Is FASTER DDR5 RAM Worth It For Gaming? - Watch This BEFORE You Buy!

[@DannyzReviews]

MSI has done a good job with their bios updates. When I tackled this topic before and did my testing it became clear that the XMP was setting some obscenely loose timings degrading performance and hence we saw a large gap when I tuned the ram myself. This time the XMP profiles seem to be doing a good job keeping up with their tune counterparts.

[@alchemystn2o]

I don't know anything about tuning the DRAM timing. I just set to XMP and hope it works without crashing. Too complicated and intimidating for me, sadly.

[@Bang4BuckPCGamer]

Nice work mate.

[@brianrobinson3961]

I grabbed a 48 GB 6800 kit from GSkill and the latency is horrible. I don’t know why but it’s in the mid 70ns range.

[@IIHydraII]

XMP usually has trash tertiary timings. Try tRRD S/L 4 & 8, tFAW 16, tXP 4 or 6

[@rulik006]

MSi sucks at DDR5 oc on their 4-dimm boards much worse than Gigabyte and Asus

[@Helios.vfx.]

Thank you for your content. I'm doing some research over ram and last products to build next year a 9950x + 5090RTX build.

[@nathanwhite617]

Is there any reason to buy a more expensive z690 over a z790 that doesn’t have stuff like post codes? Z690 is so much cheaper.

[@impuls60]

Well done video! Some well known Pubg streamers run 7200Mhz Xmp with tons of crashes. I tried to tell them that its ram beeing not completely stable since Pubg demands a lot from your ram. Drr5 runs hotter than ddr4 so you need a fan on them and hot air from the graphic cards leads to tons of crashes. Running Karhus or similar and a 3d benchmark at the same time is a must to see if the ram is truly stable.

[@StuartBoyer]

On one of my rigs a Z790 Apex Encore ,14900k (my bin performs the same as a KS) and 4090 strix all custom liquid cooled and direct die, yes I even cooled the ram, had two 420 rads in push pull config and one external 420 MO-RA3 in push pull and I found a 1 fps difference at 4k which is how I play across several games between 64gb DDR5 at 6400mhz and 48GB at 8200mhz that was in average fps and 1% lows. I just run the 48 GB as its a good middle ground for gaming and other task and use the 64 GB when needed for DCS World.

[@Kacpelux]

a really well done test

[@jediknightdiscomike22]

I have three of the same motherboards running gskill 7200mts on three different cpus running xmp. No stability issues what-so-ever.

[@Yasaro007]

Motherboard Manufactors just milk the i9 14900k 8000mhz Cod Bros then they enable XMP and are wondering why there game is crashing

[@kevinmccrea1335]

What does everybody think DDR 5 speed will go up to? Where will it top out at before DDR 6 comes out.

[@Shadowangel09]

I think somewhere like 9600. Teamgroup already has a 8000 kit but i dont anything can handle it and be stable

[@shannonbaker2069]

The top guy from each AMS and Nvidia performance division have flat out stated cas latency is far more important than frequency speed for gaming,and the opposite for productivity software

[@G1NZOU]

Yeah I heard from a few sources that 6000MHz with CAS-30 is a nice sweet spot, any faster than that and the latency is usually lower. X3D processors also seem to perform well, with differences in RAM speed for gaming performance relatively negligible.

[@eleven99]

Considering that you can still use DDR3 perfectly fine despite its age, I would say theres no urgent need for upgrading to it just yet.

[@ctjmaughs]

Ddr3 usually has better latency than the top speeds with recent memory

[@advancedmomuser-kb2qn]

what shrooms are u sniffin dude?ddr3 is dead platform like for five years already..

[@youtubeaccount7544]

Ddr3 is trash

[@ItsBigEz]

The only scenario ddr3 wouldn't be that bad is a 4790k @4.5ish and ddr3 2400-2800cl11

[@eleven99]

@@advancedmomuser-kb2qn DDR3 is still perfectly usable today. Maybe you should try it out before coming to such a conclusion.

[@uberhalla]

How much difference will RAM tuning make for 4K resolution gaming? I'm running a G.Skill Trident Z Neo 2x16 32GB DDR4 3600MHz CL16-16-16-36 kit. I just use the XMP profile provided by my motherboard. Probably could tighten the timings, maybe they are a bit loose...

