Intel i-7 620m or i-7 720qm

Now, given the choice of the i5-520 or i7-720 I'd much rather have the quad.

You could always buy an i5 on ebay and put it in yourself (and you can sell the 720 as OEM cpu's are an easy sell).

 

I'd get the 520M over the 720QM any day too.

 

I'd rather have the i5-520M due to lower power consumption, but if I was offered an i7-720QM for the price of an i5-520M I'd take it and sell it.

 

between the i5-520M and i7-720QM, i will go for the latter.

However, between the i7-720QM and the i7-620M, i would really really have to think as to what i value more.

For me the decision shouldnt be difficult, as stormeffect points out, the market essentially forces you! so, just get forced, since you dont have many options!

 

It's funny because these new cpu's from Intel are so darn good that even the mid-range stuff is powerful enough for most people's uses.

 

Have a look at Shirleys benchmark of the i7 620M .

http://forum.notebookreview.com/showthread.php?t=455347

 

Even the "lowest end" i3s are good enough for more than 90% of the market...

 

I think that's why they have the Atom?

 

well trust me , Atom doesn't cut it compared to an i3.. even a core 2 solo is better than a atom and a dual core one...

 

well that makes the QM look totally irrelevant.

and with the i5 being only about ~12% behind the i7 dual, seems to me that the best buy really is the i5. nominal speed loss, significantly less power usage/TDP.

interesting.

 

Have you played GTA IV or Flight Simulator X?

 

also, my understanding is that Adobe Creative Suite 4 and 5 take advantage of more than 2 cores if available.

but to be fair, that leaves the other 800,000 windows applications that are NOT designed to take advantage of more than 2 cores (or more than 1 core, for that matter).

so really, unless you're using apps (1) are confirmed to support 4+ cores and (2) you push those apps to their limits on a somewhat regular basis, there's no need to buy into quad core in 2010, just like it made no sense to buy into AMD x64 processors back in 2006 (not because 64-bit wasn't the future, but because its standardization was so far off, by the time it hit mainstream, those processors would be too slow to mention. same here)

 

Using 4 cores is not the same thing as using 4 cores efficiently, which you have to do for a 1.733GHz quad-core to be at all faster than a 3.067GHz dual-core.

I'd buy a quad-core on a desktop assuming it fit the budget, likely the Core i5-750, but it's clocked at a much faster speed (with Turbo) of 2.8GHz on all four cores or 3.2GHz on two cores, which is also quite competitive with the faster desktop dual-cores. Consequently, there's no real downside to the desktop Core i5-750. However, because of power consumption and low clock speeds, I don't see the point of the i7-720QM in laptops. The i7-820QM and i7-920XM are more justifiable on specs, but then the cost is too high.

 

Thank the lord that i have a core each for each of my eve online clients !

Quad ftw.

 

No, that's wrong, because the i7-720QM is weaker than the i7-620M overall.

CPUs are always about paying lots of money for slight increases in power, especially at the higher end. However, if you've got much more money than time, you could justify the steep costs for, say, the i7-820QM or even the i7-920XM. However, the i7-620M is almost always a better pick than the i7-720QM.

 

If money is of no concern than the 920XM all the way

 

You can get the MSI GE600 with an i5 and HD5730 that supports switchable graphics.

 

Read this on Fudzilla regarding i7 940XM that made me smile


Intel is preparing to launch yet another extreme mobile processor, something to top the current performance king the Core i7 Extreme 920XM. This CPU is an upgrade for Calpella high end performance notebook and it comes in Q3 2010.

The clock speed stops at 2.13GHz but this quad core automatically overclocks all the way to 3.33GHz. This will only happen if it doesn’t get that hot. TDP remains at high 55W and this quad core still supports the quite useless Hyperthreading.

This Clarksfield 45nm CPU comes with 8MB cache and 1333MHz DDR3 support. The only ironic thing about this CPU is that 2.66 GHz Dual-Core Core i7 620M that will overclock to 3.33GHz might easily end up as fast as the much more expensive Core i7 Extreme 940XM that will launch soon.

