Pairing RAM and a 6700K

[Michael]

I am probably not going to push the 6700K too much with overclocking. I bought some highly rated DRAM that is on my MB's recommended list, but I am wondering if the high price for 16GB is worth it considering that I can get twice the DRAM for the same price if I use standard spec DRAM (DDR4).

Any opinions?

One thing I do know that you always get a small drop in reliability when you populate 4 memory slots as compared to 2 and 16GB is plenty of RAM for all but video intensive applications.

Michael

 

[nerp]

I am in the camp that says a little bit faster memory is not worth it. It's very hard to see a difference and the advantage in gaming is usually only a few FPS. You're better off rolling that cash into better GPU over time.

 

[Majcric]

IMO, 3000 is the sweet spot. A few games benefit now and should even more in the future you'll be covered. Personally I would want my i7 6700k setup as a one and done deal that would hold be over until next platform upgrade. (no worries about faster or more ram later)

 

[Carfax83]

I am probably not going to push the 6700K too much with overclocking. I bought some highly rated DRAM that is on my MB's recommended list, but I am wondering if the high price for 16GB is worth it considering that I can get twice the DRAM for the same price if I use standard spec DRAM (DDR4).

Any opinions?

One thing I do know that you always get a small drop in reliability when you populate 4 memory slots as compared to 2 and 16GB is plenty of RAM for all but video intensive applications.

Michael

If you care about performance, you'll want to use high frequency DDR4 with your CPU. Skylake can take advantage of fast DDR4 due to it's aggressive hardware prefetchers, and its high performing cores will definitely enjoy the bandwidth.

I agree with Majcric that DDR4 3000 is likely the sweet spot in terms of price and performance. However, I'm not sure but you might have to use 125 strap with DDR4 3000, as 100 strap doesn't work well with that frequency; at least on Haswell-E.

Skylake might have changed it though, but DDR4 3200 works very well with 100 strap, as that's what I have my RAM running at..

A lot of people, including myself don't like running higher straps due to the fact that it messes with your idle clock speeds and voltage, uses more power and prevents the use of adaptive voltage overclocking.

 

[Dasa2]

faster ram will improve the efficiency of the cpu about 10% in games which is roughly the equivalent to overclocking from 4.5ghz to 5ghz
or a previous gen 4790k with fast ram is quicker than the 6700k with slow ram

these are some of the best value kits currently available at newegg
Team Vulcan 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 3000 (PC4 24000) Desktop Memory Model TLRED416G3000HC15ADC01 $66.99
G.SKILL Ripjaws V Series 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 3000 (PC4 24000) Intel Z170 Desktop Memory Model F4-3000C15D-16GVR $69.99

this would be one of the fastest kits available but the value is dropping
G.SKILL Ripjaws V Series 16GB (2 x 8GB) 288-Pin DDR4 SDRAM DDR4 3200 (PC4 25600) Desktop Memory Model F4-3200C14D-16GVK $99.99

the strap problems mentioned above are limited to haswell-e so you dont need to worry about that with z170\skylake

what ram do you have now exactly?

 

[UsandThem]

I am in the camp that says a little bit faster memory is not worth it. It's very hard to see a difference and the advantage in gaming is usually only a few FPS. You're better off rolling that cash into better GPU over time.

I agree. Faster ram helps in synthetic benchmarks, integrated GPU gaming, and a few other tasks. However, if you are gaming using a discrete card, there is next to no extra performance.

Plus, depending on what you use your PC for, you might not need more than 16 GB anyways.

This is a good article showing how the different speeds perform, which might help you:

http://www.legitreviews.com/ddr4-me...finding-the-best-ddr4-memory-kit-speed_170340

 

[Carfax83]

I agree. Faster ram helps in synthetic benchmarks, integrated GPU gaming, and a few other tasks. However, if you are gaming using a discrete card, there is next to no extra performance.

A common misconception unfortunately, which continues to proliferate due to faulty testing methods used by most hardware websites.

To measure the real impact of faster RAM, or even more RAM, you need to measure frametime variance. Using average framerate only, especially with canned benchmarks like what legitreviews used which are designed to stress the GPU, will likely not show any differences; but that doesn't mean there are no differences.

Infact, framerate is not really even a good method of judging performance in a game because it doesn't tell you what's really going on. Only frametime testing can do that.

The best way to test RAM performance is to find a CPU intensive circumstance, perform a consistent run through lasting at least one minute, and measure via frametime analysis.

 

[UsandThem]

A common misconception unfortunately, which continues to proliferate due to faulty testing methods used by most hardware websites.

To measure the real impact of faster RAM, or even more RAM, you need to measure frametime variance. Using average framerate only, especially with canned benchmarks like what legitreviews used which are designed to stress the GPU, will likely not show any differences; but that doesn't mean there are no differences.

Infact, framerate is not really even a good method of judging performance in a game because it doesn't tell you what's really going on. Only frametime testing can do that.

The best way to test RAM performance is to find a CPU intensive circumstance, perform a consistent run through lasting at least one minute, and measure via frametime analysis.

One thing with computers, there is always something you come across that you had no idea about.

