Would a slower quad core cpu generate more heat than a fast dual core?

Would an i7-720QM (1st generation quad core) at 1.6 GHz generate more heat than an i5-560M (1st generation dual core) at 2.66 GHz? Help would be appreciated because I want to upgrade my cpu on my vaio f series but don't know if it would make a lot more heat.

 

yes, the quad parts are rated for 45w, agains the 35w from the dual cores

 

Although the TDP in the i7 is higher and will run hotter, the difference isn't egregious.

 

However, it is worth checking whether your laptop can handle it, some laptops are designed in a way that barely dissipates the heat from a 35W CPU, others have a lot of headroom and can easily take higher TDP parts.

 

The vaio's are fairly good systems, but yes thanks for supplementing that. Definitely worth taking note.

 

The load is the same really in my experience, a given task requires X amount of processing power and if you split it between 4 cores working 1/2 as hard as a dual core it balances out.

The key is the quad core CAN get hotter but it can handle more load and get a given task done in less time.

 

And the flip side is that with more power - it won't go unused (your workload will increase to take advantage of the higher performance).

So yes; the chassis probably won't take the increased heat well and could actually give you a system with less performance than the one it was designed for originally.

 

I agree with most everyone here but want to input that; Unless running apps designed to run more than two threads or running several active apps to make use of more than two cores the dual core will be substantially faster. And active does not mean iExplorer and photoshop open in the background doing nothing. Open apps in the background are Memory intensive not CPU intensive.

Check if that Sony model offered quad as an option. While some may have offered different cooling solutions I think most kept it simple and made CPU a drop in mod to keep it simple and costs down.

 

Keep in mind that even if the swap is feasible, it won't be much of an upgrade. A 560M is much faster than a 720QM on any workload with 1-2 threads and slightly faster on workloads with 3-4 threads. It's only when you hit 5+ threads that the 720QM pulls ahead.

 

You might as well jump to the i5-580m or i7-640m if possible. If the i5-560m is the only choice then thats good too. I would choose the i5-560m for the reasons I'll mention shortly. NOTE: the following comparison is only possible because the i5 and the i7 have identical architectures with the only differences being core counts and clockspeed.
Basically, the i5-560m and the i7-720qm have the same theoretical peak performance. Hypothetically, assuming there is a program that issues one instruction per Hz and perfectly scales with cores, the i5-560m would output 3.2 billion instructions per core, and 6.4 billion instructions with both cores.
Doing the same math, the i7-720qm would issue 1.6 billion instructions per core and 6.4 billion instructions with all 4 cores.

However, in the real world, the i7-720qm is very unlikely to reach its maximum potential due to the difficulty in scaling a program's performance between cores, otherwise known as Amadahl's Law Amdahl's law - Wikipedia, the free encyclopedia. The i5-560m is much more likely and easier to program to reach its peak performance therefore surpassing the i7-720qm even in multithreaded apps due to the higher clockspeed (which improves performance irrespective of programming).
The scaling issue is influenced by many things of which a discussion would span several pages. A simple example is the sharing of a single piece of data in the memory between threads running on different cores. The analogy is bankers accessing a single money vault, a quadcore would have to divy up access to that vault between 4 operators vs 2 in a dualcore. The performance hit is obviously greater if you have 4 cores than 2 cores.

Since much of the i7-720qm's performance is derived from how well threads scale with the core count, the i5-560m is much more consistent as its performance is derived from clockspeed which has near perfect scaling. And should there be a worst case scenario if you have a program which doesn't scale well with core count (i.e. older games), the i5-560m takes a huge lead with the higher clockspeed on single cores.

The last consideration is the power consumption. The i5-560m has a magnitude of lower power consumption than the i7, this is because the i5 is manufactured on the 32nm production vs the 45nm process of the i7-720qm. This means better battery life and a reduction in at least 10 degrees of temperature.

As an interesting point to note, the i7 is rarely able to exploit its boost clocks on 2 or single cores for some reason which baffles me. The i5 actually has an unlocked TDP/TDC parameter which can be set in Throttlestop to essentially enable unlimited Turboboosting.

 

You could of just clicked that you liked the last two posts that kind of said all that stuff you said? You do know I am joking. You did put out a good read. I did learn a little some things I might of known but maybe not why.

I will ask two questions. Do all quads suffer from the lack of effiency you mentioned. Second your final two sentences talk of the boost. My ULV can run at 2.8GHz all day can run at 3GHz if single thread. That is really the CPU's we are talking about or inherent in increased threads?

I am really asking as you seem to have some knowledge.

 

All existing and future multicore processors will suffer from some degree of inefficiency as not all programs/workloads can be parellelized (e.g. a program with lots of conditionals where the outcome of one thread depends on the outcome of a preceding one). Technically, dualcores are also affected but the magnitude is minimal compared to quadcores. The degree of inefficiency also gets worse as more cores are added.
The memory example I cited will be addressed in part by the Haswell architecture with the new TSX memory but this is one among many inefficiencies.
However, program scaling with clockspeed used to be limited by memory bandwidth but this issue all but vanished with Dual Channel RAM. Remember, in the old days, you had to OC the RAM in addition to the Pentium 4 to gain 100% scaling. Unfortunately, clockspeeds have pretty much peaked now with Ivy Bridge as higher frequencies require an exponential increase in voltage. This makes it too heat and power prohibitive. I won't go in to IPC as it is getting quite boring now.

