For those who ask the question Which is best AMD or INTEL

[KH85]

this is all i have to say on, o yeah one more thing. PLEASE dont ask that question again.....

 

[DAPUNISHER]

In before the lock

 

[cow123]

:Q

 

[Ticktanium2038]

Lets also run some benchmark utilities without 3DNow, SSE, and SSE2 code because if programmers arent using the specialized code, which is most programs out there. This will show what we can expect from real world applications between the two CPU's.

Lets quit comparing fully optimized code for a paticular CPU family.

With the current benchmarking utilities filled with SSE and SSE2 code is like saying NVIDIA optimised drivers which detect the benchmarking programs are fully acceptible. Despite the removed detail levels.

You know intel is going to win on the SSE and SSE2 fronts so why are benchmarking programs gospel when real world application tell a different story? Because not all programs use SSE and SSE2.

 

[Steven the Leech]

nef of that!

 

[Richdog]

HAHAHAH pic duly saved for future use.

 

[robcy]

I think that the picture is offensive, and in poor taste. I saved it just as a refrence, and will show it to all of my friends so that they understand the dispicable filth that floats around the net.

p.s. sorry the wifey was over my shoulder.

 

[KH85]

Originally posted by: Richdog
HAHAHAH pic duly saved for future use.

LOL i liked it to, i just had to share it with you lot on here couldnt resist the urge

Hope you liked it

KHGamez

 

[Ray5150Ray]

WHAAAAAAAAAA!!!!! I love it!!!!

 

[OddTSi]

Originally posted by: Ticktanium2038
Lets also run some benchmark utilities without 3DNow, SSE, and SSE2 code because if programmers arent using the specialized code, which is most programs out there. This will show what we can expect from real world applications between the two CPU's.

Lets quit comparing fully optimized code for a paticular CPU family.

With the current benchmarking utilities filled with SSE and SSE2 code is like saying NVIDIA optimised drivers which detect the benchmarking programs are fully acceptible. Despite the removed detail levels.

You know intel is going to win on the SSE and SSE2 fronts so why are benchmarking programs gospel when real world application tell a different story? Because not all programs use SSE and SSE2.

Are you kidding? Obviously Intel's compiler has SSE(2) support, Microsoft's Visual Studios also supports it and it's one of the most widely used compilers in the professional Windows arena, Borland's compilers support it, even GCC, one of the most widely used open source compilers, has SSE(2) support. I'm not even going to bother researching the multitude of other compilers (but I'm pretty sure they all have SSE support) because the 4 that I just listed definately encompass the majority of programmers out there.

 

[orion7144]

Dam AMD trying to scr$w more money out of Intel....

 

[Corporate Thug]

in before lock

 

[DivideBYZero]

This is in your future

 

[MDE]

Who would lock this gem of a thread?

 

[Ticktanium2038]

No not kidding the majority of programs out there have not been compiled using SSE and especially SSE2 code. Granted a few of the larger camps compile with it. (Photoshop, Lightwave, and a few others.) And it makes sense to because I would like to see code compiled for a paticular processor in mind to use its full performance abilities.

But every benchmarking program there is loaded with SSE and SSE2 code. Making it a skewed representation of performance. Granted if every program was compiled with it there would be some improvements but not all programs are.

This is why benchmarks programs show that there is no possible way in the world an AMD chip can perform nearly as well as an Intel chip.

But when you run program like Unreal and many others you see that they somehow manage to beat an Intel chip. This should not be possible according to the benchmark programs which show that Intel chips are vastly faster than AMD.

No what these benchmark programs are showing you is if you compile your programs for SSE and SSE2 loaded code it could potentially be faster on an Intel based machine.

So benchmarking programs are skewed in favor of Intel based CPU's and do not reflect true world processor performance but merely its capability when compiled for a paticular CPU.

This is identical to NVIDIA optimising its drivers for paticular games and benchmarking utilities. The current benchmarking utilities are compiled in a way to favor Intel CPU's. So if you support that then you support what NVIDIA is doing to benchmarking to trick you into thinking they are making a superior product.