3000mhz vs 3200mhz ram, is there any point?

[Craigathorn]

I'm upgrading my ram to a 32gb kit and was wondering if there's any point in going for 3200mhz over 3000mhz, or whether I should just save myself the £20?

 

plus what would happen if I matched the 3200mhz with 3000mhz that I already have in my system? Both sets running in dual channel, not quad  

[ObliviationGNx]

best to have it all together at the same speed altho running at dual channel. I doubt it'll affect much if you use a higher clock speed on the other channel. But for me i'll keep it all the same speed. 

[Imakuni]

3 minutes ago, Craigathorn said:

I'm upgrading my ram to a 32gb kit and was wondering if there's any point in going for 3200mhz over 3000mhz, or whether I should just save myself the £20?

 

plus what would happen if I matched the 3200mhz with 3000mhz that I already have in my system? Both sets running in dual channel, not quad  

Depends on what you do. Assuming you are only doing some gaming, you'll be more than fine with 3000mhz.

 

As for mismatches, it would likely render XMP useless and you'd be forced to manually OC your RAM.

[SkilledRebuilds]

Higher frequency ram only really makes a difference when your cpu is fully tapped out.

Eg: Rendering/Encoding and heavier workloads can see benefits.

 

Having both in the system can be done.

You either run the 3000mhz at 3200mhz or run the 3200mhz at 3000mhz if the first option isnt stable.

[Craigathorn]

1 minute ago, Imakuni said:

Depends on what you do. Assuming you are only doing some gaming, you'll be more than fine with 3000mhz.

 

As for mismatches, it would likely render XMP useless and you'd be forced to manually OC your RAM.

Cheers dude, gaming and video editing is what I'll be using it for, doubt it would help with video editing either though right?

 

never through about the xmp profiles, overclocking ram (from what I hear) is meant to be a nightmare if you do it manually, so 3000mhz looks like it wins haha

[Nena Trinity]

Anything over 2800MHz wont give any noticeable gains...

[Imakuni]

1 minute ago, Craigathorn said:

Cheers dude, gaming and video editing is what I'll be using it for, doubt it would help with video editing either though right?

It helps a bit. Though between 3200 and 3000, it's nothing to flip about. You'll be good anyways.

1 minute ago, Craigathorn said:

never through about the xmp profiles, overclocking ram (from what I hear) is meant to be a nightmare if you do it manually, so 3000mhz looks like it wins haha

Uh... manual RAM OC isn't THAT bad.... once you know what you are doing....... and have a lot of time on your hands.........

 

But seriously, it can be as difficult as you want it to be. It can be very complex, but it can be a bit simpler tooif you don't want to push it too hard.

[typographie]

2 hours ago, Nena360 said:

Anything over 2800MHz wont give any noticeable gains...

Why would memory performance just arbitrarily stop improving at 2800 MHz? That seems random.

 

It may be more accurate to say a difference of only 200 MHz won't be a noticeable improvement, but a larger jump definitely can.

 

2 hours ago, Imakuni said:

Uh... manual RAM OC isn't THAT bad.... once you know what you are doing....... and have a lot of time on your hands.........

It didn't even take much time, in my case. I bought Corsair's DDR4-2666, and once I manually set to voltage to 1.35 V it seems to like anything between 2133 MHz and at least 3200 MHz just fine. This is just a guess, and obviously I know everyone's mileage varies with this stuff, but it may be really simple to overclock DDR4-3000 to 3200 MHz in many cases.

[Imakuni]

1 minute ago, typographie said:

It didn't even take much time, in my case. I bought Corsair's DDR4-2666, and once I manually set to voltage to 1.35 V it seems to like anything between 2133 MHz and at least 3200 MHz just fine.

V

2 hours ago, Imakuni said:

But seriously, it can be as difficult as you want it to be. It can be very complex, but it can be a bit simpler tooif you don't want to push it too hard.

[smokefest]

No point for 3200 mhz over 3000, or 2400 or 1600..

All the same performance in games..

It will be better for encoding, rendering, thats all.

 

For games, as long as you have 1333 mhz ram, you have reached max performance.

[Imakuni]

7 minutes ago, smokefest said:

For games, as long as you have 1333 mhz ram, you have reached max performance.

Faster RAM totally doesn't help, right?

 

I can give you many more examples, if you want.

[smokefest]

Where did u take thart chart ? is so wrong. Ive done it myself. 1333 VS 1600 VS 2400 , i had 0 difference fps in all games tested.

Also go check youtube, il llink some..

 

 

Here again, 1333 VS 2400, 0 fps changes

 

And by the way, you're chat dosnt even use the same CPU's that chart is false.. go check some benchmark on youtubes, from 1333 to 3200 , at max you will have 1 fps gain.

 

So no difference.

[Imakuni]

28 minutes ago, smokefest said:

Where did u take thart chart ? is so wrong. 

EuroGamer. I can pick from other sources if you'd like, though.

28 minutes ago, smokefest said:

Ive done it myself. 1333 VS 1600 VS 2400 , i had 0 difference fps in all games tested.

I've done it myself as well and there can be anywhere from margin of error difference to 10%.

28 minutes ago, smokefest said:

Also go check youtube, il llink some..

Right... so you pick a potato GPU such as the 660ti, put it on Ultra 4X MSAA to create the biggest bottleneck you possibly can, test on 2 games and then come the ballz to say it doesn't matter? C'mon... you can't be serious.

 

Also, I suggest you ignore all videos from that dude, for he clearly doesn't know what he's doing.

27 minutes ago, smokefest said:

CPU? GPU? MSAA levels? Only 2 games? Same story as the other video.

