Is DDR5-6000 CL26 worth the premium over DDR5-6000 CL30?
Posted by SyncFail_@reddit | buildapc | View on Reddit | 106 comments
I'm currently picking parts for my new build, and I'm unsure whether it's worth it in the long run to choose the best RAM possible or just settle for CL30 and forget about it.
-UserRemoved-@reddit
RAM is used by your CPU as fast temporary storage, it plays into CPU performance and doesn't perform on it's own. You never told us what CPU you have, and you never told us what workloads you use it for.
Assuming you have AMD and this is for gaming, then CL30 would be just fine and won't make a noticeable difference.
SyncFail_@reddit (OP)
Oh yes, I'm sorry. I will be running a 9800X3D and the purpose is primarily gaming. I read that low latency RAM affects 1% and 0.1% Lows the most so that's why asked in the first place. My current rig is ancient so I thought if I upgrade everything, I might as well pick the best RAM, but I am still debating.
NippleSauce@reddit
Sorry for the late thread revival...
But you were correct. But it also provides smaller frame deviations on AMD systems - which is definitely beneficial for PvP games.
What did you end up going with?
I'm planning on purchasing the X870E OCF from ASRock, as well as the G.Skill 48GBx2 6000CL26 Royals kit (once it becomes available in the US, presumably at the same time as the OCF board in late Summer).
SyncFail_@reddit (OP)
I went with 64 GB 6000 CL26. It was def. on the expensive side, but I might as well buy decent RAM if I was gonna upgrade anyway. No regrets so far.
buildspacestuff@reddit
It does help a lot with the 1%lows. If your trying to get 4k out of a XX70 series card than will definitely notice smoother gameplay. If your rocking a XX90 series card you probably wont notice
-UserRemoved-@reddit
6000/CL30 is just fine, lower CL is unlikely to make a noticeable difference, and you have massive L3 cache on that CPU anyways.
Content_Camel5336@reddit
What about the 9950x3d? One ccd does not have the 3d vcache.
TryingHard1994@reddit
What about something like a intel ultra 285k? Currently using cl30/6000
Rayrleso@reddit
Intel likes higher memory clocks better,compared to AMD. Can pretty reliably run with 8000MHz RAM
TryingHard1994@reddit
Alright, just tend to see the higher MHz ones have higher CL aswell but is that okay?
lurker17c@reddit
CL is measured in clock cycles, so a higher clocked ram will have a higher CL for the same effective latency (in terms of time)
beirch@reddit
6000MT/s CL30 is the same as 8000MT/s CL40: https://notkyon.moe/ram-latency.htm
Kitayama_8k@reddit
If I'm not mistaken in the past, 3d chips responded much less to improvements in memory latency than other chips. The large 3d cache means your frame rate is less dependent on your cpu waiting for data coming from the ram because the information is loaded into the cache, rather than a cpu with a normal amount of cache that is waiting for the ram more.
Moscato359@reddit
9800x3d hides ram latency, aggressively
breakConcentration@reddit
I’m running this CPU with a 32GB 6400/CL32 setup but I have no 6000 RAM to compare with.
Dressieren@reddit
You also will likely have your fclk changed if you’re running 6400 unless you had manually tuned your timings. Running 6400 at memclk = fclk would likely require you to have your timings quite a bit looser or your IMC is very strong. Most motherboards default to fclk/2 when running above 6000mt.
Either way if you’re running a 6400mt kit there’s a high likelyhood that it’s Hynix A die. It’s just validated for 6400/CL32 from the factory and all of the 6000/CL26 and CL30 are also running A die so they will perform identical (assuming the timings are stable) if you manually put in the timings and sub timings. The only difference is one kit is validated for tight timings and the other is validated to run faster.
Moscato359@reddit
That's the same latency has 6000/cl30
It just has more bandwidth
Rell955@reddit
No with the x3d chip it won’t make a difference because of the extra cache on the cpu.
