Hey guys, we've now updated the testing to include a DDR4-3600 CL14-15-15-35 kit as well. You can find that data in the TechSpot version here: www.techspot.com/review/2529-ddr5-ram-for-budget-pc-build/

@chovekb

Жыл бұрын

Cool but the fastest setup of DDR4 3600 cl14 DR is missing xD

@fatcatpeople

Жыл бұрын

I need your help. Im upgrading from a 1155 Intel Motherboard, I know super old 😂. I want a z790, can you recommend a good one for under $300. Im thanking msi Z790 a pro. I'll end up holding onto it for 10yrs. I held onto that 1155 forever 😂. Im getting 32gigs, intel 13700k and right now I have a 2080 super i plan on keeping. ✌🏾

Non-K 12th gen CPUs have a VCCSA locked to 0.95 Volts. This prevents those CPUs from running anything past ~3600MHz in gear 1. That 4000MHz RAM kit was running in gear 2, hence the low performance.

@TrueThanny

Жыл бұрын

And? Since that's the CPU being tested, and those are its limitations, the test represents reality.

@WrexBF

Жыл бұрын

@@TrueThanny You clearly did not understand what I wrote.

@ab-lymphocite5464

Жыл бұрын

I thought that was only in H610? I thought B660 and H670 had full ram tuning and OC support.

@lechf

Жыл бұрын

@@CrispyLimit the controller runs at half speed in gear 2 mode crippling performance

@TrueThanny

Жыл бұрын

@@WrexBF No, you clearly failed to understand the complete lack of relevance of what you wrote.

No wonder DDR4 4000 isn't much faster than DDR5 4800, you're running it in Gear 2, wtf. The only good thing still going for DDR4 is the fact that you can run it in Gear 1 for much lower latency and you don't use it, lmao.

@ActuallyHardcoreOverclocking

Жыл бұрын

non-K intel 12th gen chips can't run more than 3600 in gear1

@Yahfz

Жыл бұрын

@@ActuallyHardcoreOverclocking That's the point, doing the highest your CPU can in gear 1 if you're using DDR4.

@thepatriot6966

Жыл бұрын

I'm running my DDR4 Patriot Viper Steels at 4000mhz 15, 16, 16, 21, 36. Latency is 51.2ns. GDM Off (1T.) Along with my Ryzen 5600x and GTX 1080Ti with EK Water block. Good enough for me for the next few years. At least until. AM5 and DDR 5 matures. No need to upgrade until then.

@KryssN1

Жыл бұрын

@@Yahfz You are responding to a longtime hardcore Overclocker in the community and telling him how it all works. It's funny. Just like Buldzoid said, non-K Intel 12Gen CPUs cannot run more than 3600 MT/s in Gear 1, especially early samples that HUB have. Those even early on the release had issues running 3600 MT/s in Gear 1, like 12400 or 12500 SKUs. Also 3200Mhz CL14 is equal to 3600 Mhz CL 16 or 4000 CL18 / CL19 in most games and sometimes even faster (for games that like low latency), so it does not even matter. You would probably see some benefit of going with 4000 CL 15-16 though with a Samsung B die, but then (if mobo cost is not much diffrent) just look for DDR5 6000-6400 CL37/40 kits. I have the same approach as @The Patriot , waitng for DDR5 and AM5 mobo to get cheap and upgrade from Zen3D in 2 years to Zen 5 V cache.

@Yahfz

Жыл бұрын

​@@KryssN1 Do you really think you have to be a hardcore overclocker to know how gear modes work? My comment was just to reaffirm that me and BZ were on the same page, not to tell him anything that he doesn't already know (this is really basic knowledge afterall). Your comment makes absolutely 0 sense.

[Insert comment complaining about not buying DDR5]

@ShockingPikachu

Жыл бұрын

Same thing happened between ddr3 and ddr4 growing pains

@TheAETHER22

Жыл бұрын

Price

@SlenderHime

Жыл бұрын

It's actually okay to complain

@HazzyDevil

Жыл бұрын

@@SlenderHime yeah if you want to complain go for it. But I’m specifically making fun of those who don’t watch the video and then complain in the comments. This is about building a new PC, not upgrading an existing one.

@SlenderHime

Жыл бұрын

@@HazzyDevil maybe those new builders are from non western countries and they are complaining because the price of cheapest DDR5 and mobo compatible with DDR5 outside western countries are still ridiculous high? Because in Indonesia for example , it's definitely is right now

I'd rather take the $200 saved from a DDR4 system and build with an i5 12600k - people rarely upgrade their memory kits. They build their system with enough memory and by the time they need more capacity, it's time to build a new computer with faster memory speeds

An obvious disclamer: This greatly depends on prices in your region. In mine, buying the cheapest 2x8GB DDR5 is 2.4 times more than a 2x8GB DDR4 3200 kit

Reminds me of the 6700k basically being the same as the 4790k with early DDR4 and then pulling away once there was better DDR4 available a year later.

@rayhmann

Жыл бұрын

Not really, that 6400mhz memory is just slightly better. Also, scenario with 3090ti on medium is unrealistic. Ultra would further diminish gains of ddr5

@ShadowMKII

Жыл бұрын

It was only relatively slight, though, yes? A lot of the data I have seen for the 6700- and 7700K only showed about a 5-10% performance improvement over the 4790K.

@Dark.Syndicate

Жыл бұрын

@@rayhmann well using a 12100 with a 3090 ti probably won't happen either but then how would they show differences unless they remove gpu bottlenecks??

