Didn't AMD demonstrate a 3 Ghz Phenom quad core a few weeks/months ago? Why in the hell can't they make more of these 3 Ghz Phenoms just like that one?

It's gotta be the 65nm process. AMD needs to just phase out 65nm and start ramping up to 45nm .... FAST. I want to see a big announcement no later than March saying that they have 45nm Phenoms available for sale that are just as good or better than the 45nm Intel Core 2 Extreme QX9650 chips.

Who cares if the 45nm Phenoms are priced at $1,300+ each? Just get it out there for the deep pocket extreme gamers to play around with. AMD needs to get some good P.R. about its chips or the gamers are going to buy QX9650 chips in droves and forget about AMD. I hear you can overclock the QX9650 to 4 Ghz with air cooling ... no problem. THAT right there is a BIG problem for AMD.

How can AMD compete with 4 Ghz Penryns? Imagine the panic that would set in at AMD if Intel released 45 nm Nehalem chips in Summer 2008! If those Nehalems can overclock up to 4 Ghz like the Penryns, AMD is going to be in bigtime trouble. The benchmarks would be unbelievable for a 4 Ghz Nehalem!

AMD needs to QUICKLY add some high K hafnium into their new Phenom chips, crank up the FSB to 1,333 Mhz+ for DDR3, cram 12 MB of cache in there and get those frequencies UP, UP, UP. Get it to overclock to 4 Ghz like the new QX9650 already can.

I'd love to see AMD shock the world in late 2008 and maybe crank out 32nm Phenoms with even more real estate for cache and other processor improvements. I'm sure it's just a fantasy, but everyone just loves to see the little guy (AMD) take on the big guy (Intel) and beat the hell out them. Getting to 32nm first would sure shock Intel.

this chip should really kick AMD's "performance-per-watt" balls with the improvement of power consumption

Lots of idle cores in those game benches.

Now would be a good time to find ways to make our supposedly multi threaded OS's actually work that way. Its stupid that every benchmark gets full control of the PC for its test. Stupid that Win3.11 beat OS/2 in one-at-a-time benchmarks when it was clear to those of us using it in the lab that OS/2 beat the tar out of Win3.11 as a true multithreaded OS. Its not multi threaded unless you can bench a game, crunch linpack or PCA, rip a pile of MP3's, transcode a video, run that big SQL application, and serve up your webpage all at the same time.

I've updated the graphs on the folding page for the Amber and Tinker WUs. The previous graphs had incorrect scores for the Core 2 Quad Q6600.

It's rare when I feel giddy over a processor. Sure it was nothing more than a evolutionary step of conroe but dang, I certainly wasn't expecting anything better. Though I won't be spending 1k on a processor I anxiously await the cheaper offerings.

Great job Scott and Intel.

EDIT: you guys missed out the 95% off sale yesterday, not it's back to $1000+
-------------------------------------------
"buy it now for $70
http://www.tristatecamera.com/lookat.php?refid=8&sku=INTBX80562QX6850
BX80562QX6850 Core 2 Duo QX6850 Extreme
Item #: INTBX80562QX6850
Mfr Part #: BX80562QX6850
Suggested Retail Price: $1,399.99
Price: $70.87"

Awesome, when can I buy one of these for $200?

"Extreme" is so five years ago.

About the Overclocking, i think u got unlucky with the chip u reviewed : http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3137&p=7
on stock cooling too!!

And im sure itl get pushed harder by serious overclocker.
But from all reviews i've read so far, this new chip is pretty awsome.

#1 Awesome ending to your comment Krogoth.

#2 I think there may be more to this processor than what was tested. We all know that no one processor is the same as the next. I would like to think that the one you received, just may not have been one of the better to come off of the stack. I would imagine (at least, going by past record) that there is probably more overclocking potential with these processors. It just seems like this is one of those unlucky, few processors that could not handle the extra push.

Penyrn's architect isn't that impressive. It just a Conroe with a more refined SSE engine and may have more L2 cache. I do admit it is hard to beat already potent and proven Conroe design.

However, the 45nm process is what helps separates Penyrn from its Conroe predecessor. 25-33% power consumption reduction is nothing to sneeze at.

The more important question is whatever Intel can get good yields of it. I am sure that Intel has already build a large stock of Penyrn based chips over this past fiscal quarter.

