DDR5-6000 CL24 with Ryzen 7000!
HTML-код
- Опубликовано: 28 авг 2024
- My Patreon: / buildzoid
Teespring: teespring.com/...
Bandcamp: machineforscre...
Twitch: / buildzoid
Twitter: / buildzoid1
Facebook: / actuallyhardcoreovercl...
#overclocking #ryzen #ddr5 #amd #aorus
This right here is #1 DDR5 content online. And it's free... Buildzoids a fuckin legend!
Actually Hardcore Overclocking is the most underrated tech channel on YT. He knows what he's talking about and talks about what he doesn't know, no fluff, no exaggerations, just facts. Learning what he's saying will give you more knowledge than Gamers Nexus or LTT or any other mainstream channel.
Here we find that extra bunch of extremely rich content majority don't seek
ONLY SECOND TO FRAMECHASER OF COURSE, RIGHT? RIGHT???
GN and LTT are very good channels. One is not necessary better than other, just different audience/target
Imagine lumping Gamers Nexus with LTT
Idk I started with ltt
Moved on to gamers Nexus and hardware unboxed
I ended up here and debaur
If anyone has any other overclocking channels like this recommend them
Nice to see you back my man!
Have learned so much from you :)
Nice to see reasonable descriptions for just about every option I've seen scrolled through here. That's rarer than it really ought to be, from what I've seen here and there.
Very nice O.C congrats. 6000/24 is 250 very high, i have 7750/32 is 242. The 6400/26 is 246, but as we know is faster than 6000 even with cl24. There was a guy and he had scored with 7800x3d 14800points in timespy cpu having 6400cl26, so it is super fast. But higher memory frequences allow for better score over 15k.
@@ur_very_angrysame issue here homie. Trying right now to get 6400
My Time Spy CPU is 14888
R7 7800X3D - 105 BCLK - LLC:4
2X 16GB 6200MT/s 28-36-32-30 (tight subs) 2200FCLK.
VDD: 1.45v VDDQ/IO: 1.30v VSOC: 1.24v
Aida64 latency: 57.7ns
PYPrime: 7.534s
@@erickelly4107 Powerdown mode off I'm guessing? In safe mode or normal Windows?
@@erickelly4107o
Awesome video, thank you for explaining in such great detail! I was having so many problems with getting my memory kit running well, with this, I'm sorted 👍 Cheers!!
Insane OC
Would like to see this in game vs 6000 CL30
My 2X16GB Dual Rank (Thaiphoon Burner / OCCT / etc states 2 ranks) Adie kit can run 6000 cl26@ 1.50v VDD
6200 cl26@ 1.55v VDD
Both stable via 30K Karhu cache enabled, 4 hours VT3, 2Hours OCCT mem, 2000% HCI, etc.
2X16GB 6000MT/s CL30 38 38 96 - 1.35v EXPO from GSKILL
Haven’t tried CL24 but I imagine 1.7v VDD would work.
Yes there are indeed 2X16GB "Dual Rank" Hynix Adie Dimms, they use the dual rank timings (no post below 4 / 6 at 6000) just like the 2X32GB Dimms. These don't appear to be very common but they do infact exist, I have two sets from GSKILL
Currently: 6000 CL 26 34 32 36 @ 1.50v VDD 1.30v VDDQ/IO 1.195v VSOC
Aida64 latency: 57.3ns
PYPrime: 7.567s.
R7 7800X3D
Have a great Xmas BZ
Crazy OC ! Im a bit more cautious with the voltages . (2x60mm Noctua fans on the sticks anway ) . Im running 6400 CL28 -37-37-37-30 /410trfc with 1,5VDD / 1,42 VDDQ perfectly stable.
Did you try lower tRP?
@@JJFX- Yea , 36 is not fully stable.
@@csabauri351 Damn, bad luck. It's interesting how much tRP can vary on A-die. My current kit easily runs 30 at 6400. Can you not get closer to 120ns tRFC either?
@@JJFX- Nah .128 ns is the best i can do with this voltages.