@rayhmann

Жыл бұрын

@@Dark.Syndicate only proves my point even further

@m8x425

Жыл бұрын

Yeah.... for sure. I remember seeing those launch day 6700k tests which made Skylake look unimpressive. By 2016, the 6700k was pulling ahead of everything in benchmarks. What this really reminds me of is running 1866MT/s or 2133MT/s memory with Sandy Bridge. When Sandy Bridge was new there were no games that could take advantage of RAM that was clocked above 1333MT/s, but after a few years games emerged that would run better with faster memory. I ended up with a 3770k paired with a kit of DDR3-2400 RAM and running the RAM in XMP made a big difference in some games.

This is exactly the follow-up video I was looking for. Thanks.

Your channel has really taken off the last couple of years, your content is absolutely fundamental for a PC gamer. Thanks guys.

@Fermen80

Жыл бұрын

Agreed. Sometime to much pro AMD biased but still one of my favorite tech channel.

@GTTMRC95

Жыл бұрын

@@6210classick a mf can't even compliment a couple of dudes working their asses off that immediately comes another mf calling him a bot Smh

@AdamWebb1982

Жыл бұрын

@@Fermen80 to be fair its pro whatever is the best. (performance + value)

@christopherjames9843

Жыл бұрын

@@Fermen80 If you think they are biased you don't know Steve or Tim very well.

@AnabolicSaagAloo

Жыл бұрын

🤣

Thanks, good job as always. Loved the “test system specs” sheet

Seems weird not to test ddr4 3200 cl16 given that it is the most popular budget option and that it costs nearly half as much per gygabyte as Ddr5.

@gameison4813

Жыл бұрын

Why not add b-die overclocking as well. 3733-3800 cl 14-14-14-28-42 with low trfc and other timings?

@christopherjames9843

Жыл бұрын

@@gameison4813 Because hardly anybody tweaks their memory besides setting XMP. Most people don't even do that. Add to the fact b-die sticks are expensive if you want 2x16 which defeats the purpose.

I have B-die 4000mhz ddr4 steel viper ram. Amazing things; got them manually overclocked to 3766cl14-14-14-28 paired with my 5800x. I'll def get ddr5 when am5 and intel matures with ddr5 as well. First gen doesn't really appeal to me, more bang for buck skipping a gen or two especially when we have more than enough performance on our pc parts these days.

@Trancelistic

Жыл бұрын

Nice, I bought the 3466 mhz B-die sims back then . They are nice chips.

@surpriseblueviana3803

Жыл бұрын

Must youtubers dont have a clue about tunning RAM to that level.

@Levent_Ergun

Жыл бұрын

@@surpriseblueviana3803 it is really not that hard. Just takes A LOT of time (talking weeks atleast) to fully tune it and a bit of luck. My 5800X somehow has a worse IMC than my 3600.

@BNR_248

Жыл бұрын

@@surpriseblueviana3803 Agree, lots of performance left on the table. If buildzoid did gaming benchmarks based on his overclocks, he'd draw quite a crowd. There's not much info on heavily/properly oced ram gaming benchmarks.

@surpriseblueviana3803

Жыл бұрын

@@Levent_Ergun I put mine 3600mhz cl17 to 3733mhz cl16 love tunning a little

I still dont think ddr5 makes sense for a build at this budget. A 12100F build with DDR5 4800 CL40 is 340.96 USD, while a DDR4 build with the same cpu, ddr4 board version and 2x8 of 3600 cl18 is 278.97. Frankly 3200 CL14 makes zero sense as a purchase when the cheapest is 72.99 vs the 47.99 3600 CL18 kit. So no, DDR5 does not make sense here.

@cjsawinski

Жыл бұрын

And it’s cherry picked ram and the wrong cpu to run the tests on… the 4000 cl19 ddr?!? And in order to even take his point seriously about “future proofing” you need a $600 MOBO for ddr5… all the “budget” boards are capped around 6000 for ddr5. Just bad bad advice

@ShadowMKII

Жыл бұрын

DDR4 3200 CL14 compared to 3600 CL18 is not an apples-to-apples comparison (the best comparison would be 3600 CL*16*), but I agree nonetheless. I think things have improved, but not enough to warrant a purchase like that for the price - at least for Intel. For AMD and its AM5 platform, DDR5 seems alright, but even that's challenging to justify - especially with prices possibly decreasing significantly after the release of Zen4. This might be a different story in six months to a year, however - with prices being good enough to warrant not buying a Zen3-based computer; opting for Zen4 instead.

@lookitsrain9552

Жыл бұрын

@@ShadowMKII Yea i agree, but atm 3600 cl18 is just a few dollars more vs 3200 cl16 so its a good deal.

@ShadowMKII

Жыл бұрын

@@lookitsrain9552 Gotcha. I'm not sure if that's the case where I live, but that's fair if that's true.

I did a 12400 with 3200cl16. It was so cheap and most everything I play is going to either be GPU bound or my monitor doesn't refresh that fast. I may swap out the motherboard, cpu and ram in a couple years but very happy with what I have for now.

@Chris-hw4mq

Жыл бұрын

if you will get a rtx 4060 next year it will be great for a couple of years

@dereksmith6713

Жыл бұрын

I did the same, I would argue that going from 16gb to 32gb of ddr4 would give the vast majority of us the performance boost we want. Whatever you do get 32gb going forward.

