[SOLVED] I5-10600K memory can run at 2933 ?

[JeanP.]

If I understand correctly...


The I5-10600K are based on the I7-10700K (but with 6 cores vs 8 cores).
The I7 memory "naturally" run at 2933, the I5 is 2666. (I understand Intel do this to make a difference between the 2 processors).


My question:

It is any "hardware" (or other) reason/setup the I5 memory must run at 2666 (or it can run at the same speed the I7 do).

 

[Eximo]

With an appropriate Z class motherboard, you can set the memory speed to whatever you want. Processor should easily handle speeds up to 3600Mhz with relative ease.

 

[JeanP.]

Thanks for your reply.

I am not a pure overclocker.

I just ask myself if the "natural" memory speed is not more 2933 (like is "mother" I7).

In my understanding the I5 is an I7 with 2 cores not "on" (can be by choice, or because one(2) of this 2 cores can also have a problems),
Instead to put in garbage this I7, Intel just setup them as I5 (and put a lower memory speed not to lost too many customer to the I5).

I am correct ?

 

[Eximo]

2933 is just the speed they officially released 10th gen with, 11th gen supports 3200Mhz. Skylake launched with 2133, but even then the CPUs were more than capable of running 3600/3800.

10th gen is a little weird in that they chose to restrict memory overclocking to only the Z class boards (at a time when AMD was not doing this, with better low end CPUs...) They reversed this decision, and B and H class boards have memory overclocking again. (Which is silly when you think about it, the memory controller is on the CPU)

No, the i9/i7/i5/i3 are distinct products. There is a 10-core package, an 8-core package, a 6-core package, and a 4-core package, and a lowly dual core package.

Typically chips that don't make the cut are tossed or come out as the F chips without graphics (a significant portion of the chip) Far as I can tell, they don't release silicon with fused off cores or anything like that.

Within your package you have the different grades. 10600 oddly represents the target chip.
K series chips have high leakage, which is suitable for overclocking and not very power efficient.
T series chips operate very efficiently at lower voltages, so become your low power variants.
And the worse they behave under applied voltage and power the lower down the stack they go. Your 10400 and the like.

The main reason Intel has so many models is this package structure. Whereas AMD can take their chiplets and make all kinds of configurations with a single design. Which Intel is moving towards copying in 2023, though they intend to stack the chips.