How much worse is i7-4702 compared to i7-4700?

When I ordered my system it was advertised as having an i7-4700 MQ which runs at 2.4ghz and goes up to 3.2ghz with TB in quadcore mode. Instead I got the i7-4702 MQ which is meant to be a more power efficient but slower version, it also runs at 2.2ghz and only goes up to 2.9ghz with TB in quadcore.

How much worse is it and is there anyway round the fact that it doesnt go up as high with Turbo Boost? The extra speed matters in my case as I like to play games using PCSX2 and Dolphin.

 

I doubt there'd be much difference depending on what speed the 4700MQ throttles at. Try using Intel XTU to get another 200MHz out of it. Otherwise the performance difference is minor.

 

Like 5%. The only thing that the 4700 has going for it is that it can be overclocked to 3.6ghz.

Both CPU's have the same idle power consumption, the only difference is that the 4702 will run much better when under load.

If your 4702 can't hit 3.2 under max load it means that you either have multipliers down or you are throttling, in which case i would even advise downgrading to the i5-4200M. PCSX2 and Dolphin (i play both) are very GHZ dependant, and every 0.1ghz can give a noticeable boost in framerate, hence i suggest having a 3.0 ghz processor or higher. Also, remember to let your laptop's dGPU (if any) handle the graphics for these emulators, otherwise you will see lots of lag due to the integrated graphics adding extra heat (and throttling) to the main CPU fan.

Source: 4700 Owner, PCSX2 and Dolphin gamer

 

PassMark shows that your i7-4702MQ delivers 96% of the i7-4700MQ's performance:
PassMark Intel vs AMD CPU Benchmarks - High End
I go by the 10% rule: you'll notice a difference if it's greater than 10%. Even if we did straight arithmetic, 2.2GHz is about 92% of 2.4GHz - still not going to be noticeable. Using those numbers in perspective - if the 2.2GHz chip took an hour to do something, the 2.4GHz chip would do it in 55 minutes.
Note that the chip will most likely be unable to sustain the full turbo boost clockspeed for extended periods; it's not meant for that, just the short spikes now and then. The CPU engages Turbo Boost only when thermal conditions permit. Additionally note that the actual Turbo Boost is dependent on the number of cores being utilized; four cores under load are going to run at less frequency per core than if one just core were loaded. The 3.4GHz maximum Turbo Boost number for the i7-4700MQ is assuming perfect conditions and probably for one or two cores only; four cores would be a few hundred MHz less. Still much better than nothing.

I wouldn't lose sleep over it. Enjoy the game & the system.

 

Thanks for the replies guys, Ive got the laptop now but even then it was by far the best available for its price. What advantages are there with a lower TDP, battery life is hardly an issue for me but does it mean it runs cooler?

Also PCSX2 has run well for me so far but just how long can a system maintain TB?

 

For emulator gaming, i advise getting Intel Extreme tuning utility. First of all, max out your first multiplier to 32x, and then lower your 3 other multipliers to 22x. This will give you the best performace in games like PCSX2. Also, the reason why the 4702 runs better than the 4700 under load is because it uses less power and thus doesnt generate enough heat. The lower wattage also allows your laptop to use a smaller power supply. Undervolting also helps, i can explain it into more detail if you want.

The biggest killer for emulators is integrated graphics, because they often do not support advanced textures well and reside on top of the CPU, generating extra heat making the computer throttle faster. Even at the lowest settings, which puts no strain on the GPU, there is a day and night difference in FPS between my 4600HD and 755M found on my Lenovo Y410.

 

IIRC the 4700 and 4702 are about the same price so it should not impact the overall cost of the laptop if it uses the same cooling for both CPU's.

Try running Linx with 4 threads using Windows task manager to set affinity to separate cores i.e. 1,3,5,7. Use a small problem size of 10000 so runs will only take ~5 seconds and select 3 runs. Check your temps while running this with RealTemp and see what you get.

Running the extra 2 bins is a mute point since it can be done on both CPU's. If the 4702 does run a lot cooler for the same load then IMHO that would be a good advantage.