@Dark.Syndicate

Жыл бұрын

@@Chris-hw4mq u don't know how good or bad a 4060 may be. or how expensive. maybe amd will be better. keep ur eyes open when buying.

@Chris-hw4mq

Жыл бұрын

@@Dark.Syndicate if we know the evolution of 60 series cards it will be on the level of a 3080

@siphi7583

Жыл бұрын

Where I live the super cheap 4800mhz ddr5 memory the guy on the video uses is like 50$usd the kit 2x8gb.

You need to use tighter timings on ddr4 4000. Your kit was also in gear 2 in bios Essentially making it a 2000 mhz kit. The reason why 3200 cl14 was beating 4000 in every game. You should get buildzoid to help you out with some of these ram videos unless there’s some other agenda

@cptnsx

Жыл бұрын

No agenda dummy - they are using the USE CASE of a regular Joe that simply enables XMP and lets the motherboard decide. No TWEAKING.

@siphi7583

Жыл бұрын

$$$$ is the agenda.

@elus89

Жыл бұрын

Realistically, 90% of the audience is going to do exactly as Steve did. If it's essentially a 2000 mhz kit, how come it beat out the 2400 mt/s kit that had lower timings? The 3200 cl14 kit was essentially the best case scenario, and they put it in the test (despite a bunch of 'smart people' thinking the higher mt/s automatically = better performance).

@mircomputers

Жыл бұрын

That is a 2000mhz kit, even at gear 1. But yes gear2 gives much worst latency

3200-16 and 3600-18 are probably more representative of what someone looking to save money by going with DDR4 vs DDR5 would go with. If you are willing to pay the ~$50 premium to go with low-latency DDR4, most of DDR4's cost-saving argument is already gone. The real entry-level fun with DDR5 will start next year where next-gen kits will bring 4800-30 or better to the masses.

Your camera work is amazing as always. Enjoy watching!

I really hope all this 12th gen with DDR5 testing low-key means we're gonna get Zen 4 benchmarks soon? Love your content as always Steve, this will definitely help me decide my next build

@Chris-hw4mq

Жыл бұрын

embargo on zen 4 lifts on 26 september

@edue_mrb

Жыл бұрын

@@Chris-hw4mq cool, didn't know there was already a date! Can't wait to see how Zen 4 does

@odizzido

Жыл бұрын

@@Chris-hw4mq Oh, that's good then. Retail availability will be on the 29th I think so we get benchmarks early which is pretty pro-consumer.

@felipe7466

Жыл бұрын

@@Chris-hw4mq isn't on 27 of september?

So proud to be a Floatplane member. You guys rock!

For the Ryzen 7000 review, I'd love to see a comparison between 5600CL36, which seems to be the sweet spot now, and 6400CL32, to understand how much performance will be lost at maximum. I won't expect many people who buy a $300 and up CPU to go for a 16GB DDR5-4800 kit. 32GB 5600 sounds like a more reasonable choice.

@hiimcortana1568

Жыл бұрын

Isn't 6000 the sweet spot for perf/dollar?

@ETophales

Жыл бұрын

@@hiimcortana1568 I guess it depends where. Checking Newegg prices, 5600 is $160 for 32GB, while the cheapest 6000 is $218.

@nimbulan2020

Жыл бұрын

Not to mention that 32 GB of this basic 4800 memory is the same price as the premium 5600.

@GTTMRC95

Жыл бұрын

AMD said that 6000 MHz is the sweet spot to achieve that 1:1:1 ratio between core, cache and RAM, just like 3600 MHz was on AM4.

@GTFour

Жыл бұрын

@@GTTMRC95 only if you kept CPU FCLK speed at stock 1800mhz for 1:1, if you pushed that to 2000mhz you needed 4000mhz RAM for 1:1.

Dude you have 4000 sticks being slower than 3200 sticks...you're doing it wrong lol

Thanks Steve, I was just asking this exact question the other day, and you answered it perfectly for me.

@libertyprime9307

Жыл бұрын

Well he answered it wrong. This is the worst video HU ever made.

If it’s similar performance and similar price and you need one or the other you may as well go with the new platform now so you can upgrade easier down the road

We need the full ram timings for all of these… tuning goes a long way as does holds back alot for ram comparisons. Especially when its plugNplay

@elus89

Жыл бұрын

@@fluphybunny930 He's doing what 99% of users would do. Plug n Play, bigger numbers = better.

Great info! Thank you!

I don't plan to do a build in the near future but your recommendation has me thinking to just go for it. It should be interesting seeing the benchmarks for AMD and Intel on their new CPUs and then I'll think more about it.

@amirhb7531

Жыл бұрын

Id recommend waiting until new motherboards arrive it only makes sense to go full on next gen with 13th

@YataLocK

Жыл бұрын

You're going to upgrade an entire build just for 10% performance at 1080p in a couple games? Lol

@movdqa

Жыл бұрын

@@YataLocK I'm not a gamer. My current Windows system is an i7-10700, 128 GB RAM, 3 TB NVMe SSD, GTX 1050 Ti.

@YataLocK

Жыл бұрын

@@movdqa what part of this video persuaded you to upgrade? Lol it's all gaming benchmarks

@movdqa

Жыл бұрын

@@YataLocK I'm not upgrading anytime soon. But, when I do, it will be with DDR5. I didn't actually look at any of the benchmarks. I just appreciated his opinion on being a good time to move up. My 10th gen motherboard was no good after 11th gen anyways. I was rather annoyed that it was only good for two generations.