@@csabauri351 Mind if I ask what kit this is?
G'day Buildzoid,
Merry Christmas from Australia 🎄🎁🎁🎁🎁🎁🦘🦘🦘🦘🦘🦘🦘🦘🛷🤶🍹🍺🎅🏖
From my findings
Cl26 6000 requires 1.6 anything lower it wont boot/error
Cl26 6400 requires 1.650 some kits might need more depending on the kit bin aswell
What board r u using?
Yeah that's about right, give or take with some additional tweaking. I've seen 6200 C24 for benching over 2V. Hynix takes voltage like a champ but I wouldn't run it for very long... it'd be an interesting to test long term though...
Yeah I needed 1.67v for 6400c26 with my X670E hero, this was with mdie
@@Jas7520 is asus boards problems with am5 r over
Or u still have any
And after using the hero do u recomend it over the orus master
@@Adel-sj4ue x670e hero asus
8000 a die 6400 cl26 flck 2133 running 1.65 on the dimms been dailying no issues regrading windows etc but yea soild so far can boot 6600 cl28 but yea not worth it
Hello, new to ram overclocking here and been learning a lot from this channel. I just recently built a 7800x3d on a used Strix B650 E-F board paired with some Lexar Ares 2x16 ddr5 6000 cl30. I knew from other forums to get 6000 cl30 as the sweet spot when I was searching for parts. I must've been real lucky because I only picked it out over the looks like the idiot I am. Since then I got curious and am now learning about ram and how to tune it, I've managed to get 24-34-28-30 stable at 1.75 volts. Or atleast pass Y cruncher a few times and not crash after a few hours of gaming so far with no obvious signs of instability. It may go further but I'll stop there for now. The stock xmp timings were 30-36-36-68 I believe for reference. Idk what 'die' they are or anything but I was wondering if you'd like to make a video on them as they seem to be doing really well or I just got lucky. I got the pair from amazon.
You're severely limiting your sticks lifetime with this kind of voltage! I wouldn't say it was worth it, personally
hi BZ i know you have sponsors like asus and gigabyte and stuff but you should also try Asrock on AM5, its really nice for bclk OC and extreme OC like DDR5-8600+ with tight timings
Running the stock cooler on the 7900x , totally doable, totally reasonable, still hilarious to me 😂
Buildzoid went to the future for these timings 16/09/2024
Damn, is this like akin to 3000 CL12 on DDR4?
wenn du so willst
The most og of all og, tech channels 🎉
I have literally taken the sub timings (even some are lower) but running it on CL26 as my Gigabyte X670 Aorus Elite AX rev 1.2 couldn't apply more MEM VDD than 1.65v. HOW EVER I'm running them currently on 1.48v or was it 1.46v? I'm writing this from work so I got to go back home and verify this later to not mislead anyone here.
My timings =)
6,000 MT/s CL 26-35-35-28
tRFC1 360 (BZ is at 500)
tRFC2 360
tRFC 120 ns (latency)
tREF 65,535 (SAME AS BZ)
GDM (Gear Down Mode) ON (SAME AS BZ)
Power Down ON (SAME AS BZ)
Memory Context Restore OFF (BZ ?)
FCLK 2,133 MHz (SAME AS BZ)
Hynix A-DIE
TeamGroup T-Force RGB 32GB 2X16GB DDR5 (6,000 MT/s CL30-36-36-76 1.35v EXPO 1 STOCK RATED)
6,000 MT/s CL 26-35-35-28
MEM VDD 1.48 - or 1.46 (need to verify)
MEM VDDQ 1.36v or 1.38v (0.10v lower than VDD)
My VDDP RAM Voltage is the same as BZ 0.9975v
CPU 7800X3D BCLK stock
Vsoc 1.15v (Same as BZ)
LLC CPU VCORE loadline is maxed out (EXTREME)
LLC Vsoc loadline AUTO
PBO CO (Curve Optimizer)
NEGATIVE -20 ALL CORES. (Some cores can do better, but this is stable so far)
Motherboard
Gigabyte X670 Aorus Elite AX rev 1.2
GPU 7900XTX ASUS TUF OC EDITION 24GB VRAM
I've been dailying 1.67 vdd for nearly a year now (8000C32). Also, I know people who have been dailying 1.7 vdd (at around 40-50c) for a year too and neither one has shown any signs of degradation yet. (hynix a-die)
I'm not 100% sure but I think I recall something about Hynix documentations stating vdd and vddq should stay about 300mV away? Do you have any idea why so
Quick question: Having a 6 layers motherboard vs 8 will impact the timings and subtimings ?
