the viral marketers and the "Intel retail edge" crowd, attempting to make these numbers insignificant. The numbers speak for themselves, nothing more, nothing less. It is the Intel camp turning this into something it isn't. To those I say relax, go take your SuperPi for a few laps around the track to make yourselves feel better. Those that are interested in FLOPS, which by the way DOES have real world implications, may find this interesting.

On a side note, I wonder where companies like SiSoft earn revenue? From all the benchmarkers on the web that run these programs who pony up barrels full of cash? Or maybe from the ten or so website's that use them in their reviews? Maybe it's part of the professional's criteria in determining which processor best suits their needs? I don't think it's open source? Obviously the insignificant amount of revenue generated from such benchmarks as Everest, SiSandra, and the like are not enough to warrant development of complex benchmarking programs. It doesn't take a genius to figure that out. Hint: The PCMark controversy a month or so back should be a pretty good clue.
So, the OP's compiler statement holds a fair bit of merit. When programs that do the same task vary so wildly in performance between Intel and AMD, it is irresponsible not to take notice.

Another low post-counter defending this spam thread....are we being had here?

 

[piesquared]

Originally posted by: Ocguy31


Another low post-counter defending this spam thread....are we being had here?

Apparantely you've been "had here" for quite some time.