AMD ghz vs. Intel ghz

Status
Not open for further replies.
sorry about the poor title but i figured that would make it more "noob friendly".


K8 (Athlon 64, Sempron 64) Vs. P4/P5 (All Pentium 4's including Celeron's and Pentium D's)

one of the current most popular questions on the forum goes something like this, "how can a 2ghz AMD processor outperform a 3ghz Intel processor?".

the reason is that clock speed (i.e frequency in hz/ghz) is not the only factor that determines how powerful a cpu is. basically it is a simple multiplication that determines total power. on one hand you have the processor frequency (a.k.a clock cycles per second) and then you have the number of instructions per clock cycle. AMD performs far more instructions per clock cycle which gives it higher performance at a lower frequency. normally however the more instructions per cycle the lower the maximum frequency of a processor will be. Intel realised that people thought frequency was the only factor determining performance so they lowered the instructions per clock cycle so that their processors could run at higher frequencies. AMD uses a 12 stage pipeline and the prescott core uses a 31 stage pipeline, this article explains instruction pipelining
http://en.wikipedia.org/wiki/Instruction_pipeline

here is some more information on the difference
http://www.geek.com/news/geeknews/2005Dec/bch20051215033811.htm

also many have been asking about why AMD's dual cores are more efficient. aside from the differences mentioned before Intels dual cores are not true dual cores, they are basically 2 full processors on one chip rather than 2 individual cores on a processor chip. AMD's cores are much more closely linked and can communicate directly with each other giving a much more efficient (and dare i say, better) processor. each core on the intel core must communicate with each other through the northbridge on the motherboard which will take significantly longer than direct communication and basically results in the computer sitting idle momentarily waiting for the information to transfer. of course the delay is very small but overall it has a reasonable impact.

lastly AMD has its memory controller built directly into the processor which allows quicker transfer of data from processor to memory. Intel still uses a memory controller located on the Northbridge of the motherboard. this takes significantly longer to transfer data.

to make an approximate comparison between an AMD chip and an Intel chip just use the p-rating. the p-rating is the number with the + sign e.g. 3700+. the p-rating is approximately the number of mhz a prescott p4 would have to be clocked at to equal it. the p-rating was originally determined comparing the athlon to the old athlon thunderbird core which as it turns out has fairly similar performance per clock cycle to a prescott core.

note: this information applies most accurately to the prescott and smithfield core processors. the instructions per cycle of the mobile dothan and yonah cores is closer to that of AMD. the presler/cedar mill cores (65nm) perform better than the precott and smithfield (90nm) cores due to their doubled L2 cache but not that much higher. although the 65nm cores should yield higher overclocks

K8 (Athlon 64, Sempron 64) Vs. Core 2 (Conroe, Allendale, Merom, Woodcrest)

this comparison is almost like the Athlon 64 vs. Pentium 4 one above but with the roles reversed. the Core 2 is more efficient "clock for clock" than an Athlon 64. for instance benchmarks demonstrate that the Conroe e6600 which is the middle of the Core 2 range convincingly beats the FX-60 in virtually everything and gives the FX-62 a run for it's money. the e6600 is clocked at 2.4ghz, the FX-60 at 2.6ghz and the FX-62 at 2.8ghz. the Core 2 range is somewhere between 115% and 120% more efficient in processing than the Athlon 64's.
 
wow didn't know intel dual core were actually 2 seperate chips. i guess its a lot easier to do that than actually doing some R&D.
 
i just couldn't handle answering this question more than once a day which has been happening for the last few weeks, hopefully people will search the forum instead of posting first but thats just wishful thinking i think. i think the reason intel went with that design for their dual core cpu's is because they really rushed the design to beat AMD to the release date. AMD had been working for several years on it and Intel got wind of it and slapped theirs together in months.
 
Actually with the Core Duo's isn't the Intel and AMD around the same now regarding Ghz speed if not Intel winning a bit.
 
yeah but the question always asked regards either the smithfield or prescott cores to which this applies. its a little different with the cedar mill/presler/dothan/yonah cores. good point though i'll add that i meant prescott and smithfield.

edit:
3000posts0ci.gif


:D:D:D:D:D:D:D:D:
 
spam is nothing to be congradulated on....

btw there's an old intel vs. amd thread with some good info in it
 
i am pretty sure this does not constitute spam. also i was referring to my 3000th post so don't attack me because you misinterpreted me. i am aware there are many AMD vs. Intel threads but i have come across none that put plain and simple what the difference is. due to the recent plague of questions about how AMD processors at lower clock speeds compare to Intel cpus i chose to create this thread in the hope that it might save people some time, both for those asking the question but more for the people who take the time to answer it over and over.
 
Status
Not open for further replies.
Back
Top Bottom