Clovertown?? ... I don't get it!

[cryogenic]

Intel just managed to get a negative performance increase by doubling the number of cores from 4 to 8 on 'multi threaded software'. Can please someone explain to me how is this possible ??? Someone must have screwed up big time, but that just doesn't quite say it.

Even worse the negative performance increase is achieved by using Intel's best architecture, Core 2.

At this point I'm quite sure Intel's 45 nm upcoming chips won't make much difference in 2007, the FSB it's killing them.

And I thought AMD 4x4 launch was poor ... well actually 4x4 managed to bring positive performance increase ...

Share you thoughts on this with me, please!

 

[accord99]

Intel just managed to get a negative performance increase by doubling the number of cores from 4 to 8 on 'multi threaded software'. Can please someone explain to me how is this possible ??? Someone must have screwed up big time, but that just doesn't quite say it.

Most of the benchmarks in the review are workstation tests, that don't really take advantage of more than a few threads. Plus the comparison is with 3GHz Woodcrests vs 2.67GHz Clovertowns. Even for good scaling tests,

And I thought AMD 4x4 launch was poor ... well actually 4x4 managed to bring positive performance increase ...

The FX-74 gets beat by the FX-62 in most single-thread applications, especially in games.

 

[Pippero]

The FX-74 gets beat by the FX-62 in most single-thread applications, especially in games.

Yeah, but what did you expect?
A quad core is of no use for single threaded applications.
Sure FX-74 has a 200MHz advantage over FX-62, but with a non NUMA aware OS, it has far worse memory latency. (and with a NUMA aware OS, it probably can still have some memory latency problems)

 

[r0ck]

At worst, one would expect the 3.0 to at least be on par with the 2.8..It's not like the 2.66 Kentsfield performs like a 2.13 E6400.

 

[cryogenic]

Intel just managed to get a negative performance increase by doubling the number of cores from 4 to 8 on 'multi threaded software'. Can please someone explain to me how is this possible ??? Someone must have screwed up big time, but that just doesn't quite say it.

Most of the benchmarks in the review are workstation tests, that don't really take advantage of more than a few threads. Plus the comparison is with 3GHz Woodcrests vs 2.67GHz Clovertowns. Even for good scaling tests,

And I thought AMD 4x4 launch was poor ... well actually 4x4 managed to bring positive performance increase ...

The FX-74 gets beat by the FX-62 in most single-thread applications, especially in games.

A few threads??? yahoo messenger has around 30 threads, the winlogon process can have 10, 20 or more ... FireFox with 5 open pages has 17 ... I'm quite sure any one of the multi threaded benchmarks spawns (can spawn) more than 8 threads. The system process in windows can have hundreds of threads ... A typical instance of SQL Server running on my dev. machine has close to 300 threads ...

As a software developer that works with databases every day I sure would like to see some test on database performance.

As far as I know the best threaded software and most scalable is encoding software so I'm expecting to see even worse performance for database tests ... also databases don't even require that much number crunching power they just need fast memory and disk access to scale well.

 

[Pippero]

At worst, one would expect the 3.0 to at least be on par with the 2.8..It's not like the 2.66 Kentsfield performs like a 2.13 E6400.

Uhh, first off:
Intel: 2.66 / 2.13 = 1.25 = 25% clock advantage
AMD: 3.00 / 2.80 = 1.07 = 7% clock advantage
So as you see, the gap between FX74 and FX62 is much more narrow than the gap between QX6700 and E6400.
Second, Clovertown uses the traditional Northbridge / FSB architecture.
This means that if one CPU is idle (e.g. single threades app), the other is getting no additional latency in accessing memory, and it can use the full bandwidth of the FSB, so it can run at the best of its potential.
In case of HT / Direct Connect architecture used by AMD, sure the 2 CPUs have a theoretical aggregate bandwidth which is twice as high as a single CPU / single socket system.
However, since the RAM is split between the 2 CPUs, if one CPU needs data which is stored into the memory of the other CPU, it has to fetch it through the HT link and has to go through much higher latency than if it had to access it through its Integrated Memory Controller.
With a non-NUMA aware OS, single threaded (or poorly threaded) applications can get their memory space allocated across both CPUs, with overall much higher latency, hence lower performance.
A NUMA aware OS instead, would try to allocate the threads / processes always in the memory portion which is closest to the CPU which is going to process them.