I just bought a gpu that i was told you might be interested in. Its a gigabyte GTX 680 SOC. if you'd be interested in it then let me know.
this is insane. someone's gonna brick their memory watching this XD
brick??? Been dayling 6600 CL26 at 1.9v basically since zen 4 launched.
I really hope Gigabyte's availability gets better soon, the Masters and the Tachyon in particular are just nowhere at the moment
masters had issues (even after they fixed the vrm failures) and the tachyon discontinued
@@sparklingionOh really? Where did you see/hear that the Masters were having issues?
would love some help with my setup :( I've 2x32Gv G.Skill Trident z Neo CL30 / EXPO works for a few hours than Ramdon Bluescreen, 7950x stock settings on an ASUS b650e-e running the latest BETA BIOS 2204, I'm running the BETA BIOS as my last RAM kit Corsair Vengence 2x32gb would not work as it kept bsod even on stock settings :( I think this MOBO is total crap as I cant use any manual Curve optimatzation instant crash at login screen.
What kind of voltages would be needed to get Hynix to 8000 gear 2 with Ryzen 7000 and cl30-36?
Also, would secondary and tertiary from 6000 work at 8000?
Getting 8000 to work at all is mostly dependent on getting lucky with the silicon lottery and having the right motherboard.
@@ActuallyHardcoreOverclocking Yeah, I think I lost the lottery then, my TUF x670e and 7700x seemed like it trained the memory but then the CPU light kept illuminated after passing the RAM light and it never posted not even loosing cas to 42 with 1.43 vdd
I was hoping he would explain Nitro mode like he did Gear Down Mode.
And here I am struggling to get DDR5-6000 CL30 stabilized
I have a weird question how well do u think I could run jdec 2666 at below 1.2v? I happen to have a lot of ddr4 in a server and it is contributing to like half the system power draw.
I'd guess 1.1 would work. 1V IDK. but if you don't mind the system down time you could just test it yourself.
I dont think 18 passes of vst is any stable. I get errors on 40-50 pass or sometimes at 80-90. And if it happens then my game or system could crash too midgame
Heres to another year winging it in '24
Hi Buildzoid. I know that what i'm saying is off topic, but have you ever tried to OC ddr4 with ryzen 5600g. Recently bought one and had a lot of fun. 4000cl18 ripjaws V oc'd to 4466cl19 with 1:1 ratio. Should be able to oc to 4600 on a good chip with 1:1 ratio, but my ripjaws tops out at 4466.
Im going to be overclocking with my 7995WX qnd im decoding between the ASRock and ASUS WRX90 boards. Which of the two brands have you had better luck with for DDR5?
6200 30-37-32-30 2200 fclk works for me on 1.4V
And it is faster than 6000 28-35-30-30 2200 fclk.
So you are right. 200mhz outperforms 2 ticks on all primary timings.
6400 a no-go? My daily config is almost identical at 6400 30-37-30 with GDM and PDM disabled. That's with 1.43V, 1.38V VDDQ/IO but my chip can only scale to 2167 FCLK.
At 6200 you can likely run CL28 at ~1.43V with A-die. 6200 28-36-29 is the best I could do with full stability.
@@JJFX- nope... 1.43V is not enough to make it stable. None of what you wrote
@@proklet4694 You might need to raise tRP on your sticks at 6400 but it's possible your CPU can't go that high. The newer AGESA makes it fairly easy to run. Honestly 6200 C28 is probably the better setup if you can get that working. Congrats on 2200 fclk though. Just make sure performance actually scales that high because in many cases high fclk is actually slower but seemingly stable.