PsyOmega@reddit
X3D cache makes ram speed completely irrelevant.
https://youtu.be/37f2p9hhrtk?t=1295
Ill-Mastodon-8692@reddit
since x3d ram speed is less important. although I have a 6000cl30 kit, i was testing my old 4800cl40 with it for a bit on the 9800x3d
within 1-2% scores and fps on benchmarks/games
the large L3 x3d cache is why this works so well
so dont worry too much in the grand scheme
MclarenF1Fan106@reddit
Hardware unboxed did a video showing the benefits. AM5 with 9800X3D and 5090 1080P Ultra, DLSS Q
Cyberpunk 2077: Phantom Library
FPS Lower Numbers 1% Lows FPS Higher Numbers Average Frame Rate
DDR5 8000 CL38 (1%low109)-(Avg145) FPS
DDR5 6000 CL26 (1%low107)-(Avg142) FPS
DDR5 6000 CL28 (1% low102)-Avg136) FPS
DDR5 6000 CL30 (1% low97)-(Avg130) FPS
DDR5 6000 CL40 (1%low96) -(Avg129) FPS
DDR5 5600 CL38 (1% low96) -(Avg129) FPS
DerExperte@reddit
Small but very important correction, HU did NOT test with a 9800X3D, they used a 9700X and mention that X3D cpus will show much smaller differences.
https://youtu.be/Fr7Bfr-wPYw?t=274
Alternative_Manner_2@reddit
Thank you. I dont know why everyone thats watched that video thinks its a 9800X3d cpu being used.
ArchimedesNutss@reddit
What about: AMD gaming CL30 vs CL36? 6000mhz
lollipop_anus@reddit
Even 4800mhz stock timings vs 6000mhz CL30 is at most a 10% difference, most cases even less than that. Even less of a difference if you have an x3d CPU. Depending on the fps you are getting we are talking +- a few fps. Biggest difference is the price, get whichever is cheapest.
sSTtssSTts@reddit
Yeah this.
The "big" jump is going from 4800 or 5200 with blah CL to 6000 CL30. After that gains get rather diminishing while costs sky rocket for AMD and it doesn't really make much sense to buy.
DDR5 6000 CL30 is fairly cheap these days too so it does make lots of sense to buy unless you're really scraping to make each dollar go as far as possible.
DDR5 6000 CL30 seems to go for around $80-90 for a 32GB kit. DDR5 4800 CL40 is around $65-75 for a 32GB kit.
For Intel it can make more sense to use fast CUDIMM's but they're real pricey so even there the value doesn't work out very well.
TakaraMiner@reddit
To make the argument clear it is worth it only if you are spending a lot on everything else and plan on overclocking, but in that case, you can get significant performance uplift overclocking to 6400CL26 compared to 6000CL30. Good memory can also help compound the benefits of CPU Overclocking, and if you are already spending $1000+ on a GPU, then it is probably worth it to spend the extra money. I needed higher capacity so ended up settling for 96GB of 6400CL30, but just comparing the stock speed to EXPO enabled, the difference is significant enough to be noticeable without an FPS counter.
birdman829@reddit
On pcpartpicker the cheapest 32gb 6000 cl30 kits are $80
The cheapest 32gb 6000 cl26 kits are $180
There is absolutely no way that any miniscule performance difference between the two will be worth that price delta
thauniverse808@reddit
What about going from 6000cl40 to 6000cl26
vernSdL@reddit
Yes- there is absolutely no way that any miniscule perfofrmance difference between the two will worth that price dealta, i agree with u. no way
LingonberryGreen8881@reddit
Gaming is not the use case for low latency RAM. When gaming at 100fps the CPU has 100ms to complete it's calculations before the GPU will be waiting on it. This is effectively zero. Gaming at 4k with path tracing for example, the CPU stops being a factor.
When emulating or running an SQL server or an MMO server the system will be waiting on the CPU millions of times per second. The CPU in these cases is waiting on the RAM.
NippleSauce@reddit
It creates nearly a 10% performance difference (mostly in 1% lows, but still sometimes close to 10% for average fps), only when you're playing at 1080P.
If you're at 4K, the performance difference (both the 1% lows and average fps) is only around 1-2%.
What I have noticed in my RAM testing thus far on the AM5 platform is that my system always uses around half of my total memory capacity. 32GB (6000CL28), roughly 16-20GB was used; 64GB (6000CL26), roughly 34GB was used - but my 1% lows were much better. So, I will be testing 96GB (6000CL30) to see how much memory is used. Hopefully it's in the 50-53GB range and my 1% lows improve even more. But if that much RAM is in use and my 1% lows don't improve, I'll try modifying the memory timings and upping the voltage to reach CL26 or CL28.
Maverickman1313@reddit
Hello sir. Did you test 96GB with your 1% lows? I'm still undecided on next build if I should go 64GB like current or upgrade to 96.
