DDR3 Memory: What Makes Performance Better?

Testing Methodology

I ran each test five times, removing the high and low scores and averaging the remaining three scores.  

MORE: Haswell And Richland Memory Scaling: Picking A 16 GB DDR3 Kit

Intel System

I enabled XMP (eXtreme Memory Profile), which set all the DRAM correctly to their specification (2400 and manufacturer's specified timings) with the exception of the ADATA and AMD sets, where two sets (2x8GB) of DRAM were provided to fulfill the 32GB. This often happens when people mix sets of DRAM, and as here, even with identical sets of DRAM. Often, one can make minor voltage and/or timing adjustments to get the sets or sticks to play, but not always. This is why DRAM manufacturers offer such a wide variety of sets of DRAM in 1, 2, 3, 4, 6 and 8 stick sets; the DRAM in a given package is all tested to work together. DRAM today, especially in the higher frequencies and tighter timings, is made to such high tolerances that many factors can affect the ability of sticks to play nicely together.

AMD System

The Asus Crosshair V Formula Z uses DOCP (DRAM Over Clock Profile) — which is basically a derivative or version of XMP tailored somewhat to the frequency of the DRAM, as opposed to XMP, which pulls and directly uses the actual timings found in the SPD as programmed in by the DRAM manufacturers. AMD also utilizes other overclock profiling found on other AMD motherboards, like AMP (AMD Memory Profiles) and EOCP (Easy Over Clocking Profiles). The one exception was to use AMP for the AMD Radeon Gaming DRAM which set up both sets of DRAM to spec settings of 11-12-12-31, 1.65. Under DOCP the DRAM was set to the same as the rest of the sets: 11-13-13-35, 1.65, indicating that with these sets, DOCP simply sets the DRAM to what it is programmed to believe the best timings are for 2400 DRAM.  This occurred with both the CL10 sets and the CL11 sets, and all of the sets ran fine at these settings. However, as the specification timings are tighter on all sets, I set the timings manually to the DRAM manufacturers' specifications. 

Before getting into the testing results, I had mentioned looking at DRAM set up problems on both platforms and there were indeed a few.

  • blackmagnum
    Since the dawn of computer memory: Faster data rates=More access latency.
  • Damn_Rookie
    It's great to see you getting the chance to do some articles on here, Tradesman1 Jim :)
  • PaulBags
    Interesting. I'll be coming back here several times over the next several months, as I save up for & consider options for a new computer. Good to know 2x 2 stick kits arn't nessicarily equal to a 4 stick kit, and annoying since I don't see many 4 stick kits in local online stores.

    Also good to know frequency has such a measurable impact on compression, I probably don't need fancy high freq ram but now I'm going to be more annoyed my budget won't allow it.

    Thanks for the ponderous read! :)
  • Cryio
    This was such an eye opening read. Until now I though different ram frequencies really didn't amount to much. By those times in multitasking ? Wow.
  • damric
    Thanks Tman. Looks like you really worked your butt off in this testing.
  • Crashman
    15517997 said:
    Since the dawn of computer memory: Faster data rates=More access latency.
    Nope. Latency stays roughly the same while clock cycles speed up. DDR3-1333 CAS 6 and DDR3-2666 CAS 12 have the same latency in nanoseconds, the number of cycles doubles while each cycle occurs in half the time.

  • synphul
    Pretty good review, made for a fun read. I'm still amazed at the ram incompatibility when mixing different sets. I know it exists, it's not the first time people have experienced this. I've mixed various ram just on a roll of the dice and I guess I've been lucky. The brand mixing was random. When upgrading one system from ddr2 to ddr3 I threw 2x2gb of patriot in with 2x2gb of geil evo one without issue. When upgrading one of these systems and adding ram, I added 2x4gb of team zeus at 1600mhz with 2x4gb of geil evo corsa at 1333mhz and just let the motherboard configure it and match to the lower speeds/settings. No issues with either.
  • xFolterknechtx
    1.65V sticks on Z87/97 boards? I think the upper limit Intel mentioned in their white paper for Ivy and Haswell was something around 1.55 - 1.575V for DRAM voltage.

    Do what you want with your hardware guys, but I ll stay clear of these voltages.
  • Pierre Olivier Vidal
    I have 2 questions for future articles.
    1 - Why do memory control sugest no higher voltage then 1.65V yet my DRAM says it can handle up to 2.4V and why would they make a DRAM capable of handling 2.4V if no memory control can push safely over 1.65V? ( I'm runing on Kingston Beast 32GB 2400MHz )
    2 - Why is there some DRAMs that are not compatible with some motherboards, what makes it compatible or not?
  • Onus
    I would hate to risk shortening the life of an expensive Intel CPU by running more than 1.6V through its IMC. You raised it to 1.7??