26 minutes ago, smokefest said:

And by the way, you're chat dosnt even use the same CPU's that chart is false.. go check some benchmark on youtubes

Er... excuse me? Do you even understand what's going on?

 

You can see an i3 6100 moving from the standard 2133 RAM to 2666mhz; you can see the boost there. They also throw in an i3 4130 moving from standard 1600 to 2133 RAM; you can't compare it to the 6100, that much is true, but you are comparing teh 2 4130 tests to see the difference in RAM speed.

 

Just because you don't understand a chart doesn't mean that it is false. But okay, you want youtube benchmarks? Fine, have a video from them if you don't believe the chart.

 

26 minutes ago, smokefest said:

at max you will have 1 fps gain.

 

So no difference.

Imprecise. If you have an unbalanced setup with settings cranked WAY above those that your GPU can handle, better RAM won't help. And in some games, even a balanced setup with good settings is not going to benefit from RAM. In those instances, you'd be right to say that RAM doesn't help.

 

But generically saying that at max you get 1fps in every case is just straight up wrong. RAM does matter, and it's time for people to start realizing that.

[smokefest]

Show me one video that prooves that higher mhz ram gives better performance. Your chart is wrong.

 

Just find 1 . 1 single video on youtube when i can see  a difference in games fps

 

I have 4690K @ 4.5 GHZ with gtx 970 tryd with 1333 1600 2400

Rise of tomb raider, crysis, BF1 and many more games tested, not a single fps boost.

Also all and ALLLL the vids on youtube shows that ram dosnt give performance boost at all once youve reached 1333.

Also if you go check linus tech tips youtube video about ram speed in games, it confirms everything i say

 

 

 

For the Linus honor, Thank you

 

From 32 to 32.7 from 800 mhz to 2400 mhz...

Thats what I call no difference 

[Imakuni]

1 minute ago, smokefest said:

Show me one video that prooves that higher mhz ram gives better performance. Your chart is wrong.

1 minute ago, smokefest said:

Just find 1 . 1 single video on youtube when i can see  a difference in games fps

I just did?

8 minutes ago, Imakuni said:

Just because you don't understand a chart doesn't mean that it is false. But okay, you want youtube benchmarks? Fine, have a video from them if you don't believe the chart.

Are you even listening to me?

 

Also, no need to make separate posts. Group them up into a single one, makes the thread a lot more organized.

1 minute ago, smokefest said:

Also all and ALLLL the vids on youtube shows that ram dosnt give performance boost at all once youve reached 1333.

I just showed one proving it wrong, but okay...

1 minute ago, smokefest said:

Also if you go check linus tech tips youtube video about ram speed in games, it confirms everything i say

Sigh... if I was to ever become an LTT CEO, my first measure would be to delete that RAM video and handwrite a formal appology to each and every person that has ever viewed it. The video is just THAT trash:

1. Lack of depth (remember that "only 2 games" thing?)
2. Flawed methodology (remember "potato GPU + uber settings" thing?)
3. And it even goes as far as explaining wrong stuff (latency part is wrong).

That video should not be taken as a source. Don't quote it just because he is Linus, the dude doesn't really know as much as people might be led to believe.

[maybethisnamewillwork]

10 minutes ago, smokefest said:

Just find 1 . 1 single video on youtube when i can see  a difference in games fps

 

[smokefest]

Why posting a lame i3 6100 ????

 

Keep thinkin what you want, I can assure you and othert people here that your chart is ... actually retarded,,, they dont evne cmopare the same cpus, everything is wrong in it,

Oh i wonder why 99% of the vids on youtube shows and 99% of techs says ram dosnt change performance in games.

But your 0.00001% must be right. sure.

 

 

 

Ive seen dozen of i5 i7 with good gpus benchmarking their ram speed gaming performance and never there was more than 1 fps difference. Didnt watch about i3 as if u use i3 for gaming then ur not even worth commenting here

[Imakuni]

3 minutes ago, smokefest said:

Why posting a lame i3 6100 ????

You want the i5 test? Fine, suit yourself

 

[smokefest]

The fps are al lthe same ?

 

You even have less fps with 3200 then 3000.. zzz..

 

 

 

 

 

There are thousounds vids, they all show the same ... almost , close to almost no gain in fps with ram speed.

 

 

 

See ur i5 skylake... many games... almost ALL them have 0 gains.

 

Of course there might be a situation where it helps a bit...

This is 0.01% of reality. 

99.99% of reality , ram speed dosnt change gaming experience.

 

enjoy

[redclaw72]

On 10/6/2016 at 8:24 AM, smokefest said:

The fps are al lthe same ?

 

You even have less fps with 3200 then 3000.. zzz..

 

 

 

 

 

There are thousounds vids, they all show the same ... almost , close to almost no gain in fps with ram speed.

 

 

 

See ur i5 skylake... many games... almost ALL them have 0 gains.

 

Of course there might be a situation where it helps a bit...

This is 0.01% of reality. 

99.99% of reality , ram speed dosnt change gaming experience.

 

enjoy

 

 

 

Your a idiot lmao

[DarkSmith2]

To end this discussion. Every game that benefits from having a faster CPU will depending on how well the game can use cache benefit from faster RAM.

 

If a CPU spends 50% of it's time waiting on ram, and you make the ram 50% faster then the CPU will become 1/(.5+.5*.5)= 1.333 times faster.

If it spends 10% of it's time waiting, then the speed up is 1/(.9+.1*.5)= 1.05 times faster.

The benefit is always positive, but depending how well the application can use cache, it may be large or small.

 

This is scientifically proven since 1988 and is called Gustafson's law. https://en.wikipedia.org/wiki/Gustafson's_law

If anyone wants to bring his own scientifical Theory to supersede Gustafson's law, go ahead try to change facts in Computer science.

 

Have a nice day.