 

Well, for quad-friendly applications like video encoding, the 940XM will be a lot faster. I'd expect the 940XM to reach 2.4GHz on four cores with turbo boost, which is a lot more raw power than 3.067GHz on two cores; you just need applications that are able to take advantage of it.

 

it's time to acknowledge that there are two directions for performance in modern day processors. One is linear performance, the other is LATERAL performance.

The arrandale series has an edge in linear performance. it's clock speed edges out the clarksdale [as has already been mentioned] in single and dual thread applications.

Where the 720QM really shines is not in multi-core, but multi PROGRAM. Sure, not many programs take advantage of more than 2 cores, but what if I'm running a game one one monitor while watching a movie on the other while converting a movie for my iPhone? At this point the performance gains of the 720 become obvious [I say this from first hand experience]

And for the record, Ghz is a very poor way of assessing processor performance. We've had 3 Ghz processors for years and years now, why is the 3 Ghz P4 not performing the same as an i5 running a single core app @ 3.06?

Two different processors, for two COMPLETELY different applications. Assess your needs and use the correct tool for the job. . and stop arguing about which is better. We all know the answer to the question is either:

"the one you DIDN'T buy"

or

"the one that was released the week after you bought yours"

I believe you meant to say "still supports the quite useless TO YOU Hyperthreading". ^^

 

Hopefully Intel is diabolical enough to release a faster dual-core as part of their Calpella refresh so that we can have one final round of dual vs quad (before Sandy Bridge comes out) by comparing an i7-6x0m vs. i7-820qm.

 

You're forgetting that better linear performance also translates directly into better lateral performance.

You're neglecting the fact that if the processors can do the same amount of work per clock cycle per core, which is reasonably accurate for Arrandale vs Clarkale, a single core running at 3GHz can do the same amount of work as two cores at 1.5GHz, regardless of what your configuration of processes and threads is. Consequently, the i7-720QM still has little to offer over the i7-620M.

Consider a multitasking scenario where you have one application that would benefit from around 2.5GHz of processing power on each of two threads, and another that only needs 0.5GHz on two threads. If you're using 4 cores on the i7-720QM, the first application would be stuck with 1.73GHz at 100% utilization of two cores, while the second application would only be using ~30% on the other two. On the other hand, with two cores at ~3GHz you can allocate most of the time on each core to the first application, and the rest to the second, and have overall better performance. The fact is that it is actually much easier to distribute processing power between applications when you have less cores.

While GHz is meaningless if you want to compare processors with vastly different architectures like Pentium 4 vs Nehalem, clock speed is pretty much all you need to look at when comparing one Arrandale against another (though there are slight technological differences like AES-NI and VT-d). Similarly, Clarkdale's architecture is very similar to Arrandale's, so it makes perfect sense to compare the two based primarily on clock speeds.

 

It is not 3.33GHz on two cores, only on one. 3.06GHz on two

http://en.wikipedia.org/wiki/List_of_Intel_Core_i7_microprocessors#.22Arrandale.22_.28standard_voltage.3B_32_nm.29

 

Gah, I've gotten it right plenty of times before. Fixed now, anyway. Thanks.

 

Looking at some benchmark done with different M15x, I'd say the i7 620M performs better for gaming, but the difference is very small. Running with an overclocked gpu, the difference is about 200 points on 3dmark06...

 

For me there is one must and one soft reason to go with the i7-720 instead of the i7-620:

1) Most 15" notebooks offer 4 SO-Dimm when you take a quad core configuration, the dual core models only have 2 SO-Dimm. So while performance be on par or slighty towards a i7-620, I need the i7-720 to get the 4 SO-Dimms for more ram.

2) The i7-620 (max/min 64.7W/26.5W) seems to consume more (!) power than the i7-720 (55W max/15.4 min) as measured by this guy: ***Klick***

 

I'm doubtful about the power consumption figures; I'd want to see both CPUs tested in the same system before I'd consider the results to be conclusive one way or the other.

As for the 4 vs 2 SO-DIMMs, I don't think that's altogether as common as you claim. There's not that many notebooks I've seen where the manufacturer changes the motherboard between the dual- and quad-core configs. In fact, the only ones that come to mind are HP models - the Envy 15, and probably the EliteBook 8540w in your sig.

