2 vs 4 ram sticks

[Sunny and 75]

I do have a X3D chip

What are we talking about in performance gain ? worth taking the chance of another kit running at 3733mhz cl 16 ?

The ripjaws v are single rank ram, so that's why they are cheaper than most ram,memorys ?


I never had a X3d so im a noob regarding ram,memory for that cpu

I only have the g skill ram, delivery doesn't take long when the ram is stock, order ram 10 o'clock or just a little earlier in the afternoon an you have it less than 12 hours later today it was before 10 o'clock in the morning

View attachment 354149

There is no discernible difference between 3600 and 3733 memory speed unless you're benchmarking.

 

[gasolin]

Would another kit be worth buying for the last performance gain (single vs dual rank ram)

Does single,dual rank matter as long as the ram runs in dual channel which my kit does

I might confuse dual rank ram with dual channel ram

Nice to know if i have to lower speed to 3600mhz if i buy another ram kit

 

[A Computer Guy]

I do have a X3D chip

What are we talking about in performance gain ? worth taking the chance of another kit running at 3733mhz cl 16 ?

The ripjaws v are single rank ram, so that's why they are cheaper than most ram,memorys ?


I never had a X3d so im a noob regarding ram,memory for that cpu

I only have the g skill ram, delivery doesn't take long when the ram is stock, order ram 10 o'clock or just a little earlier in the afternoon an you have it less than 12 hours later today it was before 10 o'clock in the morning

View attachment 354149

Something to keep in mind when talking about RAM performance gain. If your processing activity already fits in the bandwidth of your RAM then improving latency may increase performance. If your activity is being starved of bandwidth then increasing the bandwidth may increase performance.

Something that seems missing from this conversation is talk about other sub timings regardless of ram kit that contribute to better performance such as being able to reduce tRFC.

One thing I find difficult in following conversation like this is understanding the performance of a particular kits people are talking about. I put together this little cheat sheet that allows me to compare the timings in nanoseconds with the frequency. When people are discussing only the primary timing however that's not a complete picture of how a particular kit will perform but a general indicator of what kit may perform better out of the box when comparing kits. (not taking into consideration CPU or motherboard limitations)

One last note I would like to add. I have never been able to tell the difference while gaming between a CL16-18-18 DDR4-3600 kit from CL22-22-22 DDR4-3200 kit (13.75ns vs. 8.89ns) - but that's just me stuck at 60fps gaming target.

 

[Sunny and 75]

Would another kit be worth buying for the last performance gain (single vs dual rank ram)

Does single,dual rank matter as long as the ram runs in dual channel which my kit does

I might confuse dual rank ram with dual channel ram

Nice to know if i have to lower speed to 3600mhz if i buy another ram kit

Look, the most important thing to consider is the ram capacity, alright. Memory speed comes after that.

Unless you've got heavy productivity workloads, you do not need more than 32GB total ram capacity.

In terms of memory speed for ZEN3, 3600 is guaranteed for the optimal 1:1 ratio configuration. Anything above that is NOT and you won't notice the difference anyway.

 

[gasolin]

So regardless of amount of ram i don't need another kit of 4000 mhz cl 18 for any 4 ram (single rank) stick performance gain ?

 

[Sunny and 75]

So regardless of amount of ram i don't need another kit of 4000 mhz cl 18 for any 4 ram (single rank) stick performance gain ?

Correct.

 

[izy]

The idea behind x3D chips is to make the CPU access the RAM less (because you have the extra cache which is way faster than RAM) so i dont think you will see any gains in games having 4 sticks SR or 2 sticks DR vs 2 x 2 SR , maybe 2 3 fps in some cases (~5-10 FPS on a normal CPU from my experience depending on the game) .. you will probably see some benefit in some applications that are RAM sensitive, as for capacity i think 32GB is the sweet spot for gaming (even 16GB is ok but if you have stuff opened in background having 32G will help alot).

Edit: What i would go for is 2x2 32GB kit dual rank (most of 32Gb kits are DR) or if you already have 16 GB KIT 2x2 going for another one its a good but you have to be sure they are exactly the same , sometimes with 4x4 will you will have to run your RAM slower or 2T or even using worse timmings (but i think you will benefit more from extra capacity even with the other downsides in this situation) .