Exactly what I am looking for. Thanks!

Hey Steve did you see the Frame Chasers response video to this DDR5 For Budget CPU's video?

@Hardwareunboxed

Жыл бұрын

I said I wouldn't watch that 'garbage' in the this comment originally. I should have been more tactful and respectful. Rather I should have just said I'm not interested in watching another react video to our content.

@LiquidSnake690

Жыл бұрын

@@Hardwareunboxed Well you might want to watch it considering he's claiming you were misleading with your DDR4 vs DDR5 comparison. He said you're comparing the best DDR5 ram kit vs the worst DDR4 4000 ram kit. He said the DDR4 4000 was running in gear 2. He said when DDR5 6400 ram is benchmarked on lowend locked intel cpu's errors don't pop up but if you run games for a long session the PC will crash. FrameChaser isn't the only youtuber who reacted to your video. The Good Old Gamer also reacted to your ram comparison and also noticed you made mistakes with your ram.

@Hardwareunboxed

Жыл бұрын

@SolidSnake680 there were no mistakes made. I'm sorry you've taken their word as gospel.

@LiquidSnake690

Жыл бұрын

@@Hardwareunboxed I'm not taking their word as gospel I'm just trying to get accurate honest information.

@Hardwareunboxed

Жыл бұрын

@@LiquidSnake690 DDR4-3600 CL14-15-15-35 (so the best DDR4-3600) memory is 3% faster than the DR 3200 stuff shown in the video. That's the difference between what was shown here and the very best DDR4 memory loading just XMP. We will include that data in the TechSpot written review. But yes it is a shame that people were lead to believe I was doing something wrong or were tying to mislead them. Chris is a low life as far as I'm concerned, he could have easily messaged me for more information but he sold out a friend for a few views, sad.

Hey Steve. I think you should do another round of benchmarks with the same budget DDR5 4800 memory but this time, tighten the timings and see how much extra performance you can squeeze out of the ram. I have a feeling it could drastically increase performance if you do so and would be a good video as it is one of the cheapest DDR5 sticks you can buy and showing how much extra performance can be had by redoing the timings would be great to watch. Please consider it.

buying cheap ddr5 now and "better" ddr5 later seems questionable for anyone on a budget

id like to see some non frame rate testing. that the factorio benchmark you include sometimes, or civ, or even any paradox game

@theplayerofus319

Жыл бұрын

For sure

Thanks a lot for the video guys! I'm building a new PC with AM5 so all this info related to DDR5 is super valuable!

@andersjjensen

Жыл бұрын

AMD has already confirmed that 6000 is the sweet spot for gaming performance. That's the highest you can consistently go in 1:1:1 ratio. From what I've seen so far even the cheaper SK Hynix and Samsung (avoid Micron!) based 5400 kits will do 6000 and reduced timings with a bit of voltage. And anyone with a fully populated mesh front case shouldn't be afraid of giving the sticks a bit more voltage. The JEDEC specs on wattage are designed for Dell-style shitboxes that cooks their internals because apparently making air holes in things is hard.

Love your VLDL-Miniaturs!

DDR5 has come down a lot in price over the past 6 months. I just built a system with a 12400F and decided to go DDR5 to give myself a better upgrade path (like you said) because 2x16GB DDR5 5600 CL36 was only $35 more than 2x16GB DDR4 3600 CL16, both kits Corsair Vengeance.

@adrianafamilymember6427

Жыл бұрын

Might have to end til' the end of the month to see if anything comes out of it.

@siphi7583

Жыл бұрын

I threw my 3600mhz DDR4 to the trash can and got myself some 6000mhz DDR5, this is the future.

@Momo.momo789

Жыл бұрын

On top of that any Hynix 5200-5600 DDR5 can easily oc to 6800@CL34 to get top of the line performance. while a pair of B-die 3200@CL14 32G cost more than DDR5 counterparts.

@ppn7

Жыл бұрын

Have you any example of testing that OC ? I want to buy a new set up. How can I know if I get Hynix kit ?

@crylune

Жыл бұрын

@@siphi7583 Eh, in my case it'll be the future once Zen 5 / Meteor Lake is out and I do a full rebuild. My 3600 CL16 B-Die is still fine. Upgrading now would be silly.

Thx for this test

Thank you so much for your great videos, I watch every single one your videos since I found your channel 4 years ago I believe. Would you be nice if you made a similar video about PCI express. Cheers from Canada!

Why are you testing DDR4 in gear2? Of course it'll be slower than 3200mhz in gear1... If you take 5min and tune that 3200c14 b-die kit to max out the IMC in gear1. i'll be faster than your 6400mhz ddr5 stock... The least you guys could do would be to use 3600mhz stock. As that's the same price and runs gear1 on locked Intel cpus.

@SKHYJINX

Жыл бұрын

I'm surprised ALL his platforms on DDR4 can't do 3600 14 14 14 DR. As a MINSPEC for testing I mean. Ryzen1/2 compatibility testing or "official spec 3200" seems off to me in today's market where 3600C14 DR is EZ for the platforms today.

@Miguel-rb8mi

Жыл бұрын

​@@SKHYJINX All 3200c14 b-die kits can do 3600mhz c14-15 easy. Also, 3600mhz is the same price as 3200mhz. If you want to run stock xmp makes no sense at all to buy 3200mhz. Anything over 3600mhz on intel and you'll be in gear2 and take a huge latency penalty. With a K SKU you can probably run up to 4000mhz in gear1 without much hassle.