 

[Action_Man]

Look at the 3D studio max 8 and cinebench 9.5 benchmarks you moron.

 

[accord99]

A few threads??? yahoo messenger has around 30 threads, the winlogon process can have 10, 20 or more ... FireFox with 5 open pages has 17 ... I'm quite sure any one of the multi threaded benchmarks spawns (can spawn) more than 8 threads. The system process in windows can have hundreds of threads ... A typical instance of SQL Server running on my dev. machine has close to 300 threads ...

Yes, but nearly all those threads aren't doing any work. They just sleep and occasionally wake up. It's rare that have you applications capable of splitting data to 8 threads and even a little bit of serial code means that scaling will suffer at 8 cores. Cinebench runs 1.8X faster with a second core, but by 8 cores is less than 5X faster.

As a software developer that works with databases every day I sure would like to see some test on database performance.

As far as I know the best threaded software and most scalable is encoding software so I'm expecting to see even worse performance for database tests ... also databases don't even require that much number crunching power they just need fast memory and disk access to scale well.

Enterprise databases like SQL Server are very scalable and should respond well to extra cores (assuming you have enough memory and a fast enough storage system). An HP ML370 scores 42% higher in TPC-C just by replacing the 3GHz Woodcrests with 2.67GHz Clovertowns.

 

[Pippero]

Look at the 3D studio max 8 and cinebench 9.5 benchmarks you moron.

Was it necessary to call him a moron?
This thread had been pretty civil and peaceful... until now. :roll:

 

[Action_Man]

Yes, this thread is stupid and trollish and he has a history.

 

[Heyyou27]

Look at the 3D studio max 8 and cinebench 9.5 benchmarks you moron.

Was it necessary to call him a moron?
This thread had been pretty civil and peaceful... until now. :roll:He did start off by saying that AMD's 4x4 platform has increased performance, while refusing to acknowledge Intel's Core 2 platform.

For the record to "Cryogenic", the Intel Core 2 Extreme QX6700 performs better than the FX74 pair while also consuming a much smaller amount of energy.

 

[Pippero]

So if i say something wrong about C2D, it's not like i'm personally insulting you, is it? :?

For the record to "Cryogenic", the Intel Core 2 Extreme QX6700 performs better than the FX74 pair while also consuming a much smaller amount of energy.

Then maybe i should call you moron for saying this.
The point of this thread, that you completely missed, even if the poster's conclusions were wrong, was about the somehow disappointing scaling of the Core Quad architecture in a dual socket solution.
The fact that it consumes much less power than the FX74 and outperforms it, has everything to do with the uarchitecture of both CPUs, and little to do with the platform itself.
In the referenced article from Tom's HW, the quad gets beaten by the dual (albeit higher clocked) core in most of the tests.
An exception is, as Action_man said, Cinebench and 3DSMax.
However it's interesting to notice that Cinebench is the only test where the FX74 outperforms C2Q, and in 3DSMax it is also extremely close in performance to it (under Vista, according to this link that Action_Man posted himself.)
Clovertown also clearly wins in 2 other tests, Linpack and Sungard.
Now Cryogenic's conclusions were questionable and fanboyish, and i've already expressed in another thread that i find this testing methodology quite questionable for a server/workstation CPU, but the thread itself presented some interesting topics for discussion.

 

[jkflipflop98]

We're at that junction again where we're waiting for software to catch up with the hardware.

 

[Action_Man]

Sigh sigh sigh.

So if i say something wrong about C2D, it's not like i'm personally insulting you, is it?

It makes you WRONG and when you post stupid threads like this one it makes you an idiot.

Then maybe i should call you moron for saying this.

Why because hes right?

Lock this stupid thread now.

 

[cryogenic]

Yes, this thread is stupid and trollish and he has a history.

A history of what?

You too have a history hero you know, a history of insulting everyone you don't share an opinion with!

Sorry man if this thread offends you but it's not stupid or pointless!

 

[bullaRh]

i havent been here so long but action_man always insult people, thats atleast what i have seen. i guess the kangaroo's in his yard pisses him off

 

[Action_Man]

A history of what?

Trolling, being a moron, etc etc.

You too have a history hero you know, a history of insulting everyone you don't share an opinion with!

Or people who are too stupid to see that they're wrong.

 

[Action_Man]

Another stupid American with lame kangaroo jokes, whats the deal with that? That lame ass crap doesn't offend or do anything except make you look like more of an idiot.