It is like P2 to P3 Coppermine over all again.

BTW, very good work Damage. I hope some Greeks don't steal your stuff again.

THIS IS TECH-REPORT! [Damage kicks the theif into the pit full of trolls]

These power consumption figures are seriously impressive. A quad-core cpu perfectly suitable for a silent system.
Too bad the 999$ cpu is the only one for now, the dual-core penryns should be very, very impressive too on that front (passive cooling anyone?)

I'd say the best chip Intel ever made.

Just fantastic.

It's like a postcard. From a vacation. In megahertz-land.

Between that and the Ann Coulter joke, I can tell you need a nap.

Looks like a nice CPU, some very nice clock-for-clock gains to be had there, and all for a reduction in power consumption

lol, just reading through the first page... damage if the whole pc industry ever does collapse, rest easy in the knowledge that you can switch to doing standup ;D

ann coulter and pat buchanan, aahahahah...

edit: typo on page 7, "proof positive"

Now, if we could just get X48 and more X38 main boards, not to mention nVIDIA's new chipset main board in the retail channel, we all could see how far Yorkfield can stretch its wings!

Also, I notice the "US coinage as a basis of measurement" is back (even if it's Intel's fault).

Amazing chip. Amazing power consumption conservation. Good review, thanks.

Looking at the various reviews so far TR seems to have got somewhat of a dud overclocker although Damage didn't get the voltage too high. Probably most impressive was HotHardware's 3.9Ghz at a CPU-Z indicated 1.14v. It looks like lostcircuits got there's to boot windows at 4.4Ghz at the same voltage.

Can't wait for the Q6600 equivalent—Q7600? Q9400??

On a side note, my B3 Q6600 hates voltages higher than 1.3V, and according to Speedfan, when o/ced to 3GHz (1333 FSB) it's running 1.25V, though the BIOS reports normal, 1.3V.

Colour me impressed. I was already sold by the solid performance increase particularly in the rendering and folding departments but the power consumption just blew me away. I can't wait until I can pick up a cheap Yorkfield quad-core.

I came to TR to see if there was a review of the 8800GT and more so for solid info on the new process 8800GTS. I find this gem of a review instead. Somehwat dissapointed but only because it wasn't what I was looking for. I wasn't really considering one of the 45nm-based chips but those power savings are considerable.

My own random thought - I believe Intel's engineers may have tweaked the 45nm process for yield over raw clock speed, at least to start. Remember that 65nm was designed for Netburst originally and with the increasing transistor counts that makes sense. Especially since 45nm will be used for Intel's first monolithic quads when Nehalem comes out. Any way to verify this speculation?

Surprise, another good review. Looking forward to lower prices on dual quad core systems eventually. Then I'll buy, thanks to AMD vs Intel for the competition.

Um, not to nit-pick, but it would've been nice if you guys included the Crysis Demo in your test suite, even if only amongst other intel C2Ds to cut down on time.

Crytek stated that the engine plays nice with quad cores, and everyone has been b**ching lately about how resource-intensive it is. Why not throw a QX9650 at it?

Edit: Wow, talk about "ask and ye shall recieve." (see Sunday Shortbread)

Compared to the 65nm Core 2 Extreme QX6850 running at the same clock speed, the QX9650 brings a 33% reduction in system-level energy use during this task.

That's truly remarkable. Just when you think they're fresh out of rabbits, Intel's process engineers pull another one out of their hats.
http://realworldtech.com/page.cfm?ArticleID=RWT012707024759

Kudos on the review. Excellent as always. Other than the lucid grammar, you'd hardly guess it was the product of a liberal arts major (especially when you throw down highly technical terminology like "cubic assload of flying doodads").

That looks like a really nice chip... looking forward to its less... ehem... premium siblings in January.

The power consumption stuff is really amazing.

Woo Hoo! New Reiview. I should be studying, but bah to that, Yay to this...

anyways, to reply to #2, yeah, I think Damage did that just the one time to shut all the whiners up. American site = American money. I'm pretty sure if Dissonance ever did a comparison, he'd use a Canadian quarter.

I like how that first wafer picture was taken with a "JVC GZ-MG555" which is a video camera. Haha.