NippleSauce@reddit
Hey there. It ultimately comes down to your motherboard and CPU for which memory kit will perform better. The 96GB kit came in pretty close second place to a 48GB, 6000CL26 kit of G.Skill Royals. And although the 48GB kit takes more voltage, it's rock stable whilst doing so.
So, although I prefer having more RAM (even when mostly unnecessary for how I generally use my PC), I decided to leave the 48GB kit installed, as it seems better/more immediately responsive for PvP games.
Maverickman1313@reddit
Thanks! That's interesting indeed. Makes sense I guess. Lower timings trumped the additional RAM. But what happened to the 64GB, 6000CL26? That's more RAM and same timings as the 48GB. It was worse than the 48GB? I've usually just left my RAM on XMP profiles thus far, and plan on going EXPO. I'm assuming the timings were tweaked heavily to arrive at your results?
I'm creating a 9950x3d with 870x-Apex for general use/gaming, so probably similar to your use case.
NippleSauce@reddit
Indeed! With the better 1% and 0.1% lows being mostly provided by the lower CAS Latency (CL26). As for everything else when it comes to low CL kits, I generally just use the EXPO timings to play it safe (as I also have a 9950X3D). However, for the 96GB kit, I mimicked the G.Skill Royal's tRFC timings and lowered the CL to 28 due to its lower operating voltage and default EXPO settings (6000CL30, 1.35V). But even still, that kit struggled with CL28 and took a hit to its 1% & 0.1% lows. It did better in those categories when EXPO was left stock at its stock values. Also, I never modify voltages if I don't have to. Lower voltages generally results in cooler operating temperatures (depending on the part and the situation specifics).
As for the 64GB Royals, that kit did indeed perform the worst. However, that is because of my motherboard not supporting more than 48GB of higher voltage kits with a lower CL (X670E Taichi Carrara). A good majority of X670E boards have the same limitation. IIRC, there's only 1 or 2 x670e boards that can perform better with 64GB 6000CL26 kits over the equivalent 48GB kits (can't remember which ones off the top of my head - but I think one of em' was from surprisingly from Gigabyte or MSI).
Maverickman1313@reddit
This info is gold man! Tyvm. I was actually looking at those specific G.skill models of 64GB 6000CL26 1.45v vs 96GB 6000CL28 (new release ofc) 1.35v. My MB supports both/all kits because of the new chipset most likely. These RAM kits came out 2 years after the x670s so it makes sense some of them didn't make the QVL list, heh. That tip to input my MB specs into G.Skill website was helpful, just to confirm the mostly updated QVL list on Asus website. I also read reviews with people saying these kits worked with their x870Es. :D
I'm also getting the Silverstone Icemyst 420 AIO cooler and it has these 3 attachable fans (onto the block) that cool the MB components including the RAM. It's very neat, you should look it up. Anyway, do you think it's worth buying the 96GB 6000CL28 and trying to get CL26 on it by pushing the voltage? lol. Because the 64CL26 kit already starts at 1.45v. I'm not too privy on how well RAM can be pushed, especially at these higher capacities. It's going to be 1 of those 2 kits that I end up getting, 99% sure.
NippleSauce@reddit
No problem, friend!
Regarding the 64GB CL26 kit; I didn't try running it at CL30 to compare it to the 96GB CL30 kit. I figured I'd be better off returning it to Newegg since it was around twice the price of the 96GB kit =x.
One thing I forgot to specify about the G.Skill website's RAM config page and DDR5 RAM in general:
Memory on the list with the correct motherboard and CPU series selected just shows memory that should have a 100% chance of working with your setup. And whilst that doesn't necessarily mean that a memory kit that isn't on the list won't work, it (in my experience) also means that it won't work as good as the fastest memory kit that is on the G.Skill list.
As for the custom timings and the voltage differences between the kits that you mentioned; this is likely due to memory age, the different capacities, and the memory sticks being manufactured by different memory fabricators (Hynix, Micron, Samsung). With the fabrication process being slowly refined, the DDR5 timings are being gradually tightened. With more precise/enhanced fabrication, the more tight the timings can be whilst maintaining stability. But, more voltage may be required to push those tighter timings..... And luckily with enhanced fabrication, more voltage won't hurt stability =). But unfortunately with the 96GB kits, the voltage can't be safely pushed much further than it currently is with EXPO due to the increased capacity. Anything above 1.4V would likely cause instability, whilst also not being enough power for CL26 (and I would personally be a little fearful of an OC'd modified EXPO memory kit causing my 9950X3D to fail due to AMD and ASRock stating that the 9800X3D failures were due to memory incompatibility issues).