@Hardwareunboxed

Жыл бұрын

@HYJINX because we know the sweet spot is DDR4-3600 CL16 which is slightly slower than our DDR4-3200 DR CL14 kit which works with everything.

@SKHYJINX

Жыл бұрын

@@Hardwareunboxed Forgive my ignorance, is 3600C14 DR not easy for each platform?

@Hardwareunboxed

Жыл бұрын

@@SKHYJINX probably no, I suspect a lot of CPUs will fail at that frequency, especially older models used for comparison.

It'll be interesting to see what the memory differences are with the coming GFX cards that will likely appreciate as much bandwidth as you can feed them, and then throw in the Ryzen 7k with that and its good times ahead!

@h0stile420

Жыл бұрын

Good times ahead indeed my friend!

Love the little VLDL easter egg you got in there, Steve!

Could you test CybperPunk 2077 with Ray Tracing to see DDR5 vs DDR4 impact? Regarding SimRacing games, since they like low latency, does that mean that for simulation games 5800x3D could still be faster than DDR5 systems? Cheers and thanks for the hard work.

I am targeting the second half of 2023 for my next build and by that point I will go for DDR5 for sure.

Watched your previous video and moved to DDR5 because pricing is getting really close indeed and the ITX motherboard I was thinking about for a while just happens to be DDR5 only. And even though memory prices are expected to fall a bit further still, for the next few months DDR5 may not lose much market value considering the release of new AMD CPU/motherboards which all will require DDR5. Quite a few. That, or I may "lose" some 20EUR at worst. So timing seems as good as any.

There is 0 chance that DDR4 4000 kit is running in gear 1 like the 3200 kit is. That's the only reason that 3200 kit is winning.

@fabiusmaximuscunctator7390

Жыл бұрын

The 4000 kit isn't B-die. The B-die has CL 19-19-19-39. I guess the sub timings are way better as well.

@SamLoki

Жыл бұрын

@@fabiusmaximuscunctator7390 not relevant lol

@fabiusmaximuscunctator7390

Жыл бұрын

What do you mean by not relevant? Why would people then go for Samsung chips if they weren't any better? 3200 CL14 is way better than CL16 garbage

@skorpers

Жыл бұрын

Yeah I have been loosely following this from the drama it's been causing and... Memory tuning is really not complicated and it's dishonest to suggest to viewers that they should buy DDR5 when unobtainable DDR5 speeds barely outperform the 3600 C14.

@fabiusmaximuscunctator7390

Жыл бұрын

@@skorpers Do you have any advice for RAM OC? I haven't tried it as I'm not sure where to start. I have a 4x8 3600 CL 16 19 19 39 kit G.Skill Trident Z RGB

I am planning a new build from the ground up but the price of boards are a lil bit too price for me especially in the ddr5 itx category

I see a great benefit in RAM finally pushing lower resolutions higher because with upscaling tech running internally at closer to 1080p means even better upscaling as it keeps the CPU out of the way.

Can you use Intel's XMP 3.0 designated memory in the new AMD system's that would normally use EXPO memory? I want 6000mhz with 2x32GB sticks but they only seem to be available for XMP 3.0.

How tall is the cooler included with the i3-12100? I'm wondering if I can fit it inside of a node 202; or if I need to buy a Noctua.

Funny non story from me. I was running a benchmark for Cyberpunk and wondered why my numbers were so garbage compared to what I've seen around the place. Went into my BIOS and found my ram was running at 2133 -_-. so, I kicked it into the factory; 3600 at cl 15, turned on resizeable bar and holy shit the difference was great. Anyway, cheers for the video.

Very good video. Can you overclock this sticks? That would be very interesting for me and others.

I'm curious how tunable these DDR5-4800 kits are. Can we up the voltage and get the CAS latency down to 36 or 32? Get a few extra hundred MT/s?

I do not know what to think about it, on the one hand very nice material showing the difference in performance, on the other hand so detached from reality. The vast majority of computers based on the i3-12100 and even the i5-12400 are computers that will never work with a card better than the RTX3060 / RTX3070Ti or their equivalent. Maybe somewhere a small group of user for unknown reasons and in the unknown future they will pair this hardware with something better, but this will definitely be a margin. The question then is whether in an average set based on these processors DDR5 makes sense and significance, or is it just a surcharge to the non-existent difference in performance, when the real limiter is the card.

I hear that so long as you avoid Micron DDR5, and get yourself some Samsung or SK Hynix DDR5 sticks, your chances of hitting 6400 is pretty good.

@andrewschmier6749

Жыл бұрын

@Skyfox nah micron DDR5 is just trash all around, can’t reach high clock speeds or low latency

@cptnsx

Жыл бұрын

No - you need a good motherboard as well.

So are you going to do similar comparisons how different DDR5 speeds affect Zen 4 processors?

Seems like you're going out of your way to make DDR4 look like absolutely sh*t

Would be nice to see a comparison to oced DDR4 3800/CL 16 which alot of cheaper DDR4 kits like the Crucial Ballistix can do.

My dude, great comparison, not thinking about ddr5 yet but it's good to be updated. That said, take care of your voice, a lot of reveals incoming

I put my I3 12100 on a 50% off Z690 Classified board with cheap ddr5 to ready myself for 13th gens mid/high end parts as im into running home servers, ARK S.E and Minecraft servers, as well as virtual machines. As well as to wait for better encoders on 13th gen whether that be the 770 or some 800 series uhd graphics Basically I bought in cheap-ish to reap the benefits in a year

What about stability on ddr5? Is there any major stability issues I need to worry about?..and is OK if I got 2 separate 16gb sticks not x2 kit

Stuff that streams textures seems to like more memory bandwidth. I bet you'll find this for Arma 3 as well.

