Intel Vs. Amd

[clocker]

Originally posted by 3RA1N1AC@8 September 2003 - 23:42
well, the effect of a CPU's brand on games is relatively intangible. upgrading your CPU may or may not contribute to an increase in frames-per-second, it may or may not make the game run more smoothly. i would not expect a person to be able to identify Intel or AMD (like a Coke vs Pepsi taste test) if they were given a chance to play the same game on two unmarked computers.

in contrast, you upgrade a 3D video card, and you get more frames per second, you can turn up the detail levels, you may be able to enable more special effects that your older card couldn't produce, etc. some people can pretty easily tell which picture is produced by which card, because of certain quirks or characteristics of each brand's display methods. the characteristics and features of the video card are obvious.

but CPUs? i would be incredibly surprised if someone could identify the two brands in a "blind" test, to the point where a preference is justified on performance or stability alone. their brand-exclusive features are entirely speed-related (aside from throttling/idling behavior). both brands perform well, both brands make 100% stable CPUs. CPUs either work or don't work, period-- improper cooling, shoddy motherboards & RAM, etc are completely separate issues. there just is no obvious difference to identify the CPU brand, if you haven't already been told which one you're using. "i play Quake on AMD because AMD makes Quake look better, sound better, feel better." anyone who claims such a thing (about Intel or AMD) is just fooling themselves.

So would that make Lamsey's assertion that "it all comes down to money"* valid then?


* paraphrase

[ilw]

Of course when choosing a chip the money is all important. If youve got tons of money to spare and are just looking to buy the very best and latest system then its all about the top spec chip, otherwise you will want the best chip you can afford (without sacrificing the quality of the other components). Also the point about discussing which chip is faster is not only looking at what could use 100% cpu now, but also how long the chip will be able to cope before its obsolete/can't play hte games being released.

[Xanex]

Originally posted by ilw@8 September 2003 - 09:31

Depends solely on the cooling, there is no figure for this. What u want is the rate of heat output (or consumption) in watts, which i dunno offhand

Thankyou for that really nice post I enjoyed reading it =)

To clarify: Intel P4 3.06HT with the OEM cooling.

Basically a prebuild that you would buy from any major suckage retail shop eg PC world etc

Thanks

[LaVaNaNo]

AMD ownz all

[Cl1mh4224rd]

Assumtions: "Intel" = P4 3.06GHz; "AMD" = AthlonXP 3000+ (2.1GHz)

Cost (from NewEgg.com)
Intel: $379.00
AMD: $269.00

AMD is cheaper. (Intel - 0; AMD - 1)

Intel = $0.12/MHz
AMD = $0.13/MHz

Intel squeezes by AMD at $0.01 cheaper per MHz. (Intel - 1; AMD - 1)

(Following info gather from Tom's Hardware Guide: "Benchmark Marathon: 65 CPUs from 100 MHz to 3066 MHz")

OpenGL Benchmarks (1024x768 / 32 bit / 85Hz)
Intel: 351.3 FPS
AMD: 311.2 FPS

Intel gives you more FPS. (Intel - 2; AMD - 1)

Intel = $0.92/FPS;
AMD = $0.86/FPS;

AMD offers better price/FPS than Intel. (Intel - 2; AMD - 2)

Intel = 0.11 FPS/MHz
AMD = 0.14 FPS/MHz

AMD offers better FPS per MHz than Intel. (Intel - 2; AMD - 3)

Direct3D Benchmarks (1024x768 / 32 bit / 85Hz)
Intel: 15878
AMD: 15655

Intel squeaks by with a higher score. (Intel - 3; AMD - 3)

Intel = 41.89 pts/$
AMD = 58.20 pts/$

AMD offers more points per dollar. (Intel - 3; AMD - 4)

Intel = 5.07 pts/MHz
AMD = 7.28 pts/MHz

AMD also offers more points per MHz. (Intel - 3; AMD - 5)

Audio/Video Benchmarks (1178 MB Wave / VBR)
Intel: 72 seconds
AMD: 110 seconds

Intel beats AMD when it comes to encoding. (Intel - 4; AMD - 5)

Intel = 0.043 (MB/sec)/dollar
AMD = 0.039 (MB/sec)/dollar

Intel offers slightly faster encoding per dollar (Intel - 5; AMD - 5)

Intel = 0.0052 (MB/sec)/MHz
AMD = 0.0050 (MB/sec)/MHz

Intel also offers slighltly faster encoding per MHz. (Intel - 6; AMD - 5)

