SaHaRzZz

10 hours ago, Blue4130 said:

I have a e5 2650 and gtx750, my load tests with both at 100% are 177w.

 

I don't see your setup being double that. 

Great to hear
Thank you!

I want to run this CPU coupled with a GTX 550TI for the display, and use it as a 24/7 server, that will be on 100% sometimes.
The question is.. is a 450W power supply enough? because whenever I check with PSU calculators, it shows that the max consumtion is under 400, and the recommended is 400-450W PSU. will that be enough for my purposes?

Also while at it.. I'm going to need a budget CPU cooler to go along with it, I'm thinking about the Hyper 212, is there a better suggestion? because I don't really know how this chip behaves.

45 minutes ago, thrasher_565 said:

you can go to pcpartpticer and select case that can do 280 and 360 rads oh nvm guess you cant my bad

 

newegg you can with the fan options

 

https://www.newegg.ca/p/pl?N=100007582 601317102 601317057 601317254 601317175

Yeah, it's not very specific still but thanks.
Trying to find something that will fit me is hard tbh, the only perfect one seems to be the 1000d, but it's so damn expensive

18 hours ago, Tech87 said:

I use the coolermaster nr600, and find it great for cooling with its front mesh design.

I've heard good things about the phantek p400a digital as well.

18 hours ago, davemc said:

Phanteks P360A, Corsair 4000D Airflow

I appreciate your replies, the phanteks p360a, p400a, and the Corsair 4000d airflow, are roughly the same price in my country, but as for radiator support I'm thinking about something.. will it be possible to put my 2 radiators (280mm & 360mm) both as intake? that way the CPU & GPU will get cold air, right?
Something like what the Corsair obsidian 1000d is capable of I think, but well.. not that ><

So my CPU temps are never low for some reason, I swapped to a new motherboard, downgraded to an 3600 from an 3900x, and even swapped from an X62 to an H150i, and temps are still.. not the best, 50c at idle.

which makes sense because my setup is crap, I'll explain why.

I have the H710, with a 2080ti + Kraken G12 & X62, mounted at the front, with push&pull configurations, 48-50c at load with the GPU, I'm very pleased with that.

After that, the CPU, which is now a 3600 + H150i, mounted at the top, with push configuration (can't really do pull on that one), 50c idle and around 75~ish at load.

The H150i liquid temp at idle is 40, which is pretty high for idle, and probably is because of the front intake giving hot air.

And of course I have a rear 140mm fan, BTW all of my fans are pretty good, Corsair ml pro series & noctua, only the H150i has the stock 1500rpm ml120 lies edition fans.

What I thing I need is a new case, because I see everyone bashing the H710 for poor airflow capabilities, and if that is what will bring me to sub 40 idle degrees, then that's what I want.

I'm completely open to hear case suggestions!

8 hours ago, jaslion said:

More paste will not help. The biggest issue here is that the cooler simply does not cover the whole cpu. It wasn't a huge issue back with old intel chips but amd uses a 4 chiplet layout on the 3900x causing it so that the actual parts that get hot in the cpu are maybe covered for 2/3 and that is it. The rest that the cooler covers is basically empty space. This is because it simply is a pre ryzen cooler and not meant for ryzen cpu's. They made a adapter plat after the fact so you can use it but that does not change the fact it was not made with ryzen in mind hence the obvious deficit here.

I've seen more and more videos of people online with the combo of the 3900x with the X62, and they are reporting idle temps of 35~ and up to 70~ on load,
makes me worry I have something wrong with mine, but I already am planning on getting the H150i as a replacement, hopefully to solve this issue.

8 hours ago, jaslion said:

A sycthe fuma 2. Don't need watercooling at all when a fuma 2 is better than 95% of 280mm aios anyways.

Thing is, I really like the silence and the fact that the heat gets blown directly outside the case instead of inside.

5 hours ago, icase81 said:

I had an H115i Pro on my 3900X that did a very good job with the included fans fairly quiet. I had them at 600rpm until 80C and they never ramped up, with idle being around 38-40C. This was in a Fractal Define C mATX, so the front is fairly closed off. Also had a 5700XT with a G12 and a 120mm EVGA AIO on that in the exhaust. You can pick them up for around $100 recertified from Newegg or Amazon.

Beautiful PC!
Something about the G12 just makes things look more prettier IMO.
I'm thinking of getting the H150i, hoping it'll do the job better.