[@DannyzReviews]

Not a lot for 4k if at all, but 1% lows will probably benefit resulting smoother gameplay overall

[@Solrac-Siul]

i can tell you from experience that at 4k the difference between my ddr4 4266-15 and my wife's ddr5 6400 , with the same processor (12700k ) and gpu, is around 1 to 2 fps . once you go above 1440 as long as your cpu is not the bottleneck, fps will be dictated by the gpu. So if you are playing a more than 1440, there is really no point in thinking in ddr5 presently and probably for the next pair of years. You will also not notice much improvement from tighten timings beyond resolution of 1440.

[@uberhalla]

Thanks! @@Solrac-Siul

[@Winnetou17]

Well, technically, if you play Starcraft 2 at 4K, then the RAM will certainly matter. But for graphically demanding games, the GPU is, by far, the bottleneck, and in those cases it rarely makes any difference. Sometimes a game engine that's not that well optimized might still benefit from it. I'm thinking of Microsoft Flight Simulator for example. But for most new games, the GPU is where the invesments need to be made for 4K gaming.

[@IIHydraII]

That's a b die kit, you can quite comfortably run 4000 16-16-16-32 or 3600 14-14-14-28. Try lowering trrd s/l to 4/6 and tfaw to 16

[@johnfrankster3244]

Nice testing, however it would have been great to see how the ddr5 performs without any XMP profile on at all, non overclocked. Realistically with those kind of specs the only things that really would matter to me is 1% lows. The difference between 156 & 170 to me borderline seems like a waste of time considering that its going to reduce the duration of the hardware. Still would have been useful to see defaults, otherwise very solid content 👍.

[@IIHydraII]

Overclocking memory won't reduce the lifespan, so long as you keep it cool!

[@rodhester2166]

Just my experience I have had with current setup. 7800x3d, 7900xtx and 6400 speed ram, running at 6200 with zero issues.. when running at 6400 it would crash once or twice a day. I have not updated my bios on my asus mobo yet and once I do I will try 6400 to see what happens.

[@emilemil2613]

WHAT MOTHERBOARD U HAVE?

[@fr_danji]

what "tuned" means?😭

[@PatrickH-u8w]

2:50 there is no gear 4 in amd

[@allyjane2545]

I think DDR5 can be really useful for APUs.

[@XX-_-XX420]

I doubt so, as the same is always being said with new ram, but I'm open to be proven wrong.

[@anitaremenarova6662]

TL:DR - 6000 CL30 RAM is good enough. Spending any more on memory is pointless since you could improve your performance a lot more by using the same money for a better CPU or GPU.

[@FIVESTRZ]

I flipped on XMP and was running 7200C34 with the new AMD bios, performance is within the ballpark of tuned RAM without all the work. Went ahead and left primaries and tuned the rest of the timings and saw a nice boost in performance. It was exciting because tuning secondaries doesn’t need nearly as much voltage as working on the primaries.

[@daverhodus]

If I just click on XMP (7200 C34) without touching anything else, will that be worth it? (7800X3D)

[@FIVESTRZ]

@@daverhodus Worth it, ehhh debatable versus tuning it and keeping things 1:1 BUT for me I was able to keep lower RAM voltages while getting the same ballpark performance as my 6200C28 set up. Hate to say look at my channel in someone else's comments, but i did test out the new RAM configuration AFTER a recent bios update that allowed for higher frequencies.

[@jessefisher1809]

I have a 5600 cl36 kit that I overclocked to 6400 cl30. I see that 7200 kits ( that I imagine are hynix A) are pretty cheap right now? Think its worth it for gaming and general computing? I Have a 13600k and a 4090 but I just ordered a 14700k its on the way. That 7200 kit is even on my shitty b760 mobos qvl.

[@FIVESTRZ]

@@jessefisher1809 not bad then. I wouldn’t have expected that on the B series boards. I run my 7200 kits at 7200C32 on my MSI and 7800C36 on AMD

[@Frozoken]

​@@jessefisher1809Less so if u just got a 14700k ironically seeing ur getting some of the benefits mitigated with the extra l3 cacje of that part. Also the QVL means sh*t all on ddr5 and is very misleading. Especially if ur board isn't an msi there's almost no chance that works stock despite the qvl. Look at gskills qvl with a similar speed, that's the actual realistic one and I'd bet it'd be quite eye opening how little boards are there.