Synthetic Benchmarks - CPU (PC Mark 2002) (1280x1024 / 32 bit / 85Hz)
Intel: 7571
AMD: 6646

Intel gives you a higher score. (Intel - 7; AMD - 5)

Intel = 19.98 pts/$
AMD = 24.71 pts/$

AMD offers more points per dollar, however. (Intel - 7; AMD - 6)

Intel = 2.42 pts/MHz
AMD = 3.09 pts/MHz

AMD also offers more pts per MHz. (Intel - 7; AMD - 7)

Synthetic Benchmarks - Memory (PC Mark 2002)
Intel: 8120
AMD: 6853

Again, Intel gives you a higher score. (Intel - 8; AMD - 7)

Intel = 21.42 pts/$
AMD = 21.76 pts/$

Wow, a close one, but again, AMD offers more points per dollar. (Intel - 8; AMD - 8)

Intel = 2.59 pts/MHz
AMD = 3.19 pts/MHz

AMD offers more points per MHz, also. (Intel - 8; AMD - 9)

Applications Benchmarks (File Compression: WinRAR 3.1) (178 MB)
Intel: 53 seconds
AMD: 69 seconds

Intel beats AMD in file compression. (Intel - 9; AMD - 9)

Intel = 0.0087 (MB/sec)/dollar
AMD = 0.0096 (MB/sec)/dollar

AMD offers faster file compression per dollar (Intel - 9; AMD - 10)

Intel = 0.0011 (MB/sec)/MHz
AMD = 0.0012 (MB/sec)/MHz

AMD also offers slightly faster file compression per MHz. (Intel - 9; AMD - 11)

Conclusion:

Overall, AMD just barely beats out Intel. However, if you like big numbers, Intel is the way to go (Intel - 6; AMD - 1). If you're more concerned about bang-for-your-buck, AMD is a must (Intel - 3; AMD - 10).

There you have it...

[Lamsey]

Oh great, another person who's obessed with clock rates.


That's a very nice post and well thought out, but you really need to lose the clock rate fascination.

[Cl1mh4224rd]

It's not just about clock speed. It's the performance per advertised clock speed, which AMD clearly wins at. If it was about clock speed, Intel would win hands down.

So, how do you suggest comparing these processors without getting into their clock speeds? I don't see anyone else actually trying figure anything out. Just a lot of meaningless "you're wrong," "no, you're wrong," "you're all wrong" bullshit...

[Amarjit]

Hey, I'm not saying that Intel's perfect. In term of Front Side Bus Frequency, they are beaten, by IBM, with their G5 processer which has a 1GHz FSB.

[Lamsey]

Originally posted by Cl1mh4224rd@21 September 2003 - 11:43
So, how do you suggest comparing these processors without getting into their clock speeds? I don't see anyone else actually trying figure anything out. Just a lot of meaningless "you're wrong," "no, you're wrong," "you're all wrong" bullshit...

AMD have a rather useful thing called a model rating. If you actually read my posts above on the subject, you'd see why clock speed is often irrelevant.

[Cl1mh4224rd]

Originally posted by Lamsey@21 September 2003 - 14:46
AMD have a rather useful thing called a model rating. If you actually read my posts above on the subject, you'd see why clock speed is often irrelevant.

I prefer AMD myself, but the model rating is useless. I know you know it's just the comparative clock speed of what an earlier AMD chip design would've been capable of. If Intel also used a model rating like AMD, I'm sure the P4 3.06GHz would be the "Intel 4000+", which means clock speeds are just obfuscated even more, and no other choice but to look at the actual clock speed of the chip.

Seems as though you're trying to say that comparing CPUs is like arguing the existence of God. So, I ask you once again... how do you propose showing comparative CPU performance, without involving clock speed?

I'm not trying to say, "OMG!!11 INteL sPEed > AMd SpeED! iNTEl Win!!111!!" and that a 2GHz Intel chip = a 2GHz AMD chip. If I had done that, then you say I was obsessed with clock speed. Look over my numbers and you'll realize that this about how much performance each chip squeezes out. Granted, FSB is a factor, too, but I wasn't sure how to add that into the comparison (although Intel's 533MHz FSB may have helped in the breakdown, it still wasn't enough to beat out AMD, with only a 400MHz FSB, in most areas).

If you can figure out a way to include FSB, L1 cache, L2 chache, and any other aspect of each chip into the comparisions, I'll gladly redo everything.