Aktív témák

• #1 e-biza őstag

  2003-03-02 00:43:13 #1

  Összes hozzászólása itt Válaszok az összes hozzászólására itt Válaszok erre a hozzászólásra

  Privát üzenet küldése

  

  e-biza

  őstag

  [L]http://www.overclockers.com/tips00302/[/L]
  
  Amit a cikk tartalmaz, hogy a P4 architektúra miért tud olyan magas GHZ-re kuszni, és hogy az AMD miért használja a számozást.
  
  Hard as it is to believe, people are still buying what AMD calls ''the MHz myth.''
  
  You might think AMD has been stretching it a little lately, but what they say is essentially true.
  
  To keep this really simple:
  
  Until the PIV came out, Intel and AMD chips essentially were the same MHz for MHz. That means they did roughly the same amount of work per clock cycle, so an AMD Athlon 1GHz CPU did about the same as an Intel PIII 1GHz CPU.
  
  This changed with the PIV.
  
  Generally speaking, when you design a CPU, you have to decide whether you want the CPU to do a lot per clock cycle, or a little.
  
  If you choose a lot, that means a MHz from that design is ''worth'' more than a MHz from a CPU designed to not do a lot per clock cycle. So a ''heavy work'' 1GHz CPU is actually ''faster'' than a ''light work'' CPU.
  
  On the other hand, you can't crank up the MHz on a ''heavy'' design as much as you can a ''light'' design.
  
  So the choice is more work, less MHz, or less work, more MHz.
  
  Intel decided the latter for the PIV. They decreased the amount of work a PIV does per clock cycle from the PIII/Athlon standard. This let them increase the maximum MHz to a level above that possible with a PIII/Athlon design.
  
  In short, they devalued the worth of a MHz so as to get more of them.
  
  That's why an Intel CPU can go over 3GHz while an AMD processor can only do 2.5GHz. At .13 micron, for the work an Athlon does per clock cycle, AMD isn't going to reach 3GHz. Neither could Intel with that kind of design.
  
  That's no lacking on AMD's part; Intel wouldn't be able to do any better than that with a PIII-type design, either. It's solely because a PIV MHz is worth less than an Athlon MHz that a PIV can go higher.
  
  Let me put it another way. Both the United States and Canada value their currency in dollars. The U.S. dollar happens to be worth more than the Canadian dollar.
  
  What Intel essentially did was stop paying in U.S. dollars and started paying in Canadian dollars. Now there's nothing wrong with Canadian dollars, it's good money. But only an idiot would think that a dollar is a dollar is a dollar, and so a Canadian dollar is worth the same as a U.S. dollar.
  
  So Intel is giving four Canadian dollars and AMD is giving you three U.S. dollars, and plenty of people think Intel is giving them more money because four is more than three.****
  
  And these people think they're smart.
  
  I know, nitpickers, the proportions are wrong. A Canadian dollar is now worth about sixty-seven U.S. cents, while an Intel ''dollar'' is worth roughly eighty-five AMD ''cents.'' I'm just trying to get the general principle down. When this happened, AMD introduced PR, which essentially says, ''Now wait a minute. Our U.S. dollars are worth more than Intel's Canadian dollars. We need an exchange rate.'' That's all PR is.
  
  At most, maybe AMD is saying their dollar is worth a little bit more than it really is, like saying the U.S. dollar is worth $1.60 in Canadian dollars rather than $1.50.
  
  But only a monumentously mentally challenged one would say, ''This is all BS. Exchange rates are BS. I'm taking them 1:1.''
  
  

Aktív témák