Undervolting Ryzen 7000 Guide (How To Max Out Your Undervolt)

Undervolting Ryzen 7000 Guide (How To Max Out Your Undervolt)

Beefy Tech

1 год назад

49,186 Просмотров

Ссылки и html тэги не поддерживаются


Комментарии:

@OnitsTV
@OnitsTV - 02.02.2024 17:44

How to check the best core??

Ответить
@rafalprzyb5988
@rafalprzyb5988 - 31.01.2024 13:42

I own a 7600x. Before UV the CPU temperature after 10 minutes of OCCT test =76 degrees centigrade. I performed everything according to this video. I reduced the voltages core 1 - 25, core 2 -15, core 3 -20 and core 4,5,6 - 30. after this procedure, the CPU temperature in the OCCT test rose to a borderline 89 degrees within 15 seconds. WHAT is wrong? I understand that after UV the processor can be unstable, but rather should not heat up more

Ответить
@noname-qx8en
@noname-qx8en - 15.01.2024 14:44

will a 7900x at 4.1ghz, work stable at 1.0v core voltage? with a 6000mhz ram cl30. (cpu mining randomX-windows and linux)

Ответить
@svan71
@svan71 - 01.01.2024 21:43

Bios under volting or nothing

Ответить
@artemskovronskyy9105
@artemskovronskyy9105 - 22.12.2023 14:44

AMD master says my Ryzen 7945HX is not supported, are there any alternatives? Thanks

Ответить
@Snafu2346
@Snafu2346 - 21.12.2023 10:40

Is there any benefit of doing something like this through Ryzenmaster? Or through bios? Or no difference?

Ответить
@kreamypie
@kreamypie - 04.12.2023 03:51

Nice video explaining how it is on idle is when the instability happens. 😃

Ответить
@rjukan79
@rjukan79 - 29.11.2023 20:25

1. Theory about best cores. I have -29 to -31 on all cores on ccd1, runs perfect, gain from 5150 to 5350 on full load.
2. Theory about no performance gains on ccd2. Lower temp-higher clocks. I have -29 all cores on ccd2 and gain from 4950 to 5250 on full load. CPU 7950x, cinebench r23 stock 37500 and after CO record is 39966, ambient temp 17C

Ответить
@Serexityy_
@Serexityy_ - 27.11.2023 16:42

Just did per core today, -15 on fastest and -25 on the rest gave me 484 , 309 , 250

-15 fastest and -30 on rest gave me 472 , 318 , 251 which one would u keep as the daily overclock?

Ответить
@whiteou7
@whiteou7 - 25.11.2023 12:00

Unstable no matter what I put it at -15 -20 it sucks

Ответить
@leemarco1136
@leemarco1136 - 06.11.2023 16:22

When i use boost overdrive, i can't open my comp[uter. Why?

Ответить
@oldschool_bamagatrader9960
@oldschool_bamagatrader9960 - 28.10.2023 23:58

thanks for your time to make such an informative video. Since we use X3D Ryzen9 .. our concentration is at CCD0 to CO and concentrate at the best Core from Ryzen mastery. This will reduce Temp accordingly .. but is there any issue if we apply the same for remaining CCD1 cores ? i mean it suppose to reduce Temp more but in a price of Overall performance ? u suggest that ? personally i dont mind if i loss little performance to gain extra wattage reduction ? ... also what u think about reducing PPT TDC EDC ?

Ответить
@PaulyDownUnder
@PaulyDownUnder - 16.10.2023 01:50

On Ryzen Master utility in Basic View (not Advanced View) you can have the utility find the best values for you by running the "Curve Optimizer Mode" on "Per Core". It will take around 2hrs to run the test and give you the correct undervoltage value for each core.

Ответить
@dr.tobiasgoodfellow5174
@dr.tobiasgoodfellow5174 - 28.08.2023 22:49

Since you were using Ryzen Master, why didn’t you just use the curve optimize mode to calculate the best values, since it has that feature. Or at least show that in your video and your reasoning for doing it manually?

Ответить
@cheesyvoid3143
@cheesyvoid3143 - 31.07.2023 15:13

Just built my 7950X3D system last week, and unfortunately, I didn't do enough research about DDR5 compatibility, hence why I accidentally bought 2 sets of 32GB DDR5 6,000 MT/s (CL36) RAM having a total of 64GB. I have been tweaking the memory for a couple of days and gave up using the EXPO profile because of instability issues, initially at least.