@@JJFX- I think it is mostly due to the fact that I did not try SoC voltage over 1.275. I probably need 1.3, but I did not care to try
@@proklet4694 Possibly, but in that case simply lowering SoC to test may be enough. My chips run really low SoC so I've only had to push it for 6600. You shouldn't need to raise SoC for 6200 C28 though.
Hi, I have 6000 32 38 38 96 trident z5 neo rgb. Its 1.35V by default. I dont want to add voltage. What is minimum CL?
X670e e vs x670e orus master
Which is better for ram oc?
Gigabyte has been the the most consistent for me RAM OC wise. Though the BIOS is kinda clunky and the ECLK implementation is really awkward to use compared to how ASUS/ASrock implement it.
@@ActuallyHardcoreOverclocking thanks man🤗🤗
2x SR vs DR fully tuned Hynix A die for AM5 if cooling is not a problem?
2 sticks of single rank? Would you not want dual rank?
Hi can you please suggest me a video from your oc to ddr5 G.Skill Flare X5 DDR5-6000 BK C30 DC - 32GB (AMD EXPO)
@actuallyhardcoreoverclocking I find 6000mhz (my ram is rated for that) a challenge. It's supposed to be for intel but it works if you set the voltages. Also I've been getting windows corruptions.
I tried the newer bioses for my board (tuf gaming x670e plus) but the 1653 bios seems to be the most stable so far. I had to bios flashback it to the old bios because they really don't want people using the old bios files.
The symptom of windows corruption is the system will boot up fine but the nvme will just stop working after about 1min of the system being on. The windows bootup fails to fully load. I'm not overclocking and it still fails. Reinstalled bios same issue. Reinstalled windows and fixed it. I think it might have been the newer 5.12.x.x amd device drivers that might have helped but I'm not sure. Also ryzen master is junk. It says "your -30 overclock is all good" and then also freezes the system when it boots up... useless. OCCT has been a good way to test for stability so far (the power test is awesome for example).
Since it is a Asus board I would verify that the VSOC is 1.2V and not 1.3V-1.4V like some auto settings on older bios.
How much RAM do you have and what config? Also my exp with ASUS AM5 boards has been that they have the most of the weird stability problems. Though I've also seen AM5 boards from all brand some times just miss-train and then get stuck with bad memory training until you clear_CMOS or force a re-train
@@ActuallyHardcoreOverclocking I have 32gb of g.Skill trident ZS F5-6000U4040E16GX2-TZ5S. 2 sticks. As far as I can tell it's Samsung.
It's not supposed to be "amd" ddr5 but you know... It works and it works well and can be stable for weeks. OCCT has been run for 8hrs and zero issues before. All tests stable.
Also there's the mounting of am4 coolers on am5 sockets being a bit iffy as well. If I run OCCT (or cpuz stress test) and look at the temps and move the heat sink around and try different pressures on the heat sink then I also notice there's a band of higher steady clocks. I.e I screw the heat sink down a little more and it improves and then it will get worse and I have to back it off. It can be as much as 100mhz or more. I.e it's a noticeable difference.
I have some cooler master 240 AIO. If I have the cash in the future I'm planning to get a 420 or some such. I have a lian li D8000 so cooling is not an issue.
The weird memory training sometimes happens (50-50-50 etc) if you don't use expo or don't manually set the timings. Also it's imperitive to set ram voltage to 1.3.
Before the 1654 bios I had to set Soc to 1.28 (it settles around 1.3 to a smidge over) and vddio/mc to 1.28. Now I just set VDDIO/MC to 1.28 and soc is a bit lower on auto.
