Hyper-threaded dual core processors for notebooks?

[Scali2]

An actual quad is always going to be faster than a dual+HT.

Only if they use the same architecture.
The Nehalem architecture is considerably faster than the Penryn one.
Add to that the fact that you generally don't get super-good scaling out of extra cores, especially past 2 cores, and two fast cores + HT may just be faster than four slow cores.

 

[grimreeferx09]

for me, a laptop just needs 2 gigs of ram and an nvidia 8400 and a decent dual core cpu.
just enough so i can watch 1080p videos with haali renderer.
anything more taxing needs to be done on a real computer aka desktop.

 

[ilkhan]

Only if they use the same architecture.
The Nehalem architecture is considerably faster than the Penryn one.
Add to that the fact that you generally don't get super-good scaling out of extra cores, especially past 2 cores, and two fast cores + HT may just be faster than four slow cores.

Well in my mind the question here is a dual core nehalem with HT on vs a quad nehalem with HT off or a yorkfield. Either way, the quad nehalem with HT off is going to be fastest at any given clockspeed.

 

[Scali2]

Well in my mind the question here is a dual core nehalem with HT on vs a quad nehalem with HT off or a yorkfield. Either way, the quad nehalem with HT off is going to be fastest at any given clockspeed.

That's not what the opening post said.
Besides, what sense would that make? Nehalems all have HT, so even if you have the quad, you can still turn on HT if you like.

 

[Digital Viper-X-]

Seriously, Nehalem is what 10-15% faster then a Penryn at the same clock? and you think that a dual core is going to out perform a quad where multi-threaded apps are involved?

 

[Scali2]

Seriously, Nehalem is what 10-15% faster then a Penryn at the same clock? and you think that a dual core is going to out perform a quad where multi-threaded apps are involved?

You can't compare them directly.
And yes ESPECIALLY when multi-threaded apps are involved, the difference will be smaller, as I already explained.

 

[Digital Viper-X-]

You can't compare them directly.
And yes ESPECIALLY when multi-threaded apps are involved, the difference will be smaller, as I already explained.

yes, but even in multi-threaded applications the Core 2 Q and the i7 are not far enough apart to justify what you're saying

 

[Scali2]

yes, but even in multi-threaded applications the Core 2 Q and the i7 are not far enough apart to justify what you're saying

Depends on the application, get it?
Look here for example:
http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3448&p=17

Especially in Pov-ray, the Core i7 is an incredible deal faster than Core2.
If you take the 965 score and divide it by 2, you'd have a very pessimistic estimate of the performance of a 3.2 GHz dualcore Nehalem.
This would result in a score of 2101. This is only a smidge slower than the 2.66 GHz Q9450 at 2238.
It's quite possible that a real dualcore + HT at 3.2 GHz will actually score more than 2238 (scaling from 2 to 4 cores is never going to be 100%, so the score for a dualcore will be more than half of a quadcore).

So if you would have a choice between a higher-clocked Nehalem dualcore + HT or a lower clocked Penryn quadcore for Pov-ray... it may just be possible that the dualcore + HT is faster. It may even be possible that the dualcore actually uses less power.
QED

 

[CamaroZ28]

Depends on the application, get it?
Look here for example:
http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3448&p=17

Especially in Pov-ray, the Core i7 is an incredible deal faster than Core2.
If you take the 965 score and divide it by 2, you'd have a very pessimistic estimate of the performance of a 3.2 GHz dualcore Nehalem.
This would result in a score of 2101. This is only a smidge slower than the 2.66 GHz Q9450 at 2238.
It's quite possible that a real dualcore + HT at 3.2 GHz will actually score more than 2238 (scaling from 2 to 4 cores is never going to be 100%, so the score for a dualcore will be more than half of a quadcore).

So if you would have a choice between a higher-clocked Nehalem dualcore + HT or a lower clocked Penryn quadcore for Pov-ray... it may just be possible that the dualcore + HT is faster. It may even be possible that the dualcore actually uses less power.
QED

Heh... I wouldn't say it "may be possible" the dual core would use less power, its pretty much guaranteed the dual core will use less power...