AFAIK technically both CPU's have the possibility to run over 4GHz but have not seen that done yet..

 

Thanks for the info, really interesting stuff. For now its not a huge concern but from what Im hearing my the 4702 will run cooler with less power when its on TB max? Also just how save is OCing for laptops, AFAIK they have much lower thermal envelopes than desktops and can apparently lead to great wear-and-tear.

XTU says that the multiplier is already at 32x and can go higher with the cores at 32, 31, 29 and 29.

 

It runs cooler with less power because it runs slower. If you ran the 4700MQ at the same clocks it'd likely produce the same heat and draw same power.

 

Well if you run the Linx test I posted earlier that should give some idea. Haswell does seem to run hot though so might be quite limited. Here's some results of the 4700MQ at just 3.4GHz.

Obviously you have to be careful not to put too much current through the chip. How much you can achieve will vary from one system to another.

Here's a P8400 hardware overclocked from 2.26GHz to 3GHz a few years back and is still working fine today.

 

Please do not rely on Passmark benchmarks.They're synthetic benchmarks and don't give accurate results.
Interpreting PassMark Results [Solved] - Intel i7 - CPUs

 

Define accurate?

And how do you measure relative performance w/o synthetic benchmark...?

result from app benchmark would vary from each other, which one is accurate?

 

Thanks again for the info guys, for now Im happy with the chip as it is but OCing or maintaining TB will definitely be a priority for me soon. Ill keep this thread bookmarked as plenty of great stuff here!

I know this may have been answered but just for clarification, will the 4702 run cooler at the same speed as a 4700 since it has a lower TDP?

 

At the same (real world) workload, both processors will run at the same temps*. Only when pushed to the limits will the TDP come into play (and only then it will still depend greatly on the cooling design (and effectiveness) of the notebook chassis in question).




*All else being equal.



Take care and Merry Christmas.

 

Why don't you run the test I posted earlier and see for yourself how hot it runs?


Merry Christmas to you too tilleroftheearth and all.

 

From Dictionary.com

Instead of synthetic benchmarks you use games or utility programs like 7-zip.

Results from Passmark would be inconsistent because each "sample" is used in a different system with different OS, RAM, etc.

 

It will leave more thermal limit for the GPU to use. In this case actually having the more power-efficient but slower chip will mean the overall performance will be better. All the 4-core i7s on laptops are fast enough for nearly anything by this point so you're better off with the thermal limit for your GPU.

Funnily enough the i7-4702 is actually a more expensive chip.

 

As TDP is based on case temperature then chances are the 4702 will run hotter than the 4700 clock for clock.

 

Games and utility only represent one kind of workflow and hardly representative for general performance.
And we yet to come across mobile cpu benchmark ran on exact same system. All mobile CPU test are run on different machines.

 

The 4700MQ will run at full turbo cores 3.4ghz across (when OC the 2 extra turbo bins with Intel XTU) all 4 cores when demanded by the application. I ran BF3 that way before I got my 4900MQ. The 4900MQ stock CPU settings are pretty much the limit as far as power goes and the unlocked 6 turbo bins are pretty useless IMO because it needs more power. Mine will run at 3.7ghz across all 4 cores when loaded in BF4 and pretty much just stay there until I close the game.

My guess is that with the lower power consumption of the 4702 won't allow it to use those unlocked turbo bins either and you will be stuck at its max or even below that max turbo speeds when loaded.

 

Nope. I have the CPU, and the stock multipliers are 34/33/33/32.

 

Yep. I had the CPU, and if you had read what I said it will run at 36/35/34/34 WHEN OCing the 2 turbo bins USING INTEL XTU..

 

Oh, haven't oc'ed mine due to it being useless since i stopped using my laptop as my primary. But thanks for letting me know, i might try it later on.

 

My guess is they would use about the same power clock for clock. Instead of guessing though it would have been nice to see real figures, maybe run stock settings with the processor locked to 2GHz using ThrottleStop and run a benchmark such as Linpack and see what the maximum reported power is then we all might learn something.

Why, because you don't want to run the higher power?