as a graphics programmer, i'm looking to get as many cores as i can since i'm almost always compute-bound; however, i must say that amd's 4x4 solution is looking better than kentsfield for two big reasons:

1. heat - when overclocked, this chip simply produces WAY too much heat for comfort. with two sockets the thermal load is spread much more effectively. like the radeon 1900xt, this is exacerbated by compute-bound workloads that don't wait on system memory.

2. bandwidth - when i'm not specifically doing ray tracing, but other stuff like skeletal animation, texturemapping, image operations and things like that, things suddenly become the very data-hungry. being cache-aware helps and is a primary thrust of modern efficient programming, and here the total of 8mb l2 (!!) is undoubtedly very effective; however, in the worst case when there are just memory accesses flying all over the place, 4x4 should crush it (under specific conditions, see point 3).

and there's one programming-centric reason to get the 4x4:

3. numa - it's great practise to get programming on a numa box so one can learn the do's and don'ts of proper thread-aware memory management.

anyway, that's my ramble. exams are over and i'mma gonna get coding :P

Athlon 64 4600+ still the best bang for the buck.

Unless you render all day long, that's all you need.

Core2duo ( fake conroe ) 6700 model is still $500, more than twice as expensive as athlon4600+ and hardly any difference in 90% of cases.
And you could buy an athlon, nice video card and almost 2GB of memory for that.

Why no comparisons to a two socket opteron system? Since that is what 4x4 will essentially be, I'd think that would be a comparison of great interest.

Also, why not do some more aggressive multitasking tests? For example, run multiple instances of a video encoder while playing a game?

I feel like so many reviews are just cookie cutter copies of past reviews. There's not much creativity out there anymore. I wish Johan were still doing reviews at Ace's Hardware -- that guy really put some serious thought and creativity into what he was doing.

Obligatory "We can't possibly use that much processing power" post....

oh wait, gimme, gimme, I want one!!

Looks to me like Kentsfield is quite heavily FSB-bottlenecked at 3.2GHz on 266, your Cinebench score is only like 50% higher than on a 3.46GHz Conroe. Perhaps a bit of FSB overclocking would have helped. Quad-core looks to be pretty useless for most people but it does sound quite nice, hehe. I'm glad the XE will actually offer something significantly different from what people can get from the 180 buck parts so they're actually "worth" their premium.

#5, Kentsfield is fine as an XE part, its profit margin is nothing to scoff at. AMD does need 4x4, if only to offer an alternative to Kentsfield (though a very power-hungry one). I also find the 4x4 motherboard quite nice, since it's packed with all kinds of goodies (12 SATA ports, 4 16x PCIe slots, so nice) and hopefully wont be too expensive ($250-$300 is still OK for a premium motherboard but anything above that is excessive for the consumer market).

Fascinating. I was planning on a Core 2 Quadro when they come out, but looking at these results, it seems a good C2D and a healthy OC will eat up the C2Q, and run cooler and quieter while doing so.

The four cores have me all happy, but the thermals have me scared. Intel proves they can release the hottest chips on Earth, thermally speaking, yet again.

E6600 plus OC FTW! By the time 680i rolls out in meaningful numbers, I hope the E6700 is down to reasonable pricing.

Given that it has to use the FSB to communicate between the two individual "cores", is there a real clear difference between this and just using a dual socket, dual core set up? I'm sure there would be a slight difference, but I would think it would be rather minimal.

This will be interesting to see once the quad core Xeons are out.

I'd really like to see the Quake 4 benchmark done with the 1.3 patch. It made a HUGE difference in multithreaded performance on my computer... which is a 3.0 Ghz Pentium D. From 55 fps to 85 fps. The 1.2 patch did basically nothing.

How 'bout it damage??

I'm normally i lot more excited abotu a new CPU release specially an increase in cores. but *blah* However I'm insainly excited about that F@H benchmark!! Where can i get a copy (sorry not a forum troll) and may i say guys I've been waiting for somethign like that for a long time.

Damage, you've done it again!!

I aways thought the term was "beckon call" not "beck and call".

wow, games need major improvements/optimizations to deal with multi cores. those benchies are not as impressive as I thought they would be.

Will Quad cores add any benefit it we are multitasking four or more resources while gaming? Ex folding/spy ware/antivirus/gaming???

A very interesting but useless hack. I'll wait for the "real" quad cores to come out, maybe by then software to take advantage of those cores might be only a year or so away.