You guys are killing it with the content! Well done.

I think I've grown out of seeing hardware reviewed in the best case scenario rather than actual real world use. It's annoying having to assume how X hardware would perform at 1440p and/or using midrange GPU. I just feel real world testing should be the norm. The current method puts X product in the best possible light in configurations the majority are least likely to have, and leaves too much to the imagination trying to account for performance lost using different hardware for no good reason in 2022 - imo ofc.

@m8x425

Жыл бұрын

I like seeing both the extreme tests and real world tests. Because of the extreme tests, too many people get confused and then recommend an i5-12600k (or in the past the 10600k) to a novice that's going with an x070 or x060 GPU. The channel called Testing Games is really bad about airing extreme CPU tests while ignoring the more realistic ones.

@theoldpcgamer77

Жыл бұрын

@@drunkhusband6257 4k with dlss performance, same internal res as 1080p but pretty much 4k image quality on a 4k 144hz screen for max fps. Or 1440p 240hz with dlss quality is just under 1080p.

I would like to know what gear is the DDR4 4000 kit? This shows its slower than 3200 c14. Somethings not quite right here. Its good that explain the difference b/w GPU & CPU bound games but most of the tested games does not shine with faster RAM. The only shooter is CSGO & its an old engine. I dont mention Rainbow 6 because its a builth in benchmark which doesnt reflect the multiplayer. Im not sure if CSGO is also using a community made benchmark. They're nice & easy to do, repeatable but not realistic of the actual games.

Great video, but I recall hearing issues when a second kit of memory is installed, even though the sticks themselves are identical. Was this an early adoption issue or is it still relevant and could AMD's new EXPO have this cross-over onto their new 7000 series CPUs?

@ivanbrasla

Жыл бұрын

I figure worst case scenario you could sell your kit, then buy a bigger 2x16 pair for example

I’m still using ddr3

the ddr4-3200 dr have been tested million times . I would like to see the fastest ddr4 in dual configuration so to make this testing even better.

Shuold have included Corsair Vengeance LPX 3200MHz too, as it is one of the popular option

I would really like to see some ddr4 4000mhz cl16 16-16-16-36. I feel like it would be comparable to 4800mhz ddr5 cl32.

U should add cost of contact frame for CPU :) unless u are happy with non even CPU/cooler contact.

For me it makes sense because I'm making a productivity PC with a powerful CPU. I'm also not building until next year, so I have plenty of time to wait shortages out.

Are budget ddr5 like crucial good for productivity as well like compression decompression, blender, handbrake, multiple browser instance with 20+tabs

the Alder Lake + DDR5 option is tempting ngl, no doubt DDR5 is getting more interesting with time, but no idea... with DDR4 im still fine for the things and games i play, may rather upgrade to Ryzen 5000 since my mobo is compatible and is way cheaper upgrade, then wait like 2 years more to jump to a DDR5 platform.

The thing is that ddr4 memory can go as high as 3600mhz cl14, some people managed to get 3800mhz cl14. I think that would be the real test against ddr5.

@elus89

Жыл бұрын

Not in a budget build.

@saputrasaputra3347

Жыл бұрын

yeah and you need expensive b-die + good z690 board

My consern with DDR5 is simply the cost of the motherboards .. everything but the extreme versions of AMD are seriously chopped down

@ColdRunnerGWN

Жыл бұрын

The AMD motherboards are expensive because their MOBO partners have put a lot of PCIe5 slots on the boards, which just isn't necessary. Hopefully, they'll come out with some variants that only have one or two slots of PCIe5, and the remainder as PCIe 3.

@NotThatGuyJD

Жыл бұрын

@@ColdRunnerGWN the partners put them on because AMD mandated it with the chipsets.

@johndoh5182

Жыл бұрын

Well then that's not really a concern with DDR5 now is it? I'm concerned about DDR5 because the sky is falling. See how those aren't related?

@ColdRunnerGWN

Жыл бұрын

@@NotThatGuyJD - I didn't realize it was mandated. I hope they relent because I can't see anyone using PCIe 5 for anything right now.

@NotThatGuyJD

Жыл бұрын

@Cold Runner they won't. Just look at the chipset requirements. Even B650 needs a pcie 5.0 for nvme storage

That GSkill 14-14-14-34 is amazing. I wonder what the DDR5 equivalent will be.

@bobsaget6720

Жыл бұрын

You can get oloy ram with the same timings & voltage for 116

Thanks this is the first time Ive seen a review comparing different Ram timings and different lvl DDR4-5. It shows very well the the CL speed is still important that the Speed of the Sticks. so the need for CL 36 or less is going to be better than CL 40 of which I see a lot of. I would like to see the difference between Dual sticks and quad sticks too

@BottomOfTheDumpsterFire

Жыл бұрын

Of course CL speed is important, both it and frequency make up the actual latency in nanoseconds. Hopefully, the sub-13ns sticks (5600CL36 and faster) become cheaper, and maybe we'll get 10ns like we have with DDR4.

@veechaeuphrates8512

Жыл бұрын

@@BottomOfTheDumpsterFireThere's a 32GB 5600 28 CL G.Skill for $270 on Newegg, which isn't bad and probably better value per dollar than the 32GB 6600 32 CL $370 Corsair. Would love see those two benchmarked.