 

[gasolin]

I had some weird stutter,lag thing in dakar desert rally

I changed my ram to 3600mhz and now, knock on wood, it runs more smoothly

I will keep it at 3600mhz cl 16 say thanks for the help and thats it.

Cheap ram that performs good and if i happen to buy used 12-14 gen next time i upgrade, i have 4000mhz cl 18 memory i can use (or 3600mhz cl 16 if that runs best).

 

[freeagent]

Well.. GN, HUB, and others perhaps were using 4x8 3200C14 for their testing at one time, I think HUB switched to 3600C14. So 1R, 2R, its up to you.. Just thought I would point that out though.

 

[Sunny and 75]

I changed my ram to 3600mhz and now, knock on wood, it runs more smoothly

I say Amen to that

I will keep it at 3600mhz cl 16 say thanks for the help and thats it.

No problem, mate

Happy gaming!

next time i upgrade

Only Zen6 or the new Core Ultra lineup of processors are worth upgrading to, at this point.

With the 5700X3D that you've got, you won't feel the need to upgrade anytime soon, especially at QHD.

I have 64gb 3200mhz cl 16

About that 2x32GB 3200 kit, have you still got it?

 

[gasolin]

I only have g skill ram

Only Zen6 or the new Core Ultra lineup of processors are worth upgrading to, at this point.

12-14 gen used wouldn't be to expensive, i do have games i don't know why cpu usage isn't 96-100% or close to 100% but think a ryzen 5700x3d + rtx 4070 is a pretty good match for 2560x1400 180hz without going crazy with a 7800x3d + a rx 7900 xt or 4070ti super

 

[Sunny and 75]

12-14 gen used wouldn't be to expensive

Do not waste precious money on that, considering you now own a 5700X3D and you play at QHD.

Only another GPU upgrade will provide you with more performance that can be deemed substantial.

You've just bought a 4070 so the best path to take RN is waiting for something like a 5070.

Keep in mind, when it comes to upgrades, do NOT settle for anything less than 30% uplift in horsepower.

Happy gaming!

 

[gasolin]

xmp profile to 1, vcore to 1.450 volt, fclk (infiniti fabric) to 2000 (i don't think the available ram makes a difference in the final score)

 

[freeagent]

She isn't happy.

Edit:

You are using a way outdated version of Aida.

 

[gasolin]

Don't know why my L3 chache is mediocre

#89

Passmark memory score ?

3800mhz cl 14 is super expensive unless i know what ram can do that speed and they actually are cheap (32gb)

Chipset difference shouldn't matter to much (B550)

 

[freeagent]

Don't know why my L3 chache is mediocre

Outdated version of Aida.

Passmark memory score ?

Not sure haven't run it before.. I will try it out.

3800mhz cl 14 is super expensive unless i kn ow what ram can do that speed and they actually are cheap (32gb)

This stuff is 3200C14, cant find it on Newegg anymore though, at least not a sane price.. crazy.

 

[gasolin]

Latest aida64 Had to buy a licens first. Much Better

 

[freeagent]

Passmark memory score ?

 

[gasolin]

Second run of ibt (intel burn test) at high it passed https://www.techpowerup.com/download/intelburntest/

 

[Sunny and 75]

xmp profile to 1, vcore to 1.450 volt, fclk (infiniti fabric) to 2000 (i don't think the available ram makes a difference in the final score)

Is there a significant difference in your framerates by doing so?

 

[gasolin]

why should there be a significant difference in my framerates by doing so?

It just what i need to do for the ram to run stable with my cpu,mb

Higher min fps and 0.1% lows

 

[freeagent]

Kinda funny..

The last time I ran this I was on my 58X3D, and now I am on my 5900X.

All of my scores are higher, except my memory

But in other benches my memory scores higher with my 5900X

 

[gasolin]

High base and boost clock on a ryzen 5900x ?

 

[freeagent]

High base and boost clock on a ryzen 5900x ?

Could be..

 

[Sunny and 75]

why should there be a significant difference in my framerates by doing so?

It just what i need to do for the ram to run stable with my cpu,mb

Higher min fps and 0.1% lows
View attachment 355524View attachment 355525

The results are roughly the same.