Then, 2 days ago, I realized that my BIOS version is from July 2022, so I updated to the latest one. And lo and behold, the EXPO profile now works with 4 DIMMs, and I was able to push it to 6,200 MT/s. As for the CPU, I was able to do an all core offset of -32 and it is stable with the LLC set to Extreme. So far, cores 8 to 15 are always running between 5.20 GHz to 5.75 GHz. I guess I won the lottery.

Ответить
@lendial
@lendial - 24.07.2023 19:42

I was able to set all cores to -30, would lowering the offset on my best cores yield performance or temp gains or do i need to test every combination manually?
Also, I have expo enabled which raises my SOC voltage from about 1.0v to 1.303v which is still safe but id like to reduce it but even 1.295 still causes stuttering, if i reduce the offset my curve optimizer would this allow me to reduce my SOC voltage a bit?

Ответить
@Mexi99
@Mexi99 - 23.07.2023 03:31

So question. I’ve been doing per core. Now and trying stuff out. Do I put my best cores to a higher negative and the rest higher? So say my best cores are 1 through 5 do I put those to the highest negative code as I can before the system crashes? Right now my best comes from what I’ve been playing with is 14,12,1,10,16 and
I got them to 35 with out crashing do I go higher wtih the rest of the cores?

Ответить
@ifytch1935
@ifytch1935 - 26.06.2023 04:54

Happy belated :P
I and I think many others would comment on the video with the 13700k :)

Ответить
@xiv7413
@xiv7413 - 14.06.2023 02:09

I did exactly this and my clock speeds are lower than stock and my cinebench score is 1k lower than stock and without pbo enabled in bios. When I only used CO without pbo, my cinebench score was 13580 and with pbo my score was 12000. Temps didn’t change when I did exactly what you did in the video. My cpu is ryzen 7 5700x

Ответить
@Directx11cool
@Directx11cool - 12.06.2023 20:38

y cruncher finds unstable better. While OCCT shows no errors, y cruncher crashed in 10 sec.

Ответить
@oats7
@oats7 - 05.06.2023 19:50

Hi I'm trying to undervolt my 7950X3D. You have a previous video on undervolting this CPU, but you did it in the BIOS. Which should I use? Ryzen Master or the BIOS?

Ответить
@ApplePotato
@ApplePotato - 01.06.2023 01:23

Very good information. Especially on the fact that you need to voltages higher on your best cores to keep the frequency up. But I think a disclaimer on what you define as stability would be nice. I find that the AIDA64 SHA3 test is next to impossible to pass with any cores at -30. I don't even start real stress run before I can pass several runs of the SHA3 test.

Ответить
@carlb641
@carlb641 - 25.05.2023 06:59

I have set all to -30 and can run cinebench but bsods in prime95. Can only get -20 stable in prime. Were you able to try your settings with different stress sw inc prime?

Ответить
@rindrasi
@rindrasi - 18.05.2023 03:04

@beefy-tech why don't you use cpu boost override and infact have it on disabled? Most have that set to 200 or 150.

Ответить
@wankerka
@wankerka - 15.05.2023 22:53

Hello there quick question. I have a 5900x. Just used ryzen master curve optimizer per core. Took me a 1hr and 30 mins to finish. Lol. Well anyway is it possible to input the results manually to the bios and uninstall Ryzen master? I have done stress test seems stable . The result i got -27 almost all cores on 2 cores one was 18 and the oher was 23.

Ответить
@Pytte
@Pytte - 14.05.2023 04:39

my 7800X3D was able to do -30 on my 2 fast cores, and -35 on the others. Under that I got worse results.

Ответить
@djchotus1
@djchotus1 - 12.05.2023 16:25

Woooo wait. He said - 40?

Ответить
@jonathand7465
@jonathand7465 - 09.05.2023 06:47

did they make that you dont get the start optimizing button (curve optimizer) for the 7000x3d ? i only have validate button there

Ответить
@mich5528
@mich5528 - 06.05.2023 22:34

Maybe a noob question but why not simply enable ecomode 105w in bios and be done with it?
GN tested that mode and it was only 3% slower than ‘unleashed’ but actually almost halfs the power consumption.

Ответить
@historickeeper
@historickeeper - 05.05.2023 14:18

Very nice video I have a 7800Xd3 core 2 best core 1 2nd I did a all core not per core -30 a is stable as a rock is it worth pulling back on 2 and 1 if they are stable odd thing is don't really is much change in temps but it does hold higher clocks for longer.