[@GainingDespair]

bigger number better is always nice but games are rarely CPU bound these days, and ram rarely plays a serious role in overall performance (assuming you have enough to start with).

[@jameshill6846]

I use 5200 cl42/ 5600cl46 on the lower setting for better timings

[@nebsi4202]

i am really happy and glad that 7200 works on z790 tomahawk wifi i didnt have any stability issues as of yet with it

[@nelju08]

whats your cpu?

[@nebsi4202]

@@nelju08 13900k

[@SVT_LIGHTNING]

Same here. Msi tomahawk z790 13700k Corsair 32gb ddr5 6400 cl32 “CMH32GX5M2B6400C32 Hynix A-die” oc to 7200 cl34 34-41-41-83 1.45v. Been solid. Zero performance gain that I can notice in gaming “gaming at 3440x1440p” but the over all windows experience is faster no doubt. Opening programs and navigating in windows is just snappy and lot lot snappier than before. I haven’t tried going over 7200 but I’m sure it can easily. I have been able to set it to every single speed I’ve tried with out any issues. I didn’t have to up tue voltage until I went for the tighter timings on the 7200. But I’m sure I didn’t need to go to 1.45v from 1.40v but I was getting tired of oc for a while so I set it to 1.45v and the dims never go over 45c in any application / game in my phanteks g500a. So I’ll be leaving it at 1.45v which is low for ddr5 still.

[@daverhodus]

2:44 Please make a video covering the AMD stuff.

[@papahkiki8155]

so, is it ok to just buy the cheapest 4800 for just light gaming?

[@shrexyback3748]

more than ok, most cpus can't even handle the speed beyond 5200mhz, read the specs on the site, no need to overspend beyond that speed

[@amurysmatos2874]

with that motherboard price i can build i new whole pc!

[@BodySculptTV]

Had a 12900K overclocked to 5.3Gz all core running DDR4 4133 at 4000 on gear one. Then I upgraded to the 14900K on the same system and was able to run it at 6.1Ghz all core on the daily still with the DDR4 4000. The system was running great, then I figured it was a crime to run the 14900K on DDR4 so I purchased a new mother board and DDR5 7200. I have to admit, it was a waste of money and time. My 14900K at P6.1Ghz/E4.6Ghz ran happier with the DDR4 4000 on gear one than DDR5 7200. The DDR5 7200 also limited my 14900K overclocking I could no longer hit 6.1 for the daily had to stay with 6Ghz all core. I guess the faster memory puts a strain on the processor. And ultimately I did not see any real gains in games. But funny enough DDR4 now is more expensive than DDR5.

[@Spinelli__]

Running DDR4 at 4000-4300 MHz using dual-rank, b-die sticks at gear 1 equals about 6600-7400 MHz DDR5 sticks at Gear 2.

[@saarhadad]

hi tnx alot , can u test ddr 5200 vs 7200 ?

[@dianaalyssa8726]

Have TG Delta 6000, tuning is annoying lol.

[@ZZstaff]

Thank you.

[@wenshutang2645]

Hi，Could you tell me the T - create is A-die？❤

[@DDD-xx4mg]

Should be good on a 2 dim apex

[@txkingg]

it does if you play escape from tarkov

[@gstormcz]

So DDR5 is upto 5percent faster for higher midrange to top end builds. I am curious how much difference DDR5 vs DDR4 at gaming only. Probably still not worthy.

[@mannyc19]

My systems show a 4 to 5 % difference,Max,often less. Get 5600 c28 or 6000 c30363676 latency,and be happy. r5 3600x/4080FE + 32gb 3200 c14 14 14 31 570 versus r7 7800x3d ram at 5600 c28 34 34 72 525, vs r9 7900 with 32gb 6800 c34 40 40 62 495 tuned is not worth the cash outlay,imho.

[@Spinelli__]

Running DDR4 at 4000-4300 MHz using dual-rank, b-die sticks at gear 1 equals about 6600-7400 MHz DDR5 sticks at Gear 2.

