Intel Core i5-4200U vs. i5-4210U, throttling?

Recently I've read the Haswell ULV processors are throttling, while I've experienced the older Ivy Bridge ULV series are stable at turbo frequency. This question concerns me more and more, since I'm planning to buy a new laptop. If I'm not mistaken the i5-4210U is a Haswell "Refresh" and I wonder if it is mean improved performance without throttling issue?

Ohh, and please keep for yourself just buy a full 47W CPU. Thanks, not again in this planet...

 

The reason for this is that Haswell has brought back the TDC limits for the turboboost algorithm.
Basically with Ivy Bridge, the TDC was set to around 95A for almost all chips/laptop models. This means that Ivy bridge Turboboost is pretty much completely determined by the amount of heat being produced (i.e. the TDP limit).
However for Haswell, the TDC limit is back to being 25-37A (basically back to the old Arrandale/Clarksfield/Sandy Bridge limits) for reasons unknown, I cannot fathom why this is neccessary for 22nm, because the older 40nm/32nm generation chips could draw a massive amount of current which might damage the power delivery system. If Intel was worried about heat output due to the IVR, they could've still limited Turboboost via the TDP limit instead of touching the TDC.
It is pure speculation but I suspect Intel has heard of our dear Throttlestop program.

I think you can use XTU or Throttlestop to try and raise the TDC limit in order to circumvent the throttling, otherwise, you're stuck without BIOS mods.

 

Thanks for the plentiful information, it looks like I need to learn new things, because "TDC" is uncertain for me.. The TDP, SDP limits are ok, could you please to explain what are these 95A vs. 27-37A related to TDC?

 

TDC is the thermal design current or something. Basically, it's one of the two parameters which determine Turboboost, it essentially limits the amount of power the CPU can draw. So for example, if a TDC of 37A is set, for a Turboboost load Vcore of 1.15V, the chip is allowed to draw a maximum of 42.55W before enforced downclocking. Basically, it limits the scenarios where the cooling system is extremely powerful (or the TDP limit is totally circumvented via Throttlestop/BIOS) where the CPU can theoretically Turboboost indefinitely otherwise (and potentially damage the battery/power deliver components etc).

This is the reason why Clarksfield chips didn't Turboboost very much because the baseclock current draw of the 40nm Quadcore is so high to begin with. Now, since the Sandy Bridge era, it was possible to exceed the TDP limits for a brief time unless the TDC was violated (there were some cases where the Laptop would allow an unlocked TDP but a locked TDC). For some reason, (I suspect because Ivy didn't draw that much power to begin with) Ivy had an essentially unlimited TDC cap of 95A (which potentially means the chip can draw up to 120ish watts) with only the TDP acting as the limiter.

25-37A is quite restrictive for someone used to unlimited Turboboost since it's a fairly hard limit unlike TDP. However, to complicate things further, as I understand, most ULV laptops have a Short term TDC built in so it can technically operate at a higher TDC for brief 2-3 second spans. All this conspires to create a very bursty but efficient CPU, quite unlike Ivy which could operate at full boost for as long as the cooling can last.

Unclewebb knows more about this than I do actually

 

The high yet short turbo is part of the "race to idle" design philosophy. Get the job done ASAP, and then idle.

 

Well, thanks for the info +Rep. I'll probably go ahead to get one and hope TDC is set to 37A, because 42.55W sounds plenty enough for CPU part. The IGP won't be used in same time, I buying one with discrete graphics.