I guess if you are an extreme enthusiast, you can use that test to tell you right down to the minuscule detail of the different performances of ram at different speeds. It is sorta like when lower CAS performance ram became the "it" thing back in the late 90's / early 00's. It provided a very small performance boost compared to the cheaper stuff, but you paid dearly for it. (I bought a Mushkin 4 Gb stick from the (now) admin John on here in 2000 for $100 or so).

However, for the casual to moderate user, the FPS charts and program time charts they publish on hardware sites is good enough and gives a pretty good estimate of the different speeds. If a person is a power user and has to get every single ounce of performance out of their system, I agree that most of them will buy the fastest ram on the market without thinking twice about it.

But for someone like myself who is a big proponent of the 'best bang for your buck' items, there is just not enough performance difference for me to pay extra for fastest ram. For me, there is a clear point of diminished returns on any product.

When I went to build or upgrade my computer over the years, I would look at a review like this one:

http://techbuyersguru.com/gaming-ddr4-memory-2133-vs-26663200mhz-8gb-vs-16gb?page=0

I find the sweet spot and go with it. If the fast stuff is one sale and I can get it for the same price, great. However, in all my years of using the various computers I have built, I have never had an issue of my low-to-mid performance ram holding me back from what I do. Now if a person has a program or task on their computer that in someway benefits from the faster ram, then they should go for it.

 

[bononos]

A common misconception unfortunately, which continues to proliferate due to faulty testing methods used by most hardware websites.

To measure the real impact of faster RAM, or even more RAM, you need to measure frametime variance. Using average framerate only, especially with canned benchmarks like what legitreviews used which are designed to stress the GPU, will likely not show any differences; but that doesn't mean there are no differences.

Infact, framerate is not really even a good method of judging performance in a game because it doesn't tell you what's really going on. Only frametime testing can do that.

The best way to test RAM performance is to find a CPU intensive circumstance, perform a consistent run through lasting at least one minute, and measure via frametime analysis.

Has any memory manufacturer or reviewer put out benchmarks on frametime variances for fast ddr4 ram?

 

[Michael]

Well, I can see I over paid for too fast RAM that will not make too much of a difference. Thanks for the links to the two articles, each were informative.

Michael

 

[UsandThem]

Well, I can see I over paid for too fast RAM that will not make too much of a difference. Thanks for the links to the two articles, each were informative.

Michael

At least RAM prices are pretty low right now. It could have been much worse.

What RAM did you go with BTW?

 

[Dasa2]

However, for the casual to moderate user, the FPS charts and program time charts they publish on hardware sites is good enough and gives a pretty good estimate of the different speeds.

they really dont those reviews you linked are very misleading
there about as accurate as this test suggesting a pentium is faster than a 6700k

gtx980
Intel Pentium G3258 99.92FPS
Intel Core i7 6700K 99.6FPS
290x
Intel Pentium G3258 96.64FPS
Intel Core i7 6700K 91.36FPS
http://www.anandtech.com/bench/CPU/1321

when in reality the cpu is much faster as is the ram once you take them out of a gpu bottlenecked scenario
which is why all ram reviews should include different cpu speeds so that people can see the value in it

http://www.eurogamer.net/articles/d...it-finally-time-to-upgrade-your-core-i5-2500k

rather than make a real long post full of tests i will provide some links with just that

a collection of links to tests i put together a while back
http://forums.anandtech.com/showpost.php?p=37837493&postcount=14

An Independent Study: Does The Speed Of RAM Directly Affect FPS During High CPU Overhead Scenarios?
http://www.overclock.net/t/1487162/...affect-fps-during-high-cpu-overhead-scenarios

 

[UsandThem]

they really dont those reviews you linked are very misleading
there about as accurate as this test suggesting a pentium is faster than a 6700k

We are just going to have to agree to disagree on that. DDR4 2133 to DDR4 3200 is not going to give a person a big increase in performance. Most games 2-4 FPS more, with a couple a touch higher.

There are a lot of articles out there from reputable sites on the the memory speed issue. You are not going to convince me that all of them are misleading or inaccurate.

 

[Dasa2]

We are just going to have to agree to disagree on that. DDR4 2133 to DDR4 3200 is not going to give a person an increase in performance like going from a Pentium to a 6700k.

There are a lot of articles out there from reputable sites on the the memory speed issue. You are not going to convince me that all of them are misleading or inaccurate.

you seem to be missing the point that when gpu bottlnecked there is no difference between a pentium or a 6700k other than margin of error in the review above i linked done by anandtech you dont consider them reputable?

faster ram improves cpu performance so in any game where cpu performance makes no difference only gpu faster ram will make no difference at all

there is plenty of reviews i linked above that show the other side of the story where cpu limited sections of gameplay see large gains from ram
if you chose to ignore them thats your loss

 

[UsandThem]

you seem to be missing the point that when gpu bottlnecked there is no difference between a pentium or a 6700k other than margin of error in the review above i linked done by anandtech you dont consider them reputable?

faster ram improves cpu performance so in any game where cpu performance makes no difference only gpu faster ram will make no difference at all

If I just say o.k., can we move on? I come here to help others, not to argue.