So - even with an X870E board with its improved memory controller - I do not think you'll be able to push a 96GB kit to 6000CL26 due to the voltage requirements for CL26 to maintain stability. I tried pushing my 96GB 6000CL30 kit to CL26, however, it also required 1.45V to get there. But since all of the 96GB kits run EXPO at 1.35V, I imagine that doing so might result in a system crash if pushed to 1.45V - as that is what happened to my system. Granted, you might be able to achieve this with a 96GB 6000CL28 Royal kit on your board. However, I still don't think it would accept the presumably needed 1.45V for CL26 graciously. And so overall, I don't think it would be worth trying unless you're a Newegg Egg member (in which case you can purchase the 96GB CL28 Royal kit AND the 64GB CL26 kit, test both, and then return the one that performs worse or doesn't work). But again, I don't think it would be possible to reach CL26 and stay stable on a 96GB kit at the moment.
Of course, please let me know the general results if you give this a shot! I'd be super interested in hearing the outcome!
Maverickman1313@reddit
Thank you sir, lovely info. Unfortunate about the CL30 comparisons of 64 vs 96. That would be fun to see the diffs, but I'm guessing diminishing 1% returns at that point and you lose timings anyway. So all good.
I see what you're saying with the 96GB Kit. There's just too much capacity that CL26 could be very unlikely and it still might be worse than Stock EXPO. Might be too much hassle. But I saw this guy's vid - https://www.youtube.com/watch?v=uG2xCEUAo40 - and he increased the speed of his 64GB 6000CL26 to 6200 while increasing Fclk as well. This is probably the route I should go and just do the same thing as him. Increasing the speed +200MHz would be more practical than lowering the timings by 2 (I assume so, very rookie at RAM OC).
I think some people might find this convo interesting when they research this topic, so cheers. :0
NeighborhoodOdd9584@reddit
Yeah just depends if you want to spend double to get better 1% lows. It’d better but not by that much. I have the Gskill royal Neo 6000CL26.
Lewdeology@reddit
How big of a difference are we talking about?
NeighborhoodOdd9584@reddit
Probably a few frames more
basement-thug@reddit
Hell I'm still running CL36 , and everything I've researched says spending money on CL30 is not worth it. AMD AM5 isn't really benefitting from the lower latency unless you're someone whose hobby is benchmarking and you want those 200 extra points... for typical usage it appears negligible at best.
sauceman_a@reddit
this is a cope- check out https://www.youtube.com/watch?v=Fr7Bfr-wPYw
basement-thug@reddit
Yeah I didn't get to watch the whole thing yet, it only came out a day ago, but skipping through the charts it doesn't look like there's much difference. Single digit percentages mostly. About what I'd expect.
sauceman_a@reddit
lol what- go watch the video- 10% fps increase in some cases is NOT insignificant
basement-thug@reddit
10% max in specific cases is not significant. Especially when you spend way more than 10% more to get it.
sauceman_a@reddit
like I said nice cope. 10% is huge for competitive gamers who want to squeeze out every last frame
basement-thug@reddit
It's not a cope because I clearly stated in my initial comment, paraphrasing, "unless you're really into benchmarking and trying to chase those extra 200 points" ... so great, you fall into that category. My point was for most users it is not worth it and that still holds true for the day to day normal gamer. So I guess congrats on providing input that supports my point.
Ill_Computer_1028@reddit
honestly just go with a cl 30 kit that has Hynix a die like cl30 flare x5 6000mt . currently have mine overclocked to 6400 m/t cl28-38-32-30 idk was easy to overclock from you tube videos. i would only get the cl26 6000m/t if you are looking for overclocking scores because they are going to be higher bin memory chips and most likely will support even tighter timings. So unless you spend half your life in your bios you just need cl30 6000m/t
hfcobra@reddit
I read that average FPS will be nearly unaffected, but you may have noticeable improvements in 1% lows or 0.1% lows in very high FPS titles like CS2, Valorant, or R6S.
I have a 6000 CL26 kit on the way and I have a ton of benchmark data from my 6000 CL30 kit I already have. If you want to wait I'll have some actual data for CS2 by Friday.
Realistic-Yak-5316@reddit
What’s the verdict big guy 26 or dirty thirty
hfcobra@reddit
Just want you to know that I still haven't received my RAM sticks. Not happy but it is what it is. I will post results once I get them. I won't leave you hanging. 👍
Realistic-Yak-5316@reddit
Thank you I’m waiting on you to complete my purchase 😭😭
hfcobra@reddit
Hey so I got the RAM and ran a few tests today. I did break my all time best test with the new RAM by just using EXPO. No other tuning.
