Intel's multi-core onslaught has reached the professional space. The Xeon 5300 Clovertown allows administrators to upgrade existing Socket 771 dual-processor machines from four to eight cores. We let the features and numbers speak for themselves.

Interesting review, although I swear that in some of the tests that the Woodcrest beat the Clovertown the app itself was fairly multithreaded already... meh...

Most of these benchmarks don't do the Xeon 53xx series justice. These processors will probably be used for desktop rendering and the like, but I start drooling when I think of an 8 core 1U 2 socket box with SQL Server, IIS, or a dozen virtual application servers on it, or replacing an four racks of 4U 4 socket single core servers with three shelves of HDD's and one rack of 1U boxes, and ending up with more horsepower, less power consumption, easier management, and space for a ping pong table in the server room.

grrr why do i get http://tomshardware.co.uk/error404.html not found when articles are new :\

Thats what I think also... don't know why they always have so many multimedia type app benchmarks... probably because they're easily on hand and easy to configure... shrug.

Hmm, decent review, but where are the Opterons?

It looks like Blackford with dual FSB sux for Clovertown. I am happy to see that AMD will have at least one front where they will outperform Intel in the next 12-18 months.

Another "brilliant" review from Tom's HW.
When will they finally start to benchmark server CPUs with server tests / workloads?
The only result i find interesting of the whole article, is the power consumption test, where the 8 core monster surprisingly surpasses the competitors...

Great review, but can anyone please put an apples v apples review of a qx6700 top of the line machine (ie, workstation w/consumer parts) vs a top of the line x5355 dual cpu/quad core system? This review (and countless others i've read) aren't telling me anything useful about these CPUs performance in apps like Adobe Photoshop CS2 and Premiere Pro 2.0.....someone please put some benchmarks together for this!

Ok, I admit it. I'm not the sharpest tool in the shed. I watch Maury and still get a kick when the big ugly dude IS the father. So I'm not afraid to ask stupid questions.

Registered RAM is more expensive than the garden variety. Fair enough. Now what other reason is there to not use a twin Clovertown as my own workstation?

Will it run Vista? Has to.

Will it run all the sw that runs on Vista? Should.

Will the sw that can take advantage of megacores run like a hamster trying to get away from Richard Gere? Damn straight.

So what are the reasons why this is being marketed exclusively as a server product? What am I missing here?

Hi!

I would really like to see the values used in the Linpack configuration, specifically the value of N, P and Q.

Thanks!

Raul

Is Linpack's floating point measurement still relevant to megacores? Just askin'...

I agree...I would never run these tests for benchmarking software. The performance of the system with various loads under MySQL would be much more beneficial. I can't make a decision based on how fast it can code dvix.

They should take their time and do a server benchmark which might mean something to someone.

I'll let you know how these things pan out here in the next couple of weeks. I have a couple of HP DL380 G5's that are coming in with two quads in each box, 8 and 16 gig of ram respectively and 8 146GB SAS drives in each box...One is indeed going to be used for VMware.

I'll be able to compare these side by side with a DL380 G5 with a pair of 5150's in it. Granted clock speeds are diff on the boxes (quads will be 1.83, the 5150's are 2.66). HP doesn's have the higher end quads available in their DL series yet....

Thats cool, the point of having 4 core cpu's is to run as many concurrent threads as possible...max performance per core isn't as important especially considering the heat the top quad cores make and cost.