I never said faster ram doesn't improve performance. I said for the average person it doesn't matter, and as with any product, there is a clear cut-off price point where diminished returns kicks in.

 

[Dasa2]

nope im afraid we cant because the links you provided suggested to the op that faster ram will provide next to no gain which is only half the story
there tests are just fluctuating with margin of error they could drop the cpu speed and also see no change in fps
maybe not all the way down to a pentium but you get the idea? the only thing that will increase fps in there tests is a faster gpu

in any part of a game that is cpu limted the performance can go from 50 to 60fps just from faster ram
so for anybody that is considering overclocing the cpu faster ram is a must as it significantly improves the value of overclocking

im just trying to help to and the best way to do that is to help those that provide help to others understand how they have been mislead by one sided reviews

 

[UsandThem]

nope im afraid we cant because the links you provided suggested to the op that faster ram will provide next to no gain which is only half the story
there tests are just fluctuating with margin of error they could drop the cpu speed and also see no change in fps
maybe not all the way down to a pentium but you get the idea? the only thing that will increase fps int here tests is a faster gpu


in any part of game that is cpu limted the performance can go from 50-60fps just from faster ram

im just trying to help to and the best way to do that is to help those that provide help to others understand how they have been mislead by one sided reviews

http://www.silentpcreview.com/article1478-page6.html

Though the testing for this investigation was brutally tedious, the results were satisfyingly conclusive. On Skylake systems, the type of memory used is only important if you lean heavily on the integrated graphics chip. If this is the case, opting for a single stick rather than a pair of DIMMs in dual channel is a critical mistake that can cost more than 30% performance. Frequency is relevant as well with 3000 MHz offering around 10% improvement over 2133 MHz. The difference between C12 and C15 is negligible and should be ignored unless the price differential is minimal.

On the otherhand, if you only use the integrated graphics for pedestrian uses like video playback or if a discrete video card is to be utilized, the speed and latency of the system memory is inconsequential. Our general performance tests results depict the highest performing settings as having a one third of one percent advantage overall. High speed kits are marketed towards enthusiasts but even the most hardcore overclockers can achieve their desired goal without high frequency RAM. Skylake "K" variant chips have multiplier based overclocking and standard Skylake processors can be clocked up by increasing the BCLK frequency and using low memory dividers.

 

[Dasa2]

another gpu bottlnecked test they could oc there cpu to 5ghz and still see no change in fps with a 6700k paired with a gt640 2g
yes the majority of ram reviews done show no difference so im sure you can continue to link all day to tests where ram makes no difference it doesnt change the fact that its only half the story as all they have really benchmarked is the gpu

your still ignoring the fact that faster ram cant help in game where overlcocking the cpu will make no difference

all your showing is that overclocking cpu\ram is pointless in many games where the gpu is all that maters

edit
as for other cpu tests a lot of programs that are cpu bound see no change from ram
i believe this is due to the way they utilize the cpu cache
games are large programs and the files they need sent to cpu is not nice and orderly like it is when rendering so whenever the game needs something thats not already in the cpu cache it has to wait on the ram for it this is why the intel 5775c with a large l4 cache does so well

 

[UsandThem]

all your showing is that overclocking cpu\ram is pointless in many games where the gpu is all that maters

And that is exactly my point.

Average user with one video card to play games, do office work, and other regular desktop tasks. This person does not care about uploading results or going on a forum to talk about about the extra 1% in performance they get from 'gaming' RAM. All they care about is their computer playing the games they buy.

That person will not get a good return on their investment by buying high-end and expensive RAM.

Article after article shows the sweet-spot for RAM performance is between DDR4 2666 to DDR4 3000. However, they could probably buy DDR4 2133 and it really wouldn't make a bit of difference for their use.

If the average user (no synthetic benchmarks, or Firestrike result uploads, etc) instead of buying DDR4 2666 16 GB kit for $63 they instead bought DDR4 3200 for $119, is it going to improve their performance so much to justify the almost double price they paid for it?

With that, I am done in this thread.

 

[Dasa2]

and my point is that they have already spent ~$200 extra on a overclockable cpu\mb\cooling which will provide no benefit in the game tests you showed

its only $10-50 extra for faster ram depending on how far they want to go which will provide desent gains in cpu limited games just like overclockign the cpu will

 

[Michael]

I way overbought speed and I am thinking of dropping down to 3000 or 3200 and get 32GB. I will be trying my hand at publishing and that means manipulating image files so more RAM (and faster but not crazy fast RAM) will be just as good.

I will overclock the CPU, I just will not be water cooling and pushing it to the bleeding edge of stability.

Michael

 

[Dasa2]

32g could be a big help if your needing over 16g for your work
if not the faster ram isnt without its perks
http://www.techspot.com/article/1171-ddr4-4000-mhz-performance/

SLI 980Ti

Single 980Ti

these tests seem higher than even i would expect so i wouldnt be surprised if there was something with there test setup that is exaggerating there results
very few tests show any real gain in handbreak so unless they are using a different benchmark method\file type it may be hardware related