2 minutes ago, Chris Pratt said:

More thermal paste is not a good idea. You can try tightening it more, but if the issue is incomplete coverage of the IHS, nothing is going to fix that. You'll need a different AIO.

is there a good one that you would recommend for the 3900x? I'd rather a 280mm one, because I already have 2 fans for it, but I can also fit a 360mm if there's something that is worth getting.
I wasn't aware until now that the X62 was bad the 3900x, was completely clueless.

Just now, jaslion said:

Your case has bad airflow but still that is insanely high idle temps and you are actively thermal throthling at 95c. However the x62 is a pre ryzen cooler and if memory serves right does not cover the whole cpu and has been know to thus let it run hotter. Also with the am4 adapter kit the mounting pressure is very hit or miss.

if the mounting pressure is not very reliable, should I apply more thermal paste or something? or maybe tighten the cooler even more?
Very informative answer, thank you for that

Hello, I have an issue with my 3900x CPU, that has been going for since I first got it, it runs at around 70c, sometimes spikes to 80c, when completely idle.
if I'm gaming, it goes to around 95c, which is bad.

My case is NZXT H710, with a front mounted NZXT X62 for the GPU (2080ti with g12), and a top mounted NZXT X62 for the CPU.
My fans for everything, including the rear fan, are 140mm Corsair ML120 Pro RGB.

The GPU is completely OK, with temps sitting at 30-40c idle, and around 55-60 when gaming.
The CPU is completely NOT OK, well, I already explained why.

I have tried replacing the thermal paste but it just didn't seem to change anything, I have a few pics from HWiNFO if that would help if something is abnormal,
do you guys have any suggestions on things I should try?

1 hour ago, muito_gostoso said:

use just 1 stick of ram for testing!

 

well i was to say maybe you have thermal trottling i see 86.1 c and 87.4 c, using msi afterburner you can spot thermal throttling and frezees easily.

 

another thing is if you can plug and replug all your components to check connection issues.

 

if you have some other spare part you can try mixing components and see if frezee still happens.

 

setting your bios all to defaults could help to.

 

sorry to be to general but wen i have frezees its by drivers or bad o.s. configuration,not hardware related.

 

 

can you confirm yo are not thermal throttling like this with afterburner?

 

are you using windows 10 right?, sometimes exist also S-O throttling in windows due amdppm.sys disable in testing ,run these.

 

 

1.-RUN REGEDIT MY THROTLE STOP TO SEE HIGH PERFORMANCE TAB IN POWER OPTIONS.

2.-Browse to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\amdppm.sys and set the 'Start' value to '4' (=> disabled). AND REBOOT.

 

3.-Browse to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\CONTROL/POWER/CSENABLED --------------< set 0 and reboot.

 

cheers.

 

 

 

 

 

 

One more thing, I opened my case(NZXT H710) and each time I do I find this disturbing(image).
It's like the AIO is blocked, although there is technically a way for the air to be ventilated it seemed like it's almost nothing.
I couldn't find anything related to this on the internet, probably because the case is still new

24 minutes ago, muito_gostoso said:

use just 1 stick of ram for testing!

 

well i was to say maybe you have thermal trottling i see 86.1 c and 87.4 c, using msi afterburner you can spot thermal throttling and frezees easily.

 

another thing is if you can plug and replug all your components to check connection issues.

 

if you have some other spare part you can try mixing components and see if frezee still happens.

 

setting your bios all to defaults could help to.

 

sorry to be to general but wen i have frezees its by drivers or bad o.s. configuration,not hardware related.

 

 

can you confirm yo are not thermal throttling like this with afterburner?

 

are you using windows 10 right?, sometimes exist also S-O throttling in windows due amdppm.sys disable in testing ,run these.

 

 

1.-RUN REGEDIT MY THROTLE STOP TO SEE HIGH PERFORMANCE TAB IN POWER OPTIONS.

2.-Browse to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\amdppm.sys and set the 'Start' value to '4' (=> disabled). AND REBOOT.

 

3.-Browse to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\CONTROL/POWER/CSENABLED --------------< set 0 and reboot.

 

cheers.

 

 

 

 

 

 

I've changed the values in the registry, now when I tried running cinebench and watching the temps, I didn't hear the fans and Ive opened iCue to discovered that the commander pro section is bugged.
Something is very wrong and I really don't know what, but I'll start with reinstalling this software and go from there.

Should I do anything specific afterwards?
Like removing the ram sticks and testing with 1 installed? is it related to the temps?

2 hours ago, muito_gostoso said:

Run one by one test!