I get about 4.7 to 4.85 on linpack and 5.1 to 5.3 on other all core tests.
wait i can go down to 24 timings at 6000. now i need to see about the timing calculator
BZ, i have a watercooled rig, that includes the RAM too, (yes watercooled ram) but I run a 7800x3d. In your opinion, would it be reasonable to try to push my ram higher? I'm running 6000-32-38-38-96 (basically stock)
for gaming? there is pratically zero incentive on overclocking ram for X3D chips, they use mostly the speed of the cache, hence why its so good at gaming, for most gaming related memory activities, so the difference between even a 5200 kit and your 6000 is very very close to zero. going below C32 on 6000 is COOL to do, and I really incentive you to do it if you're interested in getting into overclocking as a hobby, but you won't gain anything from it performance wise
@@TheNubaHS thanks!
I have a question: There is something weird going on...I have an *Asus x670e-f* with a *Ryzen 7700x* and *4x16gb 6000 CL30 Hynix Ram.* I can only run it at 5600 Cl36 because I can't set the VDD higher than 1.18v and VDDQ higher than 1.11v. When I apply the EXPO profile or manually setting the voltages to 1.35v it won't even post. It posts with 4800MTs (or lower) and 1.35v but it's still unstable in y-cruncher. With only 2 sticks it runs with 1.35v at 6400 CL28 stable without a problem. It seems like I've tried almost everything but nothing works...
return your 4x16, get 2x32. 4x16 is silly.
@@tommihommi1 Ik, it's for the aesthetics only. Idrc about performance in this system but I still wanna know why this happens...
@@CrazyPinkWomen just get a set of 2 sticks and 2 spacers
Whats the best value mobo for 7900x or 7800x3d?
BZ, you say it's impossible to get error reporting on AM5, but I've found some info on Reddit about feature called Advanced Error Reporting (AMD CBS / NBIO Common Options / Advanced Error Reporting) which needs to be set to 'Supported'. The issue with it is it's up to mobo vendor to add this setting. For example, it exists on Aorus B650 Elite but doesn't on MSI B650 Carbon. I'd like this info to be more available (and I'd like to see MSI add it to their BIOSes)
official memory support is 5200 for 2 sticks and 3600 with 4. 5600 is perfect for 4 sticks
@@hristobotev9726 and my friend's got 2 sticks 5600 MHz. What I'm saying is there's an option to enable error reporting (which is not well known) but at least MSI has decided not to implement it
@@user-to2uc8jn4y I tried that in the past and it really didn't seem to do antyhing for catching IF errors.
Do asus use the same name for voltages?
Any chance we could get a mobo pcb breakdown for x670e hero?
Hello, you can talk about 4 sticks ddr5 being in January 2024, if there have been improvements with the bios or the IMC, maybe you can make a video how to stabilize 4sticks ddr5 sticks on amd please. Thank you
Zoid off script, is best Zoid
it's not like I ever have a script
It's a shame you don't have Teamgroup sticks because as I responded about in the previous vid, you can use the voltage bug (at least on Asrock) to bench voltages above 2V. I've seen 6200 C24 using this when it otherwise wouldn't be possible.
It did work on GB boards but you seemed to have trouble posting high voltage when it was bugged. I can only assume that's the PMIC. It's hard to imagine none of your kits use the same one as Teamgroup but who knows. If you try it, don't set much more than 1.44V to start.
Hey i got a question does anyone know if this Gigabyte Aorus RGB GP-ARS16G37 its 3733mhz cl18 kit could be or is bdie?
which ram test is that?
PC date 16-Sept-2024, is this video from future ??
anyone else can confirm by turning off nitro mode on gigabyte boards makes them run the timings more stable and better?
did not test yet but any timing i try that is tighter than stock expo profile at cl30 38 38 38 96 just flops no matter the voltage
29:05 I found that if your FCLK doesn’t work at 850, it probably won’t work at 950 or even 1050.
850 is spec, i run 2200 at 850mv and more voltage breaks it. Going over 1.1vSOC also makes IF stability worse.
Which voltage it is ?
