The Prescott's enhancements are designed to allow it to scale to higher clockspeeds, and do so more efficiently. They were not designed to boost performance at the same clockspeed. This is why you are not paying extra for Prescotts at identical clockspeeds to Northwoods.
That was done intentionally, why bother comparing it head to head with high-end processors when those CPUs are not commonly found in systems of people that are about to upgrade or buy a whole new system. By comparing it to an older system one can much more accurately determine just how fast or slow this processor is. The 3.06GHz has been out for about a year, and many people have a system at home based on this CPU. As our benchmarks show the 3.06GHz is barely slower than Intel's latest sibling. Ergo conclusio, if it's about as fast as a 3.06GHz running at a 533MHz it is slower or on par with faster CPUs.
Is it just me, or does Intel's mantra of "clock speed is king" seem to be creating more problems for them, at this point in the game, then it's worth?
It seems that everything Intel's done to this new revision of the core has one purpose, making it easier to scale up the processor speed.
They may be the first to hit a 5ghz cpu, but I bet they'll be the first to a 100 stage pipeline too. LOL
It's just too bad that the average computer buyer uses clock speed as a major yard stick for their buying decision, it plays right into Intel's hands.
Ok, I'm a bit of a AMD "fan boy", that's why I get cheesed at AMD having to use the stupid "performance rating", just to show that a "slower" cpu can perform as well as a "faster" one.
In my mind when marketing takes over R&D there's a problem, and IMHO this new Prescott proves it.
Athlon XP M 2500+ (12x200FSB=2.4 ghz, PR 3500+), Soltek SL-75RN2-L, 1gig 3200DDR Ram, ATI X800GTO, NEC MultiSync FE991sb, Creative Audigy OEM, Logitech 5.1 speakers, 40gig HD booting XP, 200gig@50%games, Lite-On DVD burner, LG 52x32x52x
Josh - Why is indeed the question. No one seems overly willing to discuss it. The processor draws much more power than one would expect, even given the larger caches (not very power-intensive), and additional logic. It only makes sense to conclude, then, that there's some sort of efficiency or current leakage issue with the processor. Intel seems to insist that it can be fixed as the process technology improves. I suppose we'll just have to wait and see.
why not just switch back to traditional silicon instead of strained silicon, i mean the current northwoods kick ass and run cooler than the prescott, personally only thing i like is the enhanced L2 cache that is 1024KB(1MB)but that L1 Cache needs to be more like 512KB(1/2MB)-1024KB(1MB).
You guys have heard of Virtual Memory and Multiple CPU's, what about a Virtual CPU?
The 80386 used a 32bit bus to interface with system memory. The big deal with the 80386, however, was that it could run multiple copies of real mode. This meant that the 80386 could run several virtual 8088s. The 80386 used paging to remap memory so that each virtual 8088 got recognized about 60 times a second. This is how we got multitasking and in Windows 95 the chop length dropped to 20 ns.
The Pentium - Back to the Future - using Multiple CPU's
After adding a cache and offering some other refinements the 80486 gave way to an idea the Prescott could expand on today under pressure of its thermal limitations. The Pentium improved upon the 80486 by running two 80486s in parallel. Typically two instructions could be executed at once. The advantage was obvious but required software that would take advantage of it.There were separate 8K caches, for instructions and data, split into banks, which could be accessed alternately. For all you Itanium and Athlon 64 fans it also had a 64-bit bus to allow both 32-bit CPU's to run in parallel.
The Prescott Deluxe?
Nope, its not a vacuum cleaner. With all that space .09 provides on the die... well I think you see where I'm headed... Why not quad CPU's?
I think you have a few issues with your article. First, benchmark testing of a performance equal AMD CPU would of been very useful to all readers. I don't think I need to explain why. Second, at the very least you need to add an update, probably to the "somethings rotten" page, regarding the serious heat issues of the prescott. I've read about it else where but I don't archive links so I can't tell you where. The articles are out there though and should be referenced in yours. Third, and most importantly to me as its total speculation, is the "aside" remarks at the bottom of page 5."As an aside, these 90nm Celerons may be of some interest to overclockers. A 2.53 GHz (533 MHz FSB) Celeron using the 90nm process will be introduced in Q2 2004, and depending on its price and the maturity of the process at that point, may prove to be a capable overclocker." What are you thinking? Is the 90nm process going to make a hero out of a chump? The celeron sucks and anybody with a little knowledge knows it. It's a "value",lol, based cpu that is primarily used by the OEM market to fool the uninformed buyer with clock speed and omitting FSB and Cache factors. Current celerons are 400FSB and 128k cache. 90nm celerons will be 533FSB & 256K cache. WOW! NOT! 533FSB is good, 256k is not. Is the 90nm tech really going to help this thing o'clock? Probably not much better than current .13 micron technology(only my opinion and open to discussion). You can get a P4b 2.53Ghz,533FSB,512kcache for $134. How much do you think the celeron 90nm will debut at? I'd love to see a benchmark showdown between the two. Which will prevail, 512k .13micron or 256k 90nm? How do you think the new celeron will stack up against the P4 2.4c,800FSB,512k? The P4c 2.4 sells for $155 and I bet the 90nm celeron comes close or costs more. The whole P4c nothwoods are known, serious o'clockers and the 2.4 can easily get into 3Ghz land, I know of one at 3.5 using a divider. The 2.6 sells for $160,2.8 for $167,3.0 for $189. I think you get my point. Lets not forget the 800FSB and 512k cache vs 533FSB & 256k! If i want to spend less than what a P4 costs and get a hell of a lot more than a celeron(and I bet the new one too) then I'm heading directly to AMD to find a "value" cpu.
All in all I find your article a good read but your "aside" is assanine and absurd. When the 90nm 2.53celeron arrives, benchmark it against the AMD's, P4b 2.53, and the P4c's. Then do a price comparison and I believe we'll all see that while the cost, body, and performance of a whore may change, nobody knowingly wants to marry one. lol All this "aside", I guess I'll stick to my firstname.lastname@example.org">P4c2.email@example.com o'clocked and rock solid.