 

[clairvoyant129]

Intel just managed to get a negative performance increase by doubling the number of cores from 4 to 8 on 'multi threaded software'. Can please someone explain to me how is this possible ??? Someone must have screwed up big time, but that just doesn't quite say it.

Even worse the negative performance increase is achieved by using Intel's best architecture, Core 2.

At this point I'm quite sure Intel's 45 nm upcoming chips won't make much difference in 2007, the FSB it's killing them.

And I thought AMD 4x4 launch was poor ... well actually 4x4 managed to bring positive performance increase ...

Share you thoughts on this with me, please!

No you AMD fanboy. That's why a FX-62 outpaces a FX-74 in single threaded applications :roll:

 

[clairvoyant129]

Another stupid American with lame kangaroo jokes, whats the deal with that? That lame ass crap doesn't offend or do anything except make you look like more of an idiot.

I agree.

 

[cryogenic]

Intel just managed to get a negative performance increase by doubling the number of cores from 4 to 8 on 'multi threaded software'. Can please someone explain to me how is this possible ??? Someone must have screwed up big time, but that just doesn't quite say it.

Even worse the negative performance increase is achieved by using Intel's best architecture, Core 2.

At this point I'm quite sure Intel's 45 nm upcoming chips won't make much difference in 2007, the FSB it's killing them.

And I thought AMD 4x4 launch was poor ... well actually 4x4 managed to bring positive performance increase ...

Share you thoughts on this with me, please!

No you AMD fanboy. That's why a FX-62 outpaces a FX-74 in single threaded applications :roll:

This is not the case on single thread applications running on multi core CPU but the case of multi threaded software that actually benefits from multi core getting a negative performance increase by doubling the number of cores, which is an entirely different issue.

And don't call me a fanboy!!

 

[1Tanker]

Intel just managed to get a negative performance increase by doubling the number of cores from 4 to 8 on 'multi threaded software'. Can please someone explain to me how is this possible ??? Someone must have screwed up big time, but that just doesn't quite say it.

Even worse the negative performance increase is achieved by using Intel's best architecture, Core 2.

At this point I'm quite sure Intel's 45 nm upcoming chips won't make much difference in 2007, the FSB it's killing them.

And I thought AMD 4x4 launch was poor ... well actually 4x4 managed to bring positive performance increase ...

Share you thoughts on this with me, please!

No you AMD fanboy. That's why a FX-62 outpaces a FX-74 in single threaded applications :roll:

This is not the case on single thread applications running on multi core CPU but the case of multi threaded software that actually benefits from multi core getting a negative performance increase by doubling the number of cores, which is an entirely different issue.

And don't call me a fanboy!!Good Luck with that.

 

[cryogenic]

Intel just managed to get a negative performance increase by doubling the number of cores from 4 to 8 on 'multi threaded software'. Can please someone explain to me how is this possible ??? Someone must have screwed up big time, but that just doesn't quite say it.

Even worse the negative performance increase is achieved by using Intel's best architecture, Core 2.

At this point I'm quite sure Intel's 45 nm upcoming chips won't make much difference in 2007, the FSB it's killing them.

And I thought AMD 4x4 launch was poor ... well actually 4x4 managed to bring positive performance increase ...

Share you thoughts on this with me, please!

No you AMD fanboy. That's why a FX-62 outpaces a FX-74 in single threaded applications :roll:

This is not the case on single thread applications running on multi core CPU but the case of multi threaded software that actually benefits from multi core getting a negative performance increase by doubling the number of cores, which is an entirely different issue.

And don't call me a fanboy!!Good Luck with that.

Hey I'm not a AMD fanboy, I've owned several Intel PC and only one AMD it was a K6-III, even now my rig is on Intel ... no matter how much this upsets you Intel true fan boys there's something wrong in the Clovertown picture and I was hoping to get some smart answers not this.

I agree servers should be tested using server software but I don't see any reason why a perfectly good threaded application that takes full advantage of multicore should not receive any kind of advantage by doubling the number of cores....

 

[Action_Man]

If the slower clocked cpu with more cores isn't using all those cores because the app doesn't have that many threads then *GASP* its going to be slower!

Notice how with 1 cpu the clovertown is slower but with more threads *GASP* its faster! OMG NO WAI!

 

[bullaRh]

im scandinavian which is in europe :roll: take it easy dundee and go catch some crocs insted of sulting peoples topics and call em idiots