 

cpu test intensive first like aida64, cinebench etc.

 

then gpu test like ati tool or vantage.

 

then check with afterburner and hardware info the voltage and ghz, specs.

 

Its very general but since the mix ram-mobo-cpu-gpu= frezee , you have to spot the dirty one first.

I know for sure that the GPU is alright, because it was brought from my previous Intel build

And the thing is that I think the freezes occur in RAM related stuff.
Like whenever I use high amounts of RAM, it freezes, it doesn't happened when I'm gaming.

But as for the tests, I did lots of them a few months ago with the same setup, memory held fine @3600mhz cl16 in memtest86,
I ran prime95 and cinebench and heavyload and everything seemed alright.

idk what do you mean by checking with afterburner? do you mean using kombustor or checking my OC?
I'm not OCing at the moment and my temps with my GPU are really good.

with hwinfo it gives me these values in the CPU section(uploaded an image), and I'm idling.
Already checked background proccesses and formatted, it didn't change much, I'm not pirating anything so I don't have any viruses to drain my resources.
And I am getting high temps in BIOS anyway so I don't think it's related to software.

Hi, for the past few months I am having probelm with my cpu, memory and motherboard I think, the combo I have is x3900 + (4x8) 3600mhz cl16(G.skill B-die) + crosshair vi hero.
Now I know that the MB is rated for 3200mhz, but it lets me put 3600mhz(not sure if that's bad or not).

Now the issues are that sometimes my PC freezes completely and I have to force shutdown, and when I boot back up it doesn't boot and gives a random memory related Q code,

and after a second hard shutdown it goes into safe-boot, and I restart the PC again and it boots normally.

As for now I underclocked my memory to 3200mhz cl14 with ryzen dram calc, hopefully it'll be okay.

The main issue I think I'm having though is related to voltage draw, as my cpu shows in BIOS it draws around ~1.5V and my ram draws ~1.43V.
Doesn't matter what I put in manual, it just doesn't listen, and if I'm trying offset it's the same deal.

And my cpu idle temps are around 65-75 and my cooling is the Kraken X62.

Is the MB the issue here? Today I tried to boot the PC up after not using it for around 9 hours, it gave me Q code 8, which is CPU related I think.

The build is brand new, about 4 months old, I came from intel so being tight with memory and stuff is new to me, if someone knows better how to deal with this and hopefully solve it I'll very much appreciate it.!

29 minutes ago, ShrimpBrime said:

If you actively cool the memory sticks (direct fan on them) you can run the v-mem voltage higher than 1.35v. 

Intel states 1.5v is a maximum safeness. I run mine at 1.6v (I do not suggest this for people, that's just what I do)

So at 3800mhz, try increments past 1.35v to either get it to post, or stability. You likely won't need more than 1.410v v-mem.

 

I'm curious what memory you have. 

Open this up, and show us what exactly you are running.

(pulled from my sig)

 

 

thphn160.zip 1.72 MB · 0 downloads

The kit is in the name, as well as all the information about it.
by far it's awesome, I've left the voltage at default and it's at about 1.43 at the moment, maybe I'll change it idk if I should tbh.
The sticks are at about 43c, I think it's alright ^^ will run memtest86 later to be sure of that tho

G.Skill F4-3600C16-8GTZKK.html

3 minutes ago, ShrimpBrime said:

Info - 

 

Ryzen 3000 series processors have a throttle point of 95c. Your 84c, has a little headroom left.

Want to run IF and memory clocks linked as high as possible. 3800mhz is a goal. In most cases IF won't do 2000mhz, so need to unlink at this frequency. 

Tweaking SenseMi technology to your particular rig takes time. Not so cut and dry as manual overclocking is. There's quite a few settings to play with "boosting" clocks if you will, again, takes time to tweak it all. 

 

Are you good? Well yea aside from running a Gen 1 board with a Gen 3 cpu.... I think you're doing fine.

 

The cooler you run that Ryzen Cpu the better because it will sustain boost clocks for a longer duration. My Ryzen chip is default all core boost of 4ghz, 4350mhz single core boosts. If it runs hot enough, the all core boost will throttle to 3.9ghz and I usually see this once I hit 85-90c as this is where my throttle point is.

 

High temp cpu alarm is at 70c. At this temp, the stock fan profile will put the fan to 100%. This says something doesn't it? Why did AMD boost clocks and voltages past this temp from stock?? Well, to eek out every last ounce of performance leaving absolutely no headroom for manual overclocking.