Ответить
@AndyU96
@AndyU96 - 05.05.2023 00:26

May I ask you what cooler you have on that 7950X, along with the thermal interface material you're using and the maximum temperature your CPU reaches in 3 consecutive runs of Cinebench R23? (Also what is your ambient temp)

My parameters:

NZXT Z63 RGB 280mm AIO
Thermal Grizzly Kryonaut spread on every part of the CPU in excess
Ambient temp 22C
Max temp my CPU reaches after 10 mins of C23 is 89 and it barely reaches that by the end

Ответить
@Jayy_R7
@Jayy_R7 - 04.05.2023 17:34

W video cant wait on the full amd pc overclocking vid

Ответить
@estebanrodriguez923
@estebanrodriguez923 - 04.05.2023 10:18

nice vidéo ! With 7800x3d since the bios update of 04/27 which blocks the cpu soc at 1.3v, impossible to activate the expo profile. 6000mhz as 5600mhz. CO -30 all cores crash the pc. I am in -20, PBO limit enanhcement level 3 70° and thermal throttle 85. With these parameters the pc runs well but I think I can do better to gain in temperature and performance but I don't know what to change

Ответить
@tendou4620
@tendou4620 - 04.05.2023 07:31

Ive just put -15 to ever core so I get less temp and use less wattage. didnt even bother to increase
core frequency bc 7950x3d is a beast already. Probably it aint doing much anyways 🙈

Ответить
@pipec4t675
@pipec4t675 - 04.05.2023 05:59

Can you do a video about ryzen 5000? Or it's basically the same and i can do it just like you did for the 7000? I didnt really understand really well what u said in the end of the video

Ответить
@coolbvwes1
@coolbvwes1 - 04.05.2023 04:08

Test all you're curve settings with aida 64 cpu sha 3 benchmark, that's the hardest one to succeed with a undervolt. My 7800X 3Dwas stable with - 30 on every benchmark but crashed at sha 3 😂

Ответить
@Omni-Wan_Kenobi
@Omni-Wan_Kenobi - 04.05.2023 03:44

I have -30 on every core. I'll definitely give a try your way tomorrow morning. Can't wait;)

Btw, On my MSI board there were for only 1 Bios version the Enhanced Mode Boost PBO and those were giving some good results with higher boost on every core. Normal boosting on every core is sth like 4880MHz with Enhanced there were 4950MHz. They promised they ll bring them back in the future. Maybe your Mobo manufacture will also add such future.

Ответить
@nivek318
@nivek318 - 04.05.2023 01:31

Is it necessary to change TDC & EDC? Why do you have them set the way you do?

Ответить
@Mexi99
@Mexi99 - 04.05.2023 01:30

i just put in op curv in bios to all corres -40 and my cpu is running great

Ответить
@patrickfuchs7552
@patrickfuchs7552 - 03.05.2023 23:26

It works great👍👍👍

Ответить
@tommythegun2439
@tommythegun2439 - 03.05.2023 22:49

Hello @beefy-tech. Thank you for your tutorial. I have the 7900x3d and am considering undervolting as well. I noticed you made comparison between two configurations but I wonder how much better it is against no undervolt at all? Thank you.

Ответить
@MarcoGiese-mv6cm
@MarcoGiese-mv6cm - 03.05.2023 21:41

Hi i got a Question !! I am sorry for my bad English so I try my best. I really interested for optimizing my Pc by your Hand. I saw the Formular whitch you have to rank my Knowing in Computer from 1-10 my is 3 so its better to have someone by my site who is better in english and in computer ? I got a Ryzen 9-7950X3d /RTX4090 Gigabyte OC/Venegance DDR5(2x16GB) 6000MHz C30/Rog-Strix x670E-F and my best Results in MW2 Benchmark is 310FPS. I watch all the Videos and tips for ProcessLasso and Settings you make . For an answer i wil be happy and sooryy again for my english.Thanks.

Ответить
@AndyU96
@AndyU96 - 03.05.2023 20:07

Buddy, I like your videos so far and all but you cant just compare those MW2 benchmarks and say 'one seems a little better than the other'. Why? Here is why:

What are the differences between the framerates in percentage? They are too low to be meaningful. The differences could very easily, or rather should, be attributed to run to run variance, unless you tested both scenarios numerous times and were able to assess the standard deviation of your testing method. If you told me that the standard deviation of the testing was 'somehow' SO low that even a difference of 1% between the framerates is meaningful and outside the margin of error, then I'd say okay fine

Aside from this small thing, you're being a great help

Ответить
@imrin4059
@imrin4059 - 03.05.2023 18:41

Or it doesnt work?

Ответить
@imrin4059
@imrin4059 - 03.05.2023 18:34

Can I do this with 7800x3d?

Ответить