Yes its about $30 cheaper to do that. But I wanted to avoid having to tweak my bios and settings to use quad channel and also avoind any issues with the ram not being identical even though they are the same product number settings and timings etc etc.
Well, on that tweaking angle, I got two stop-code 124 screens in the initial testing with OCCT/CPU. I took a walk on the wild side and did not delete the caching regime for my two main storage devices -- SSD and HDD. I ran CHKDSK against both after each Event ID 41, and despite the deferred writes setting, not a single bit raised any corrections.
I twisted up the Extra-Voltage-Turbo by 12mV, and ran affinitized LinX "Align 4" and maximum problem size, maximum memory, 25 iterations with the caching consuming about 5GB of RAM. The AVG GFLOPS were about 130, and the range from lowest to highest was 1.4 GFLOPS. That's 0.7 GFLOPS to either side of the mean.
As for the registered voltages under that particular stress test and setting, it looked to be about the same as it had been at 4.7Ghz before I added the RAM. The droop took it to about 1.368V, and the "turbo-idle" was about 1.384V.
I'm not even sure if I want to twist it up another 4 mV. The temperatures were the same, or an average maximum of 76C, with variation peaking absolutely at 78C.
It really wasn't a big problem, SelenaGomez. You never told us what make, model and spec 2x4GB RAM kit you were using. I already said that you won't be as likely to run CR=1, but it will not make much of a difference, or maybe 5 ns in a benchmark. But you can pair your kit with either the same or similar kit from the same manufacturer -- whatever size you choose -- if the SPD data of both overlap at certain speeds, or seem to extrapolate to an overlap.
If you find a kit similar in product code and other respects which has tighter timings at the same spec voltage (maybe . . CAS 8 instead of 9), set the timings in BIOS to the slower memory's spec. They should run at spec voltage, and you might want to increase VCCIO/VTT for the IMC by about 3 notches, because it doesn't make a difference to the IMC's health and might likely still be a setting below 1.100V. I think my own rides above reported at "auto": by 60 mV.
That should alleviate any possible shortcomings, and circumvent some testing.
Right now, on my 2700K, I would probably need to increase the turbo-voltage to just over 1.40V to get 4.8Ghz. But running this at 4.7 -- with the 4GB extra RAM, the two-tiered caching for the HDD and RAM-caching for the SSD boot-drive, it's still better than it ever was, and it was always great. Of course, different strokes, different folks and so on. . . .