It's a nice gimmick though, just like the equally useless 4x4 crap AMD will try and push.

I still think my C2D 660 (currently @3GHz) which I bought last week was a good choice.

There are only a few applications which profit from Kentsfield, and if you are not running them very often, it is more sensible to get a dual core now and a quad core later when they are less expensive and the software has advanced accordingly.

I'd love to let this thing loose on some Divx 6.4 or MT Xvid encoding. My dual core ate that stuff up when I was doing some big projects a few months back.

Here's the bit-tech article from Valve about their multi core approach:
http://www.bit-tech.net/gaming/2006/11/02/Multi_core_in_the_Source_...

A good read about some of the specific issues related to multi core programming and games.

I, for one, welcome out multi core overlords.

I wish there would have been more multi thread bechmarks, i like all the game bench marks, but i think multi tasking would be more interesting with this type of CPU for the time being, really show why i want to buy one, these tests make it a waste of time to even think about. Love the web site but gatta fix the bench marking tests for the upcoming processor show down. excuse the spelling, nerds dont spell well.

An interesting review. With a few exceptions, my overall impression is that its choked by the FSB design. It will be very interesting to see this stacked up against the AMD cuatro cores next year. I've got to get my FORTRAN compiler working again. It would be interesting to modify the linpack performance test for multiple instances and substitute singular value decomposition for gaussian elimination.

First, it's nice to see some UT2004 lovin'! A quick reminder that the TR UT2004 server is open to the public! Stop on by and play with us! http://techreport.com/forums/viewtopic.php?t=27977

Now, the thing is this quad core stuff looks great but what's the point for gamers? Dual core looks like it will be plenty good through 2007 and maybe into 2008. Am I wrong?

Will the release of this cpu trigger a conroe price cut?

As an enthusiast or gamer CPU this looks quite uninspiring - spend the money on a Core 2 Duo instead.

For a professional it looks far more tempting as an alternative to a dual-processor workstation.

AMD's 4x4 will be similarly pointless for enthusiasts or gamers, but the platform is likely to be more interesting (quad PCIe x16, 12 SATA, up to 24GB/s memory bandwidth, but 14-16 more likely) and allow better scaling for professionals. Whether or not that will make up for the deficit in core performance is another thing.

In a year's time when games get better multithreading then quad-core systems will be far more useful in general. Luckily 4x4 and this processor are laying the foundations for this.

I'm surprised there aren't more positive views on quad core. I would like to point out that in this case, these are very synthetic benchmarks. I am still running a single core, but when I upgrade, if I can afford it, I will be going quad. Even if by that time software isn't optimized fully for 4 cores, that is not the only consideration. I mean...only the newest of users runs only ONE program at a time. My computer is running dozens. This is inherently multithreaded and will make full use of 4 cores or more. While some of these may be light duty tasks, some are fairly heavy, like bittorent, media player, and various servers, db etc. I think quadcore will provide a decent improvement in performance in this more real world situation. I mean, how many cores do I need to run all of the lovely flash ads on the average website :P.

Quad-core will change every aspect of PC gaming. It will change how we create our games, how we provision our service, and how we design our games. The scalability we've seen in graphics over the last few years will now extend to physics, AI, animation, and all the systems which are critical to moving beyond the era of pretty but dumb games.

Yeah, right... All the AI, physics and animations in the world will help against the poor substance of games nowadays.. Thus, in 9 out of 10 cases, I predict it will indeed get prettier, but also a lot dumber...

Useful for about 4% of the computer buying world right now? Maybe less?

hmm... Well... I guess this may possibly be something to buy five years down the road :P

doesn't impress me at all. its really lagging behind in simple one/two threaded tasks. its not blasting thru games either like BF2 and F.E.A.R.. might as well just stick to Core 2 X6800.

can't really blame the CPU for that. game engines aren't parallelized that much yet. so for now i guess all of us SHOULD stick to X6800 :)

This would be the nuts for my work machine, but I couldn't use two of the four cores at home unless I was watching movies while working and folding and... wait, I guess I can use 4 cores. Certainly not when gaming until games catch up, but certainly when using my machine the way I like to use it. Too bad I have an AMD board right now that has a lot of life left in it!

If you don't get picked up for speeding witha C2D OC, they won't pick you up for breaking Amdahl's Law.

It's a shame reverse hyperthreading doesn't exist, what with all these multicore CPUs arround now.

Yikes. I hope we don't get cited for violating Amdahl's Law.

Tee hee