@@maveric19871 VDDG
@@PowellCat745 ok i am running 2108mhz on Fclk with 1070. How much do i need for 2200mhz? Also i remeber it van run 2000 with 950, but it is strange because flck like 2100 which is max stable in my rig need 1.3v actually 1.291 thats how much mb gives. So like for 68gb read and 99,9gb write soc need to be at 1.3v. Do you think running vddg at 1.050 with flck 2000 can cause instability? Overvoltage isnt good i guess
@@maveric19871 More voltage can definitely hurt stability. I need to lower voltages a lot to get fclk stable at 2200. SOC 1.1 does much better than 1.2
My super-manly Ripjaws S5 (96GB 32-39-39 kit) run stable at a whopping 3,583.3 °C according HWiNFO 🥵🥶🥵
That said, my Kingston kits (in the 128GB setup) also have a quirky reporting error that shows them occasionally jumping from the mid-50s to about 80 °C for a second or two.
Damn this is DDR4 level latency.
Yeah worse case scenario if your RAM craps out is your out ~ $100 and a day or two unless have an extra kit on hand.
Do something impressive. Talk to me when you have 2200mhz FCLK and 6600mhz DDR5 CL28 stabilized. Get on my level.
He did that
Whats that ram test program? I cant find it
karhu ram test
It's not free btw...
Edit: Stable version is up now.
It's funny how Gigabyte just took away the latest F20a BIOS which you mentioned that fixed the high voltage mode.
Because they released a stable version F20, but it isn’t shown on the website right now.
@@PowellCat745Ah ok, it’s not really ”released” then though.
@@PowellCat745 Perhaps, though no site that aggregates BIOSes managed to get stable F20. They still only have F20a. And on Gigabyte Website, only F11 is available.
@@Simon_Denmark Now F21a is released
What about 128gb ddr5 at 6000mhz
👍👍👍
At what max voltage is safe for a DDR5 kit? Mine are DDR5 6200 CL32-39-39 from Gskill (Trident Z RGB) (Hynix M-Die)
There's really no such thing but I'd have little concern running up to 1.55V or even 1.6V on Hynix if it makes sense. However you absolutely need a fan on the dimms which I'd recommend even over 1.4V for long term stability with tight timings. Also be mindful of VDDIO, which I'd run closer to 1.4V with VDDQ. I try the keep the difference between VDD less than 500mV but still as low as possible.
there is no 'safe voltage'. Hynix documentations only say 1.4v so it's mostly users experience going past that. But generally if you can cool higher values it seems to be fine, people have been running high voltages daily. But again, up to you
thank you both, for the answers, I will try 1.6v with a Fan over the modules
Please share AIDA ram benchmark results next time you would create a similar video
AIDA is a bad benchmark.
Why using out of sync FCLK 2133? with 6000 Speed?
It is always out of sync on ryzen 7000 ;)
@@derdaniel8261 I think he meant the half-sync FCLK of 2000? I could be wrong.
@PowellCat745 No, there is no way you can run in sync when optimising performance. Its out of the question on zen4.
I assume this won't be a for everyone copy paste and gaming.
Im using X670e hero and ryzen 7900x and corsair 6000mhz cl30 36 36 76 but managed to get 6000mhz cl28 36 36 58 and the FCLK is at 3000mhz too
What? FCLK 3000? Or you mean UCLK=MCLK=3000?
FCLK @@HoangDucNuyen
I had got a 7800x3d with strix x670e and 64gb 6000 cl36(trident neo a die). couldn't get it stable at 6000 cl40 1.4v. returned it all and put my 5800x3d back in.
user error.
some cpus can barely do 3ghz on the memory controller on single rank so i guess you had one of those
I am running 7800x3d on asrock B650e, Gskill Ripjaw 2X32GB CL30 at 6200 with 1.35V mem and less then 1.2 VSOC.
Tested for 24 hours of cinebench and multiple ycruncher runs and 4 months of daily use and gaming.
@@dwight2310 no it's just garbage. Do a 5 second Google search and you'll see loads of people having stability issues, especially with 4 sticks.
Defaults with expo wouldn't run.
Wow
Is that possible on strix e e
Or aorus master
Or its just aorus xtreme??