0.1% lows improved by around 4%, frame consistency went from 99.94% within 2ms margin with 6000CL30 to 99.98% within 2ms margin with the 6000CL26.
This is at a frame cap of 482FPS and a 3080Ti in CS2. I think that if I had a better GPU things would improve more, but this is what it is for now.
I plan to tune the RAM a bit more and see if 6400CL28 is possible and if that gives better results.
Realistic-Yak-5316@reddit
Niceeeeee I’m gonna buy the cl 26 and return the cl 30 which brand cl 26 you have ? Thanks for the info the lows are much better!!
hfcobra@reddit
G.Skill is the only brand for RAM imo.
Realistic-Yak-5316@reddit
Thanks bro!! Just bought it!
hfcobra@reddit
The shipment was delayed until tomorrow. 😒
Camluiam_@reddit
HWunboxed tests cl26 kit with 9950X3D go to 21:12 and they show graphs comparing cl26 and cl30 kits
Vandervenn@reddit
Short answer: NO
szethSon1@reddit
Wait, the lower the CL the better the ram??.... Oh sht 😭😭🤣🤣🤣
Stonesneakers@reddit
No, even cl36 makes no difference (in gaming, I don't know for anything else)
LordAlfredo@reddit
Subtiming optimization may make the 30 kit better than the 26 anyways depending on actual memory die and silicon lottery.
Old_Resident8050@reddit
M/B bios have the option to train into higher bandwidth and/or timings. I guess for 9800x3d there is more benefit going the bandwidth way then?
sSTtssSTts@reddit
Nope.
Bandwidth won't be a limiting factor for x3D chips either.
Pump you clocks as high as you can get away with and call it good.
ecktt@reddit
I cannot think of a situation where it would be worth it other than a VERY memory latency sensitive application.
For the same money as CL26 you can get a 64 GB CL30 kit that is dual rank to make up the performance difference AND have x2 as much RAM. So not it is not worth it.
If you intend to use an X3D chip, the whole point of the extra cache is to make up for slow ram so; No it is not worth it.
SyncFail_@reddit (OP)
Performance difference with 64GB Dual Rank is a very interesting observation. I wonder if Dual Rank 64GB closes the small gap between CL30 and CL26. The only disadvantage I see is that you take away the opportunity to OC your RAM to 8000 MT/s+ with the 9000 series since the IMC can't handle it well.
ecktt@reddit
Let say you are lucky and you have a golden sample SOC on the 9000 series. The best you can hope for without actually degrading the performance is 6400MT at CL32. But at that point you will be pumping the SoC to 1.2+ Volts and it will burn up eventually. At 8000MT, it will perform worse than 6000MT/s.
sSTtssSTts@reddit
From what I recall Buildzoid has a vSOC of 1.2v is fine. Its when you get to 1.3v+ that you can burn them out.
fredgum@reddit
On a X3D CPU it will make no difference: https://www.youtube.com/watch?v=37f2p9hhrtk&t=1272s
SyncFail_@reddit (OP)
Okay, that part of the video is really interesting. I didn't know the difference would be THAT insignificant. I guess the X3D cache is doing a fine job. I think I will stick with DDR5 6000 CL30 and just try to overclock it manually or whatever the DIMMS are willing to offer.
birdman829@reddit
Just turn on EXPO and literally never think about it again
sSTtssSTts@reddit
Exactly this.
Unless you're aiming for some record OC or trying to compete with Buildzoid or whatever just turn on EXPO and forget about it.
RAM overclocking in general is very tedious and frustrating (you have to test EACH setting as you change, look at how many secondary and tertiary timings there are!!) and the gains you can get at best are not going to be all that big short of resorting to LN2 cooling.
Hollowsong@reddit
With DDR5 6000 anything, you won't need to overclock.
Hell, letting it run at factory 4000 MHz is probably fine.
RAM is so crazy fast compared to what it used to be, few gains are made at the high speeds. Normally your CPU and GPU take over anything "weighty" for gaming.
Leave it base clockspeed and don't even worry about CL anything. You won't even notice.
firestar268@reddit
You'll hardly notice the difference. Spend the extra in CPU/GPU
Artistic-Savings-239@reddit
No, that money could be much better spent elsewhere so unless you have a 5090 and 9800x3d just get cl30
Redddittorio@reddit
Go with CL28, small price premium compared to 26
JakeBeezy@reddit
I can barely even find CL28 ddr5 RAM, actually I couldn't locate them at all, what am I doing wrong ?