I got a great deal for the parts and that's why I took the gen1 board with the gen3 cpu,
i'll OC my ram to 3800 cl16 and the IF to 1900 or at least try to, with the dram calculator ^^

Thank you very much!
you and @DildorTheDecent have helped me a lot !

23 minutes ago, DildorTheDecent said:

That depends on what benchmarks you're looking at. I'm basing mine off what would be daily OC results (even though that result is from HWBOT). So that's a RAM overclock that isn't anything too extreme and running the benchmark without tweaks or performance bias settings.

 

 

it was cinebench r20 too.

Every monitoring software shows different values I don't know what to believe anymore ><
I activated the docp profile in the BIOS and it got me to 7008 pts, which is about 80 pts higher.

I also found this thread:

and followed the 'Best Answered' and it reduced my voltage on idle, well according to HWMonitor..
CPU-Z  still says 1.5V - 1.526V, sometimes spikes to 1.199V, I don't even know anymore haha probably 'sleeping/parking' cores or something

idling at about 56-63 degrees, it moves a lot
seems a bit high for me but I guess that's what it is for 12 cores?

Just to confirm, am I good? ><

4 minutes ago, DildorTheDecent said:

Yes. Since there was so much confusion over it AMD engineers on Reddit confirmed it. 1.5V is what you'll see with idle/single thread load. It'll taper down under multi-thread load. Perfectly normal behaviour for Zen 2.

 

I've experienced the same behaviour on my 3950X.

I think it might be. I'm not 100% though. From what I've seen with Zen 1 and Zen + there voltage behaviour was a bit different. Although I didn't use them first hand and have only seen them on forums + YouTube. Maybe the X370 boards are better suited for those processors.

That's what a 3900X scores when it's at 4350(ish)MHz. So it's alright.

at 4350mhz it should get around 7400-7600 from the benchmarks I saw shouldn't it?
I've tested with 0 program running in the background besides ryzen master and sometimes cpu-z

as for the 1.5V, coming from an Intel user that terrifying haha

13 minutes ago, ShrimpBrime said:

Remove Ryzen master fully. Be sure to set defaults in RM before uninstalling it.

 

Shut the board down and clear cmos for at least 10 seconds. Longer if you count fast.

 

Post it up stock and see if it runs normal.

I did that, posted completely stock with these values:

idling at 1.46V - 1.526V according to CPU-Z

testing with cinebench r20:
boosting to 4000mhz at 1.352V according to CPU-Z
Thermals are at 84 according to cinebench

I am using the X62 kraken mounted on top to cool the cpu

my score was 6912, the last one was 6955 but I guess it's kinda the same, but the ram is at 2133 cl15 instead of of 3600 cl16
um.. I should probably activate the docp profile?

1 minute ago, ShrimpBrime said:

Take LLC off of auto. This is why your voltages are higher than input. 

I set it manually to level 1, tried 2 and 3 and 4 and all it did was just make thermals worse and the cpu to clock lower because of that

1 minute ago, DildorTheDecent said:

That CPU voltage is entirely normal for Zen 2. You should see it go down under load. It's best if you don't play with it too much.

 

1.45V is fine for RAM. The scary point is somewhere up nearer to 1.6V for 24/7 IIRC. I think that RAM voltage could be a board thing too. Just make sure you're either using a fully manual setup or DOCP profile.

 

I think some of the issues might be X370/board related. Would a X570 board make it better? Perhaps, I can only guess.

Is 1.5V really OK?
I just finished watching this by jay2cents:

and it describes my issues perfectly actually, besides that CPU error.

and I thought about it.. maybe it is related to my board.
I'm sorry but I'm kind of stressing out right now, it's been 2 days of booting to BIOS and changing values and watching every video I can about tinkering with voltage and the 3900x

my top cinebench score is around 6950~ did I lose the silicon lottery with my chip?

16 minutes ago, Jurrunio said:

That's how you will be behind everyone else because you locked away its ability to boost single core scores. You should adjust PBO while watching the voltage by not letting it go past 1.35V on all cores (and even that is considered too high for someone, they recommend 1.3V for all core workloads)

 

I myself use only the BIOS method though, so can't talk about Ryzen Master. You do have to clear CMOS before you can tune through the BIOS, as Ryzen Master overrides the BIOS's settings (its settings are saved under "AMD Overclocking" in the bios)

 

Lower voltage is not always better

I was overclocking through the BIOS, the Ryzen Master was only for checking the temp + clocks + voltage

Used CPU-Z to monitor voltage sometimes as well.
And by locking the cores to a specific value I meant I didn't want to do it and so manual tweaking of the voltage isn't an option, as it's just lowering my performance for some reason no matter what voltage I set it to.