The last comment I made specifically applies to Arrandale generation and Clarksfield XM chips as they have unlocked TDP/TDC. As you know, Turboboost is determined by 2 primary parameters. The TDP (heat output in W) and the TDC (current drawn in amps). Clarksfield and Arrandale had a 45 and 35W TDP limit but also a TDC limit of 37-40A. This meant Clarksfield virtually never could boost aggressively as it was already so close to the TDP/TDC limits at stock. The Arrandale chips were built on 32nm so TDC draw wasn't as restricted (still occasionally annoying) but they continually bumped against the TDP limit.
All this changed with Sandy Bridge, the chips had a very high TDC limit (so essentially TDC is unlimited) but the TDP limit stayed. As the Sandy chips were more power efficient due to the matured 32nm process and the Turbo limit was very high even for all cores, the Sandy chips had two boost modes, the second of which had a much higher TDP for a shorter duration thus allowing the chip to boost to a very high peak before settling down as it got hotter.
This two tiered system was no longer necessary with Ivy Bridge, as there was less heat output due to the 22nm process. A single TDP limit was placed coupled with a high clockspeed cap but the low heat output of the Ivy cores meant that perpetual turboboost was possible. Some manufacturers exploited this headroom for better performance by using good cooling but some used less cooling (therefore less turbo) to enable thinner designs.
Essentially, the above was a long winded way of saying your Ivy cores are not TDP limited at 2.8ghz (all cores) or 3ghz (Single core) due to their high efficiency and your superb cooling system.

 

Wow. I will reread both those posts again.

Do you know coding? I mean can you tell this app or that app is single threaded or dual or quad? I heard video is hard to multi thread blah blah. I just ask because 90% of us are idiots and the other 10% can't use computers. Kind of an odd mix.

Thanks for all the input. It really is very interesting.

 

With first gen i3/i5/i7 and previous, most of what Marksman30k holds true.

With Sandy Bridge and Ivy Bridge and likely forward, dual core won't have any advantage over quad core other than a lighter TDP and slightly less power consumption at load (well, cheaper too). Even the lowest end quad core can boost to similar speeds to the high end dual core with 1 or 2 threads. Also with quad cores, if there's any other background tasks running, Windows is good at separating the tasks, so the application will get full use of one or two cores while other processes are run on the other two cores.

Also note that most newer apps and utilities and games are being programmed for multiple threads, so while older software may be single or dual threaded, it's less of a factor with each iteration of CPU since the performance is so high to begin with.

For gaming, it won't matter much at all. See this Core 2 Duo 2.8GHz vs Core 2 Quad 2.0GHz gaming comparison I did with LaptopNut a while back:

http://forum.notebookreview.com/gam...er-np8662-comparision-core-2-duo-vs-quad.html

Even with a 40% increase in clock speed, the quad core was as good as or better than the dual core, even games single or dual threaded.

 

Hahahaha, unfortunately no, my coding knowledge is limited beyond running said app and checking the CPU usage graph. HtWingnut is right, modern Quads can boost to the same clocks as the duals so the performance difference is not as big anymore, this was the biggest thing Ivybridge brought to the table. The Ivy cores are so efficient that power consumption was below the TDP even with 4 cores at high clockspeeds.

 

Wrong wrong wrong. Your test is silly for a CPU. How about you try and isolate your CPU. If you believe your 2GHz quad beat a 2.8GHz in a dual or single threaded app you are simply wrong. I mean it is so wrong as to defy logic. The problem is your tests did not ever isolate your CPU. I am sorry you almost say you and I drove to Flint. Same road. I drove 50mph you drove 40mph and you got there first.

 

No his single thread or dual with a quad @2GHz vs a dual @2.8GHz is just wrong.

Sad you don't know coding. I hear proper execution relies on well written code and I hear that is easier said than done.

 

I said for GAMING, please read my post. And I quote what I wrote again right here.

And look at Ivy Bridge mobile. Lowest end quad core is i7-3630QM, with single or two core boost to 3.4GHz. Highest end dual core is i5-3380m, 3.8GHz peak turbo. Next bump up is i7-3740QM at 3.7GHz, so the difference is minimal. I'm doing a comparison at the moment (taking me a lot longer than planned due to other projects going on), comparing i5-3360m vs i7-3610QM for gaming and CPU performance intensive apps (Adobe Premier Pro, Blender, cinebench, 7-zip, handbrake, etc) to compare performance.

 

I apologize I guess I missed.

 

It's alright,low thread count execution speed wouldn't matter much these days anyway since ivy duals and quads boost to the similar clocks even on single or dual threads. Yeah in all practicality, games are more multithreaded now so the cpu impact is minimal
Well executed code can indeed overcome many issues including multi core scaling inefficiency or requiring high clock speeds bit it's technically difficult. I have friends who slave away for days debugging single thread creations let alone multithreaded.