If you manage to get your hands on F20a BIOS and it supports higher voltages, then it should. Unless it's one of those 6 layer ones. I'm not sure how signal integrity is affected by CAS latency, but that's still (I think) less distance and shielding from quite high current. For now GIGABYTE got rid of F20a BIOS - officially to release F20. Which is not there.
@@jannegrey593 which would u pick between x670e aorus master and
strix x670e e?
BZ 's going to cause waterblocking RAM to be normalized
i also run cl24 but at 5200
Is it possible to run R9 7900x on asus x670e gaming with 4x16gb Kingston Hynix anywhere from 5600-6000 mhz
Is there any point to this outside just winning a benchmark?
Getting latencies as low as possible is very helpful on Ryzen. By that I don't mean a big difference to the average gamer but I don't know why people coming from that perspective would watch this anyway.
No
@@JJFX- Ya, I am gaming on my new 7800x3d build. Set amd expo 6000 cl30. I know ryzen 1 scaled really well with memory. ty
@@jamesbyrd3740 Yeah the difference in gaming won't be amazing but if you're into it I'd least look into tighter timings as BZ has shown in more practical OC videos. Expo settings are often pretty terrible and it doesn't take much to dial in less aggressive settings.
Simply maxing out tREFI, lowering tRFC and raising your FCLK frequency will help latency. At just 1.42V or less you can almost certainly do 6000 28-36-36 with common secondaries, regardless if it's M or A-die. I would recommend some air flow on the memory for long term stability but it may not be necessary depending on your setup.
Once you go 4 sticks, you're back to 4800mghz anyway, so it doesn't matter.
The real goal is how fast can you get 4 sticks.
I've been trying to tweak settings up from 4800 to 5200 for the past couple of months on an AORUS Elite AX with no stable success.
I've finally given up on overclocking and just want a system that is stable that won't crash, because I lose too much current work on my IMDisk RAMdrive when Wwin10 suddenly out of the blue bluescreens.
When you lose tons of hours in work, tweaking settings, trying to squeeze out a few megahertz more here, going through all that retraining, rebooting... it's just not worth it at all anymore to make your system a tiny bit faster by an infintessimally amount. From my experience over the last 6 months, just be happy to get your Ryzen system stable and running cool at 80C or below (easy to do, set an 80C limit or turn on ECO mode in BIOS) and call it good.
Omg
Давай на русском уже. Ничего непонятно на этом недоразумении.
First!
Can you back to intel please, we implore you
No, not really. Plenty of people want the AMD content. You can wait.
Pretty sure that Buildzoid was quite fed up with 13th gen memory overclocking and I'm not sure if 14th gen (13th gen) is that much different.
@@Simon_Denmark14th gen is a better binned 13th gen (including the 13900KS) but the IMC has no improvement.
Intel is to unstable, overclocks need many hours of validation with multiple test programs.
Intel can pass for 6 hours then get 5 faults in a row and then pass for 6 hours again.
AMD is much faster and easier to test for stability, if it manages 1-2 hours there are very high chances that it can run forever with no faults.
@@larsjrgensen5975 Also it can pass for almost a week and choke up on last hours of benchmarking. But what you said is usual. Couple hours great, then couple errors and then acts as if it was stable again. Until it isn't for a (relatively) brief moment.
How important is the pcb layer count for memory overlocking? Can’t find anything reliable on the net
WHY IS your TRFC at 500 and not to 480?
He explained in the video
He explained it that performance uplift would be negligible. You're talking about reducing time memory is "unavailable", which is 0.763% of the time and improving that figure by 4% to 0,732% of time "unavailable". That is 0.03% of maximum theoretical boost in performance. Given that time gained is less than CAS latency, I would say that realistically it would be much lower boost in performance, basically near 0. Also despite the length of the video, this was supposed to be (and was, mostly) short video on easily getting DDR5 6000 CL 24. It was Extreme OC event, with scores that high, that theoretical improvement of 0.03% could change the outcome of a race.