CPU/Mobo Intel Conroe News Thread

RiO said:
I can't stomach all this jargon, so let's stick to our guns... you with AMD, and I, Intel ;)

It's not jargon.. if you want to be an Intel fanboi, read up a bit on CPU architecture.

And I'm not sticking to AMD. Where can you conclude that from??
 
Take it easy, techhead - we both benefit from the AMD/Intel and nVidia/ATi wars :) Honestly, I don't know/care about their architecture... I care about performance and OCing (but not with my prescott)... and yes, I have always used Intel.
 
TechHead said:
Architecture, not die size makes a CPU competitive.. all that 65nm means is that the distance between individual components on the CPU die is now 65nm. That's it.. allowing Intel to stuff more transistors onto the die. But if that's coupled with a 31 stage pipeline e.g. the P4 line, how is that going to help in, say, branch prediction errors?

You summarised it a lot better than I did. Exactly what I've been shouting out loud some posts ago. Fabrication is not directly related to performance! A low process simply means that the company's going to make more money as they're saving in silicon. As an effect of Moore's law, there are two ways to take advantage of decreasing transistor sizes. The first, like Intel has been doing all this while, is increasing the clockspeed. The second is adding more functionality. More execution units, better branch prediction and the like.

Then there are the drawbacks of low processes: increased crosstalk, power density and transistor leakage. Case in point? The PRESCOTT. Let's take this whole 65nm deal with a pinch of salt okay?
 
I am pro Performance at decent price be it AMD AM2 or Intel Conroe... as I have liquid cooling it goes without saying that I am gonna OC it till it takes.
 
Guys, Anandtech has taken a second look at Conroe. "We ran the AMD system with an updated BIOS, ran our own Quake 4 demo on Conroe and re-ran some of our original tests. Find out what's changed."

http://www.techenclave.com/forums/conroe-performance-preview-follow-up-68337.html

Edit after the post below : Hunt3r, nope i posted it at the exact same time as you did 4:51 PM, check out my post :) i started a new thread cause this was getting too big and outdated :)
 
Tracerbullet said:
But if that's coupled with a 31 stage pipeline e.g. the P4 line, how is that going to help in, say, branch prediction errors?
Well, intel have found one of the main culprits limiting the performance of their P4 Prescotts..and that is why they have gone back to 14 stage pipeline with the Conroe from the 31 stage of the prescotts... so cutting that deep pipeline is sure going to help in the performance, albiet, for sometime intel wud be sticking to 3-3.3 ghz barrier now..
 
Wow, on Intels, own Water Cooling Solution, the 3.46 Ghz Proccy @ 5 Ghz.

I do have a feeling that Intel will only offer this cooling as a bundle with the Conroe EE Products :(.
 
Start saving... :p Although, people who can afford an EE might just dump the stock water cooling for swiftech's - they'd end up on eBay eventually.
 
Well, if the 3.46 Ghz can do 5 Ghz under Intel's Water cooling, then i have no idea why would any1 waste more cash, unless its for Phase Change.

Imagine the Performance of the Conroe @ 5 Ghz :O.

Itll own the AMD AM2's upside down, inside out.
 
I would say that 4.2-4.5GHz is what it should do... more then that it will be pure skills + luck. I can be wrong, but from what i have seen Dothans/Yonah do, Conroe should be this much with "shorter pipeline". I dont know about software, but usually Intel optimised things are better so it can give good results with few tweaked softwares too.
 
Sandy said:
it dont say how it was tested :p they may have used 1ghz ram for testing :p

6 months more and its here :D

I seriously doubt ram would make that difference...:bleh: c'mon kudos to Intel for doing something right
I doubt even 1.6 GHz ram would make anything faster (and by that extent) cause there are other latencies involved.
 
Back
Top