It's been a long time since i tested system memory. A few years ago, high performance memory and overclocking appeared to make less difference to gaming performance than it does now. However, i am only nearly done benching GTX 690 and i7-920 at 4.2GHz and there are in more games than not, a few fps gain from going from 1200MHz to 2000MHz.

Next, i will bench the GTX 690 and the i7-3770K at 4.8GHz with the variable being memory speed, from 1300MHz to 200)MHz and then again from 4 to 8GB of system memory. That will give us a good idea about memory upgrades in Bloomfield and Ivy platforms (and no doubt Sandy's also).

How long ago exactly? Back when the memory controller was part of a northbridge that sat separate from the CPU on a motherboard I'm guessing.

P4 days


As to "exactly", unsure. i also did a platform upgrade to X58 for ABT about 3 years ago. i think i just touched on memory.

I don't know that I'd draw too many conclusions about anything (except maybe heat generating potential and highest clockspeed while disregarding IPC) from a P4 system....

Imagine how fast PC's would be if only CPU's could use DDR5 memory clocked at 6 gbps like Kepler does.

These are the kind of "fun" evaluations that i like to do.

its proving to be a very informative article all ready!

Thanx for the hard work

Thanks for your interest and help in giving this evaluation direction and meaning!

i am taking shortcuts, however. When we chart the performance improvements of faster RAM in gaming, i am not going to do both GTX 680 and GTX 690. Just one of each seems to show the barebones of what we are looking for.

Right now i'm testing GTX 690 and 3770K at 4.8GHz and the differences between 1333MHz and 1866MHz on the memory (the default and default high overclock). Then the test is between 4 and 8GB of the same RAM at 1866.
--then the "detour" is finished and i bench other than gaming using i7-3770K with 8GB RAM (1333MHz) at 4.2GHz versus i7-920 with 6GB RAM (2000MHz) also at 4.2GHz for the rest of the IPC comparison (and also the ECS MB evaluation)

the data you are collecting is amassing. There is so much that could be talked about. There is so much to cover. a masterpiece.

A lot of data and you have to present it in the best way as possible. You got a lot of topics but they are all related. You got to thread it properly so it makes a terrific read. The evolution of intel from '08 to 2012 from the consumer stand point. You are covering a whole lot, cant wait to see how you present this. Its got huge potential and it should be interesting to so so many people. Even the ones who wouldnt think so at first would have something to learn or gain. Anyone who loves the PC would be very very interested in this. It makes me excited.

thanks for the work

i think you are getting a little carried away with expectations. This is just a platform upgrade evaluation. i have to test everything thoroughly as i will be basing all of my future evaluations on until the next one.

You are correct in that the presentation is important. Thank-you for your suggestions.

i think there is a lot that can be concluded from the data. It tells the whole story