Can some explain these cpu differences to me in an easy way?

[jagr10]

Ok, i'm not that knowledgeable when it comes to cpu's. I would like to know what the most important difference is between the following types of cpu's:

AMD Opteron
AMD Barton
AMD Athlon XP
AMD Thoroughbred
AMD Athlon MP

Basically, could you list them in order of fastest to slowest? Also, could you just point out the major differences for each? Reason i'm asking is 'cause i see some sites with a 1800+ athlon xp and 1800+ thoroughbred and the thoroughbred will be more expensive. Thanks.

 

[Ronin]

Opteron - Not conventionally available yet, but the next generation of AMD CPU before the Athlon64.
AMD Barton - XP with higher L2 cache (512K rather than 256K)
AMD XP - The CPU that AMD started the PR rating with.
AMD Thoroughbred - The XP's are based on the thoroughbred core, so this is pretty much the same thing
AMD MP - This is AMD's Xeon

For the end user, right now the Barton is probably the best purchase with the added L2 cache. You'll note the PR rating for the Barton's is a little different than with the XP line (basically, a 2500+ Barton isn't over 2GHz, where a 2500+ XP would be).

 

[jagr10]

thanks!

 

[jagr10]

ANother thing, which processors have which "micron" technology? Someone mentioned to me about .13 micron or .15 micron. Which is better and which processors have which? Also, it's not too much trouble, what's the "on-die cache"? I read about how an amd processor has 640KB on-die cache. Is that the same as L2 cache?

thanks.

 

[yak8998]

'micron' refers to size. I can't exactly explain it, but basically, the smaller the microns, the more they can fit in the same size package. I believe lower micron manufacturing = higher clock speeds (rite?).

AMD MP - This is AMD's Xeon

I wouldn't go that far. The MP is just an XP that runs in dual-cpu setups if I remember correctly. The Xeon has a bigger cache (in most models) and I think there are some differences in the pipeline and stuff...

EDIT: On-die is L1. It runs faster than L2 (at the speed of the CPU I believe.)

 

[Ronin]

If you wouldn't compare the MP (which AMD has specifically stated this is what it is) to a Xeon, just because of the cache sizes (even though the MPs outperform the Xeons), then what would you compare them to?

The die size is what the micron # is referring to. The smaller the number, the smaller the die, the less heat it generates, and the higher the speed of the CPU.

 

[Intelman07]

Or you could go with Intel....
Pentium 4-Desktop/Lowend workstation
Celeron - Budget chip
Xeon- Server/High End workstation chip
Itainium - Intel's 64bit processor

Hehe...micron is the die size of the chip, usally smaller is better.

 

[pspada]

Actually, micron is a unit of measurement equal to one millionth of a meter. The older chip familys in the Athlon line like the Tbird and Pally used a .18 micron space between the leads (wires) inside the chip. Then the Tbred and the newest, Barton family now use a .13 micron space between the leads.

Then there is the Opteron. While all of the Athlon chips have been 32-bit processors, the newer Opteron is a 64-bit processor. Mostly intended for servers, at least at this point, it can directly address lots more RAM, and has other enhancements for a "server" class machine.

Now, the cache. basically, a processor works really, really fast, whereas the RAM, no matter how speedy, is very very slow in comparison. This would result in the processor have to wait for information from the RAM, resulting in much slower processing. So some high-speed memory is created on the chip itself, and runs just as fast as the processor itself. In fact there are several types, the two L1 caches of 64kb, one data and one instruction, and the larger L2 cache. So, the processor is fed from this memory, which is fed from the on-motherboard cache, which is fed fromt the RAM itself. This causes the processor to have to wait for information much less often, increasing the speed of operations. So, larger cache, better performance. But this also means greater expense, as you can not produce as many processors from the same wafer, as well as a larger number of pieces that don't pass spec and can't be used.

Now, we can speak of the Barton. While all prior Athlon chips had 256Kb of L2 cache, the Barton has twice that, 512Kb. This results in math operations running significantly faster on Barton chips. There is one other factor, called the Front Side Bus (FSB) that I will not explain in detail, but on the first Bartons and a few of the faster Tbred chips, the FSB was 333Mhz instead of the prior 266Mhz. And now the newest Bartons run a FSB of 400Mhz. As in the Mhz or clockspeed of a chip, faster is better.

Hope this clears up any confusion. :sun:

 

[Ronin]

512Mb of on die cache? Wow, I should rush out and buy one before they realize this manufacturing mistake!

 

[pspada]

Originally posted by: RoninCS
512Mb of on die cache? Wow, I should rush out and buy one before they realize this manufacturing mistake!

Ooops, edited to fix. Boy, do you think the chip would have to be any larger for a 512Mb cache?

 

[Ronin]

lol

Could you imagine? Then again, the Xeons aren't that big with the cache they have, but then again, it's not 512Mb

 

[beafer]

Actually, even though i know it was a mistake on the cache but that amount of cache would hurt more than do any good.

 

[DivideBYZero]

Originally posted by: Intelman07
Or you could go with Intel....
Pentium 4-Desktop/Lowend workstation
Celeron - Budget chip
Xeon- Server/High End workstation chip
Itainium - Intel's 64bit processor

Hehe...micron is the die size of the chip, usally smaller is better.

P4 is the only real choice out of this collection.


Get a Celeron to experience P2 levels of performance, TODAY!
Get a Xeon if you like getting loans for the CPU alone(the multi MB cache models)
Get an Itanium and join a 3000 strong band of test pilots...

 

[redhatlinux]

Better do some research on the 'space between the leads'. That's not what the 0.13 micron refers too. I'm sure a Google search will tell you what it really is.

 

[redhatlinux]

Better do some research on the 'space between the leads'. That's not what the 0.13 micron refers too. I'm sure a Google search will tell you what it really is.

 

[yak8998]

I would compare the Athlon MP to...

AMD XP - The CPU that AMD started the PR rating with.

that. Except add-on multi processer configuration compatibility.