@kirby0louise

Жыл бұрын

4 sticks is going to seriously cripple memory speeds for DDR5. According to AMD's website, Zen 4 officially supports DDR5-5200 with 2 sticks, but that drops to DDR5-3600 with 4 sticks (so you get DDR4 speeds but DDR5 timings). If you don't absolutely need the extra capacity and/or it can't be achieved with denser RAM, do not use more than 2 sticks of DDR5

@tars3249

Жыл бұрын

It doesn't show that because 4000 cl19 is poor. Try 4000 cl16 or cl14 in dual rank vs the 3200 cl14 kit then we will see how much speed matters or not.

@Tainted-Soul

Жыл бұрын

@@BottomOfTheDumpsterFire to put it into a prospective that I am looking at. I want 64GB (2x32GB) and is looking to get the 6000hz for a AM5 build to find that sweet spot and may be the 7800x3d or 7950x3d early 2023. but as all ive seen for sale in the UK have CL speed of 40 some 32GB sets of CAS32. yes it would be great to get 32 or less but I will have to wait and see over the next 3 to 4 months. also please note that this isnt my first pc lol and in the past Ive always picked the fastest Cl speed in the other RAM even way back when RAMBUS was a thing :)

Nice video, Steve 👍

You are still missing the mainstream RAM configurations of 3200 CL16 or 3600 CL18.

The x8 and x16 in the DDR5 is the density stuff?

I’d love to see the difference in performance.. with a budget / mid tier gpu, and then 1440p Sort of like, a “realistic” difference.

@teemuvesala9575

Жыл бұрын

Let me tell you, there would be barely any difference then. Shows how out of touch people get when they grow their channels to 1M and beyond mark.

@MrLincoln87

Жыл бұрын

@@teemuvesala9575 Well I wouldn't say out of touch ... its like testing the CPUs with a 3090/3090ti with a 360aio etc at 1080p The testing is all done to show us the actual differences But I would like to see some realistic testing .. which as you say, would show sfa difference (or ... would it?)

@teemuvesala9575

Жыл бұрын

@@MrLincoln87 Yeah but the actual differences are much smaller because nobody uses i3 along with 3090. They intentionally created massive CPU bottleneck and then played surprised when the difference between DDR4 and DDR5 was amplified 10 times over.

@MrLincoln87

Жыл бұрын

@@teemuvesala9575 Well I am interested to see the differences, so get rid of the bottlenecks or make it so we can see a difference. But also included testing, or a follow up.. with realistic hardware configs

@MrLincoln87

Жыл бұрын

@@teemuvesala9575 Hahaha imagine soon, i3 with a 4090 at 1080p testing ...

this plus Zen 4 3D cache means cheap ram with maximum performance. excellent test!

So considering that all tests are in 1080p, does 1440p or even 4k have any tangible differences? I am guessing not really? Would still be interesting to get an aswer on that.

Unfortunately when I built my last PC in 2020 I bought a $400 x570 Aorus Master. Maybe a bit of sunk cost fallacy going on, but I'm hoping to hit the price floor on high end of previous gen CPUs and GPUs.

@andersjjensen

Жыл бұрын

If you're mainly gaming, just keep a price watch on the 5800X3D and snap it up just before they become expensive again because of rarity. The second hand market is already starting to flood with high-end Ampere which, even if you prefer buying new, is going to have a rather positive influence on the whole GPU situation.

Still confused about 1080p Medium in ACC, 3090ti and I3-12100 fps similar to the 5700xt/3800xt/DCDR3800C14 I run. Heavily CPU limited I guess but still weird for 1080p med - lots of racers running triple 1080 or 1440p on lesser cards, high settings etc with similar numbers. Maybe one of the few games where core count makes a small difference!

@iPain3G

Жыл бұрын

ACC is extremly CPU heavy. Especially on a big grid. But corecount doesn't really matter. Most of the CPU heavy tasks are only on 1-2 threads. ACC needs high singlecore performance and tons of cache. I have an 5600x, 32gb 3800CL16 and an 3080 and can barely hit 90fps with a full grid of 50 cars. When i'm alone on the track i have over 180fps. I will upgrade to an 5800X3D(when the early adopters buy the new Zen4 CPUs) to get better rid off the CPU bottleneck.

@ryanviertel7059

Жыл бұрын

Racing sims are pretty cpu intensive because of all the physics calculations

@eugenijusdolgovas9278

Жыл бұрын

That "small" difference being a stutter on the last lap of the race with 4C8T CPU. It's so "small", that You want to run Your fist through the monitor. Core count always matters, especially if You're using more than a single monitor at 1440p.

@iPain3G

Жыл бұрын

@@eugenijusdolgovas9278 for the cpu it is completly irrelevant how many monitor you have. It only matter when your gpu can't push the pixel. The stutter on 4c8t CPU is more a cache issue than a corecount issue. For all the physics related stuff ACC only use 1-2 cores. All of the 4c8t CPUs from Intel and AMD has not that much cache. You can easily test it when you buy a CPU with big L3 cache and disable cores. On my 5600x i get the same performance with 4c8t.

could you do a video where you compare high-end 2x8 GB of ddr5 vs 2x16 of not so high end ones?

