Multitasking with Conroe

[mmj_uk]

Seem to be a lot of reports that, despite the awesome singlethreaded performance Conroe isn't so hot for multitasking due to the shared cache; and that when multithreading and 64bit finally takes off Conroe will struggle compared to its AMD counterparts?

http://sharikou.blogspot.com/2006/04/conroe-performance-claim-being-busted.html

The reason why Conroe did so well in the MolDyn test is simple: Conroe has a huge 4MB of unified cache, for such single threaded tests that can fit in 4MB*, Conroe can just run off the cache with very high speed. Since cache misses drastically reduce peformance, applications run off cache exhibit unrealistic performance numbers.

I guess that explains the insane superpi times?

However, once you go over the 4MB limit, Conroe is slower than Athlon 64 at the same clock. Both the Cryptography and STREM tests use a lot more than 4MB, larger than Conroe's 4MB cache, and Conroe immediately falls below Athlon 64 on the performance curve.

The conclusion is: clock for clock, Athlon 64 will beat Conroe in real application environments that require a working set of larger than 4MB, or in other words, larger than Conroe's 4MB cache. This means in any real multi-tasking or server environment the Core architecture will be an underdog. Even worse, for Intel's shared cache architecture, cache thrashing is a distinct possibility under heavy loads.

So have we all been conned by Intel?

I mean you buy Conroe and the new Intel quad core for multitasking and as a future proof option for multithreaded games so it's a bit of a con if they turn out to be lemons under those circumstances.

http://www.google.co.uk/search?hl=en&q=conroe+multitasking&meta=

 

[Cob]

At stock speeds maybe.

But put an average overclocked X2 3800+ (2.5Ghz ish) up against an average overclocked e6300 (>3.1Ghz) and the answer may be different.

 

[Entai]

But that is not comparing like for like then, so is not a fair comparison.

You can only get a fair comparison at the same clock speed for both.

 

[VeNT]

But that is not comparing like for like then, so is not a fair comparison.

You can only get a fair comparison at the same clock speed for both.

I'd say you should compair them at stock then at the max overclock aswell.

 

[Corasik]

Sharikou is just an AMD fanboy, and his data is all based on a single benchmark which just happens to have been written by an AMD engineer, and is specifically optimized for AMD processors.

Clock for clock the Conroe is more than a match for AMD in any 'fair' test. In fact in the large number of tests, a 2.4Ghz Conroe will outperform a 2.8Ghz AMD64.

In 64bit, the conroe is not quite as far head, but its still slightly faster at any given clock.

 

[NathanE]

The guy seems to think the L2 cache contains one application at a time. This is not the case and in fact very far from the truth. As this is a fairly fundamental oversight on his part and indeed the center of his argument I cannot agree nor can I be bothered to read further to see what else he has got wrong.

Fundamentally, 4MB is better than 2MB. Anyone knows that. He is trying to convince people that having less L2 cache is somehow faster than having more. A shared L2 cache is superior to seperate cache's on each core. AMD knows this and that's why its future designs will use a shared cache...

 

[allllec]

But that is not comparing like for like then, so is not a fair comparison.

You can only get a fair comparison at the same clock speed for both.

disagree, imo they have to be compaired by price, dosnt that make more sense or am i mad

 

[clv101]

disagree, imo they have to be compaired by price, dosnt that make more sense or am i mad

For sure - price and power consumption are the only things that matter at the end of the day. It makes no difference at all if the chip is running at 10MHz or 10GHz as long as the performance, price and power consumption are okay.

 

[Combat squirrel]

*strokes Opteron* there's a good processor, there's a good processor