i7-3632QM questions

How did Intel manage to reduce the TDP by 10W and has anyone run tests with this compared to the 3612QM?

I want to reduce temps as much as possible while maintaining similar performance to the 3610QM, but I'm wondering if the 3632QM can handle 100% CPU Load with HT-on with Prime95 at Turbo frequencies( I assume 3GHz , since my 3610QM stays at 3.093 GHz during load) and still consume only 35W..

And does this 10W drop in Power necessarily mean a drop in temperature as well?

 

I don't have a link to results, but I do know the difference in temps is not significant between the two CPU's, if anything, 2-3C at best. TDP doesn't necessarily mean less heat, although it technically should. It just depends on what threshold Intel decides to bin a CPU at. TDP is a range, and Intel classifies some chips that are within that range, not exceeding 35W TDP, where the 3610/30QM could techincally be at 37W TDP, it's too high to be binned at 35W, so it's considered a 45W TDP chip. Just my two cents.

 

Intel decreased the clock speeds as well as the maximum turbo speeds compared to the 45W parts, that also helps to keep TDP in check. On 4 cores, at full turbo, the 3630QM will clock at 3.1GHz, the 3632QM will clock at 2.9GHz. You will get slightly lower temps with a lower TDP given the same cooling, but depending on the temp threshold for the fans, you might see a somewhat different thermal behavior.

 

Thanks, if it's just a matter of reduced tested frequencies over manufacturing changes, then it's not worth the $100 extra if buying 3632qm over 3610qm.

 

No it is not worth the extra $100.

 

Yes, it is pretty much a scam.

 

I would agree but I would point out it is the retailers/OEM's as Intel prices the same $378.

In a general sense I accept the TDP that Intel lists. I don't know how Intel does it, does Intel get a sheet of silicon with better thermal charachteristics. I don't know but I do accept it is what it is. I would think the benifit comes in the case it might be put in. Say a 17 with both the benifit of size for heat dissapation and greater space for a larger cooling system. So the lower TDP might be useful even critical as we keep trying to put greater in smaller spaces. I would like a quadcore in an 13" Ultrabook but I am not holding my breath.

 

They could do it, AMD offers 19W quad cores, just performance is diminished because of the lower TDP. The i7-3632QM are hand picked samples from the i7-3630QM batches is all and clocked down a couple hundred MHz.

Intel's dual core CPU's are more than sufficient for users in Ultrabooks though today. They are already quite powerful, and it doesn't seem to be Intel's goal for the time being to offer low power quad cores, although we might see something with Broadwell.

 

@ASUS-UX32VD look into what binning is. As for how Intel makes sure the TDP is respected, the CPU can estimate it's TDP (or there's a sensor, but I'm guessing it estimates it's TDP based on some equation) and it will downclock itself to the highest clock speed that falls within it's TDP if necessary.

 

The processor and cooling system aren't inseparable, so make sure you're using the same notebook for temperature comparisons.

Just to clarify, TDP is a number provided by Intel/chip fabricators so manufacturers using the chips know how much thermal energy the heatsinks they pair with the chips must be able to dissipate so the chip stays within its recommended operating temperature range (that was one long sentence, wasn't it).

 

If it helps, my 3610m draws 35-36 W according to Throttlestop when loaded with Prime95 boend and forced to 2.9ghz. My chip might've been a good sample but i reckon the main reason for the 3612 to be classed as 35w is simply due to a lower clockspeed and a remote possibility of being better binned (requires less voltage for a given frequency).
As for price, its a ripoff since you can get oem 3610m for about 230 bucks free shipping on ebay.
Also, all the non ulv chips idle within 0.1w of each other.