I aggressively overclocked my 32gb ddr4 to 3600mhz 14 14 14 34 (including secondary, tertiary and RTOL/IOLs) following quite a few buildzoid guides. Really intrigued to see where it would stand in this benchmark when paired with a 12600k. Its a DDR4 4400 CL19 kit.

@AdaaDK

Жыл бұрын

same here, i run a 12600k and i got a DDR4 kitt Gskill 3200C14 same as the one HUB tests with, but in stead of running it at 3200C14 @1,35v i run mine at 3733 CL15-15-15 (30+ something), at 1,45v ,and its rock stable. + only reason i cant go above the 3733MTS mark, seems to be cos of the aldorlake mem controller not liking it at default voltages in gear 1:1. (i may test with more volts in the future). And testing some tuned B dies vs the rest of the lineup with be awsome for me as well.

@BNR_248

Жыл бұрын

@@AdaaDK Thats a really good OC, huge jump while still maintaining tight timings. I agree, testing aggressively oced B-Dies would be really cool to see. B-Dies are literally known for their OCing potential and its growing in popularity. Its a missed opportunity to not review OCed memory

@tars3249

Жыл бұрын

I have the same kit i'm guessing (patriot viper steel right?) doing roughly same timings (14 14 14 28) on 11th gen. That kit is really good when it was on sale a year ago because it can scale up or down in frequency with tight timings. Since yours is 32gb I'm guessing it's dual rank? is it 2 sticks or 4 sticks of ram? Mine was 2x8gb so single rank unfortunately.

@BNR_248

Жыл бұрын

@@tars3249 Yes spot on! Mine is also single rank, i bought 2 16gb kits (so 4 8gb sticks) as they were on sale for a ridiculosity cheap price. Im on 10th gen so the IMC isnt as good as yours, but i got some really aggressive OC and found 3600mhz to be the sweet spot with the most gains in both bandwidth and latency. I used the Intel MLC tester, Aida and geekbench. Gained about 30% more performance over the factory 4400mhz CL19 timings. Anything B-die is great. The secondary/tertiary timings is where i invested most my time Ocing. People dont realise how much motherboards loosen those timings.

@tars3249

Жыл бұрын

@@BNR_248 Bro 10th gen has a better IMC, I could do 4166 cl16 on a 10700k and msi z490 A Pro. You should try for higher frequency maybe? My 11700k was stuck on 3600mhz if i wanted gear 1 mode so i had no choice but to try for super tight timings instead. But 4166 cl16 is definitely better bandwidth and latency. I wish i had kept my timings for that to give to you for testing, maybe if i find them ill post them here.

What about DDR5 Sodimms for laptop? Are they better priced?

I just upgraded from z390...I have enough ddr4 to stay with it for an upgrade for now... I got a new z690 board and cpu and kept my other everything else. However if you're a new PC builder getting everything new...then you might as well opt for the newer ddr5.

I'm still on i7 4790k, Asus Maximus VII Hero, 1080 Ti and ddr3 1600Mhz. Finally but sadly its starting to show its age. Now upgrade to either AMD or INTEL except I need a CPU that has a single core High Speed because early Bethesda Games such as Morrowind, Oblivion (especially), Fallout 3 and New Vegas were optimized for single core or even dual core CPU with a high Speed. I also play Skyrim and Fallout 4. I still have problems with the occasional CTD in Fallout 4 and Skyrim due to out of memory (these don't give me any logs). Any advice on what CPU, M/B, RAM etc to get here in Australia?

Hey sir! I’m a huge 4X gamer so if you could mention how ram/cpus work with games like civilization/Stellaris/Crusader kings it would be most appreciated

@Psi-Storm

Жыл бұрын

They need high single core performance and don't scale that great with massive cores. The coming Ryzen 7600x or Intel 13600K with ddr5 ram would be your best choice. If you want to spend less, then the 12400(f) is the best value chip at the moment.

@theplayerofus319

Жыл бұрын

The 5800x3d is great for such games because of the high l3 chache

Good work !!!

My next build is a 13900K with EVGA Classified Motherboard, EVGA 1200W PSU, EVGA 360 mm AIO and definitely DDR5. I will continue to use my 3090 FTW3 till I need a better GPU.

@BNR_248

Жыл бұрын

Wait for benchmarks on both AM5 and raptor lake first. Price points look very similar so far. Don't constrain yourself just yet

@theoldpcgamer77

Жыл бұрын

@@BNR_248 Yeah but AMD take forever to get things stable. Intel are usually good to go much quicker. So it can be less hassle.

Gud thing i bought a pair of 16gb 3200mhz for less than $90 lastmonth A guddeal compared to 32gb ddr5 4800 which sells for $160-170 in my country

Anyone know if Raptor Lake will be capable of speeds more than 5600MHz?

I don't know if it's worth the work but I'd love to see a compared ddr4 vs ddr5 for productivity and not just games, to see if the upgrade it's worth it there.

@haukionkannel

Жыл бұрын

DDR5 is more important in productivity. Just looks Intel 12000 series reviews from this channel!

DDR5 already ? cooooool and here I am watching this video with my ddr3 1600 mhz ram

Would love to see a review for the new Samsung Odyssey G4 1080p 240hz monitor!

Holly shit, that’s I was looking for

I loved seeing Viva La Dirt League's Greg, Baelin, and Baradun figures there!

Adata 16gb DDR5 4800 CL40 sticks on sale at PLE for $69 each AUD - makes it a no brainer at that price

A little late for that, already got DDR4 ECC for something huge XD I wish DDR5 was in my immediate future...