[@951001ify]

me sitting with ddr3 1333mhz watching 7600mhz tests 🙁

[@killerrf]

Good video. Honestly without the little fps ticker in the top corner of the screen that 20-30 fps diff between fastest and slowest ram wouldnt be noticeable to the eye cause your doing 200fps+ already. Biggest fps gain come from gpu then cpu. Ram helps but you could easy spend a week or two just testing and adjusting. Super time consuming. Not like a cpu or gpu overclock where limits are found quickly.

[@thefilmdirector1]

its 100 percent noticeable. I remember going from 16gb 2133 to 3200 and ran fortnite as a test, instantly noticed the extra 40 fps. Now it seems when you go past 3200 the returns diminish.

[@killerrf]

@@thefilmdirector1 some games may respond different and a few people may notice some extra frames. I play games locked at 144, 180-220, then some 250-300. Less than 144 i can see the difference but above that i cant tell without fps counter. Maybe a 300hz monitor might allow you to see the difference better. 2133 mhz ddr4 is much slower than ddr5 6000xmp. So this video is showing diminishing returns arent that much faster than 6000. But to each their own whatever they feel works for them and the amount of money they want to buck up for those few extra frames.

[@Winnetou17]

@@killerrf It's not just the monitor, but also responsiveness. Well, for fast paced games. Though you have to be quite trained to easily tell the difference above 144. Still, maybe if it's stuttery, even at 144 FPS, if it's like swinging between 100 and 200 before and it's 140 to 200 after, that might still be noticeable.

[@impuls60]

@@killerrf In Pubg it makes a huge difference if you have your pc set up correctly. Ultra low latency on, network card moderation to low/off, vsync off etc. Xmp 6000Mhz have higher latency than ddr4 3600xmp so thats why many say they dont notice a difference. The got higher latency on their new pc so its actually worse for many ppl. You need 7000-8000Mhz to have a better system than ddr4 in fps games.

[@2ndLastJedi]

I wish you had tuned BDie just for us still un DDR4, which im sure there is more DDR4 users than DDR5.

[@DannyzReviews]

ya I'm trying to source a ddr4 board to test b-die but honestly I can tell you without testing that if you've got like 4000C14 memory then you're golden and there's no point for you to upgrade because performance will be virtually identical to even the fastest ddr5 config.

[@2ndLastJedi]

@@DannyzReviews sweet, I'm only at 4000c15-16-16-28. I just can't get lower!

[@IIHydraII]

@@2ndLastJedi Tune the tertiaries, they're way more important than the primaries.

[@yougod7253]

DDR-5 sucks. Perpetual training is unacceptable. just like that poison shot.

[@Patrick73787]

This video will become obsolete when Intel 14th Gen comes out.

[@Winnetou17]

Why ?

[@Patrick73787]

@@Winnetou17 Because 14th Gen CPUs (equipped with a better memory controller) will make faster RAM kits (7600+) more trivial to work with, making this current set of data outdated in regards to the state of the Intel Z790 platform and DDR5 memory. He will have to make a new video with new data that will essentially replace this one (good for him a as content creator). By then nobody will care that 13th Gen cannot handle 8000+ RAM flawlessly when 14th Gen exists and is able to do so. Making recommendations at this point in time based on a product that will get replaced in less than 2 months is not wise in my opinion.

[@Winnetou17]

@@Patrick73787 Well, there is diminishing returns. I don't think that being able to run now at 8400 MT/s instead of 7800 MT/s will give more than 2% improvement. And they will still be quite expensive. I don't think that you will be able to run over 8000 MT/s flawlessly on the existing board with any RAM kit. What I really hope with 14th gen is the 4 sticks of memory support. Right now you can barely make them run at 4400 MT/s, if you're lucky. Really terrible (and the same or a bit worse is on AM5 too). This is exactly what I'm waiting now before I get a new laptop. Getting them to be at 6000 MT/s instead of 3600-4400 MT/s will be a massive difference.

[@impuls60]

@@Patrick73787 I so hope ur right, I want to try 9000Mhz 45ns watercooled ram on Pubg. That game have really terrible lows so lower latency is the only way to improve it.

[@XX-_-XX420]

​@@Patrick73787I think you're sadly just way to optimistic about 14th gen but who knows.

[@danarcher3415]

1080P on 4090 wtf?

[@anitaremenarova6662]

There would be smaller if any differences for higher resolutions, it's for testing purposes only.

[@mrfarts5176]

Danny be bustin cheeks.