 

I believe notebookcheck updated their power consumption figures on the German site.

 

You're right.
It looks like the i7-620M is pretty close to the i5-520M after all in power consumption; we were right to be skeptical. Sadly, that still doesn't tell us how the i7-720QM compares.

 

mate you could check on the official site for the right specification, here is the link if you want to check it out
http://www.intel.com/products/processor/corei7/mobile/specifications.htm

 

Intel only specifies TDP, which isn't that useful overall because it doesn't really tell you anything specific about how much power would be consumed for various ways you might want to use the CPU.

 

In some systems, the difference is still significant. 1-5W on idle dif and when the T400 for example consumes <10W of power on idle overall...

 

I'd rather see proper battery life testing, though. The problem with figures like idle min/max is that they don't tell you how much time the CPU spends towards either end of the spectrum.
The other problem is we still can't say for sure how the i7-720QM compares.

 

Thats becaues TDP doesn't measure power consuption. It's a measure of heat output. 35W TDP is the maximum amount of heat that manufacturers will need to dissipate if including this CPU in their laptop.

 

A CPU is a device that turns electrical energy into heat, and happens to do some useful shuffling around of bits of information on the side.
The problem with TDP is not the distinction between power consumption and heat output, because all of the power becomes heat anyway, but the fact that TDP is measured under "typical operating conditions" and we don't even know what those are.

 

TDP is the maximum amount of heat the processor can withstand.

something along the lines of:

TDP = total heat generated - heat removed by the heatsink

basically, TDP is the maximum amount of heat that can 'stay inside' the core.

 

That is simply not true. You're confusing power with energy - Watts are a unit of power, which is energy per unit time.

If the heatsink isn't removing all of the heat generated by the CPU, the CPU will just keep getting hotter. You might then ask why the CPU can stay at a certain temperature under load, and why this is different to the temperature at idle.

The explanation is as follows - the rate of heat transfer away from the CPU is proportional to the temperature difference between it and its surroundings (this is also why the ambient temperature is important for heat measurements). Given a certain level of heat output from the CPU, the temperature will reach steady state when the CPU reaches a temperature such that the rate of heat production by the CPU is equal to the rate of heat transfer away from the CPU. Obviously, the more heat it generates, the hotter this steady state temperature will be.

TDP is "Thermal Design Power" - basically, it's meant as a reference figure that represents how much heat you have to account for in typical usage of the CPU. However, there is no specific set of reference conditions under which TDP is measured. According to the datasheet for Arrandale,

 

To do so, at what utilization would the i7-620M be operating at...? 100%? Seem like advantage i7-720 if so.

I didn't have time to labor over this choice when I was ordering my HP 8740w, but I labor over it now ... So when you say "NBR" does that stand for notebookreview, i.e. this site? Hope I find them if so. I might have time to change my order.

I will say my main reason for buying that machine is for Lightroom pp of RAWs, and Lightroom definitely does well with quad core. But I really worry over the much lower GHz the quads run at for processes that are single threaded.

As cited above, Adobe Lightroom as one example does utilize the cores > 2 very well.

Now, it's worth considering this scenaro. No, this is certainly not "a [single] program". I'm a software developer. In the course of developing 4-tier enterprise applications a "client request" will use GWT in the browser, async WebServices client calls using AJAX calling to WebService servlet handlers running in Tomcat, that communicate to both MSSQL and Oracle concurrently.... all running on my notebook / mobile workstation. It's when many parallel async calls are made concurrently from the client that the multiple cores would really begin shining. I think I just talked myself back into sticking w/my new quad-core equipped machine.

 

Yes, 90% utilization over 4 cores of the i7-720QM is approximately equal to 100% utilization over 2 cores of the i7-620M, as far as raw processing power is concerned. Alternatively, the i7-720QM has around 13% more raw processing power in total.

The thing is, the more cores you have, the harder it is to use them efficiently. Very few applications will be twice as fast on a quad-core as they are on a dual-core with the same clock speed.