ApacheAttackChopperQ@reddit
KingBank 6000 CL28. Paid less than 90 shipped. Very nice kit so far, used over a month with no problems.
Redddittorio@reddit
Where is it sold? I couldn’t find it earlier, IIRC the secondary timings are tighter
ApacheAttackChopperQ@reddit
Amazon had them. Aliexpress too.
Redddittorio@reddit
https://www.microcenter.com/product/689351/gskill-ripjaws-m5-neo-rgb-series-32gb-(2-x-16gb)-ddr5-6000-pc5-48000-cl28-dual-channel-desktop-memory-kit-f5-6000j2836g16gx2-rm5nrkw-white
SyncFail_@reddit (OP)
Those look very interesting. The price delta is big, compared to CL26, and perhaps you might even be able to OC those to CL26
Calm_Income6781@reddit
DDR5-6000 I think cl32 or faster is fine. I don't think you can tell the difference between that and faster bins. cl36 IMHO is just too much of a lower binned grade.
Turtledontist@reddit
You'd be better off with 6400 CL32. Marginally better than 6000 CL30 in gaming for the same price.
KillEvilThings@reddit
No
Fredasa@reddit
Only, and I repeat, only if the single biggest complaint you have with your performance is microstutter.
Pointing this out because microstutter is consistently the biggest problem I always try to squash, and even I didn't bite on those overpriced sticks.
CtrlAltDesolate@reddit
Unless the premiums like $10-20 on a like-for-like set, no.
LingonberryGreen8881@reddit
Gaming is not the use case for low latency RAM. When gaming at 100fps the CPU has 100ms to complete it's calculations before the GPU will be waiting on it. Gaming at 4k with path tracing for example, the CPU stops being a factor.
When emulating or running an SQL server or an MMO server the system will be waiting on the CPU millions of times per second. The CPU in these cases is waiting on the RAM.
artificial_me@reddit
Probably not.
Achillies2heel@reddit
It's like a sub 5% performance improvement. So no it's not worth double the price.
IntrepidContender@reddit
I just ordered some CL26 - So I really don't know but I feel like why not get CL26 / CL28 if you're really into performance
ScornedSloth@reddit
The short answer is no. You're talking about maybe 1-2 fps on a good day. You can actually squeeze more performance out of higher speed kits at higher latency, but again, you're talking negligible performance gains for gaming.
Scytian@reddit
If it's 5$ premium then yes, 10$ maybe, anything more is not worth it unless like to overclock.
Homewra@reddit
Meanwhile i just settled for DDR-6000 CL36 :(
Ill-Mastodon-8692@reddit
I tried 4800CL40 temporarily vs my 6000cl30 kit, on 9800x3d
like 1-2% difference tops, and most of the time not noticeable.
X3D cache makes the ram speed not really matter much
Vashelot@reddit
Not a big issue you lose lik 1 or 2 percent of performance.
Every_Gold4726@reddit
So you're looking at about a 1.33 nanosecond improvement with the CL26.
In real-world usage: Gaming: Typically 0-5% FPS improvement, more noticeable in CPU-bound scenarios Productivity: 1-3% improvement in apps like video editing, rendering
filthmcnasty1@reddit
That 6000 cl26 is really only a good purchase if you plan on overclocking it. And even then, it's not a noticeable difference.
Celcius_87@reddit
I was doing that same comparison lately too and it’s not worth it
Zuokula@reddit
The silicon lottery will probably give much more variation than the cl26
TimmmyTurner@reddit
if you are running any x3d cpu, then it wont matter much go for cl30
itpointz@reddit
On an X3d chip cl30 is the money spot. Can you squeeze a few fps, maybe, but you'd need to be running a 5090 to make it noticeable in benchmarks, not necessarily noticeable in gameplay
Stargate_1@reddit
It'll be a 1% difference, if even, depending on the game and how RAM sensitive it is.
If the price is not the same or almost the same, it's not worth buying the CL26, that kind of RAM is really meant for benchmark chasers who very much do care about the difference between 150 and 151 fps
GNRZMC@reddit
It shouldn't make any noticeable difference. Low single digit percentage points in absolute best case scenario. RAM is definitely a case where once you hit "good enough" you don't see much gains after