I saw that video but i'll go through it again, maybe I missed something?
 

1 minute ago, DildorTheDecent said:

I'd clear CMOS and update the BIOS if it isn't on the latest version. Might help with some issues.

 

Also, calm down on that undervolting. -0.5V is far too much. A more realistic scenario is about -0.2V at most.

 

Those CCX clocks of 4850MHz are not going to happen. Don't know why they're there.

BIOS is updated to the latest version as of 2 days ago,
my MB boots my CPU at 1.5V and 1.45V for the ram as default settings and that's why I set this aggressive undervolt, I really don't know that my MB is thinking, but it's just going nuts with the voltage, also getting 1.875 for the 1.8 PLL voltage and weird things like that.

And yeah I know that the 4850 is not going to happened lol, whenever I enabled PBO for increasing my maximum current it also gives me these optimistic values.

If you guys want, I can clear cmos and boot as it is, you'll see these ridiculous values, it's worse than what jay2cents showed in his 3900x performance enhancing/overclock guide

Hi, this is my first Ryzen chip and I'm used to intel OCing so..

I'm just about to lose it haha, ive been tinkering and trying all sorts of values in my BIOS in order to somehow tame this CPU,

My MB is the ROG x370 crosshair vi hero, my CPU is the R9 3900X

My issue is that I can't seem to get boosts(I'm testing with cinebench r20) beyond 4000mhz, I am using offset voltage because there is something weird with using manual in this MB for some reason..

my values are never what I set them, they have like this imaginary offset, like, my DRAM voltage should be 1.35V for my 3600mhz cl16 kit, but I have to set it to 1.27V in order for it to be 1.35V, idk why, it happends with almost every other setting, including SOC voltage and some other ones.

As for the CPU core voltage, it's just weird, whenever I set it to manual and 1.4X / 1.3X / 1.2X or really anything, it never sets itself to it, I go to the Ryzen Master software and it shows as 1.1V and it boosts for like 2.5mhz or something stupid, offset is the only way for it to work properly if I don't set a specific clock for all cores(which I don't want to).

I am right now with -0.5V offset and stable, I noticed that everytime I apply less voltage to my chip I get better boosts (weird as hell) and my thermals aren't that different, always around 80ish when benchmarking and around 50~ when idle, when I'm at -1.0V offset or more, I noticed that the OS crashes whenever i'm idling for too long, but if I stay with prime95 open it holds well.

it's hard lol, i'm struggling alot with this, my goal is to get the 'stock experience' and not to be behind everyone else

I have attached a quick print screen of CPU-Z and Ryzen Master,
I am also confused of why the Ryzen Master shows these kinds of clocks and then CPU-Z shows X42.25

19 minutes ago, Jurrunio said:

depends on the fan, but it's usually you need about 60% to give half the airflow and pressure and at just over half the RPM. This still varies between fans as some do not scale properly at very low RPMs and PWM duty cycle.

That's very informative actually, I am just now watching tons of videos by this YT channel called 'CoolingTechnique'
The videos are all about testing noise & performance, no talking at all.

That makes me think that I should aim for a PWM fan that I'll run at lower speeds and will yield better results, what do you think?

1 minute ago, Jurrunio said:

I just mean 20dBA can be reached by manual tuning, don't need to limit yourself to fans around that.

I really don't understand much about it, let's say..
if a fan provides 59.5CFM at 15.5 dBa @ 100%
does it provide half the CFM & dBa @ 50% ?
it has different behavior for every fan.. right?

Just now, Jurrunio said:

just dont connect the RGB cable and those LEDs will turn off.

 

fan speed control is a thing, use the BIOS.

 

I recommend the Phanteks MP series, their pressure is high for the volume which is pretty important. Airflow is a bit low, but so is every other pressure-optimized fan.

I wrote that I am using the commander pro you know, I am aware of PWM.
About the RGB I just mean that I don't need this feature and prefer to invest my money on performance & silence rather than L.E.Ds.

I will check the Phanteks out, thank you!

2 minutes ago, Juular said:

2x280mm radiators ? What are you cooling ? My list would be be quiet! SW3, NB eLoop and Noctua NH-A, really cheapest of those. be quiet! ones are somewhat limited on RPMs tho, so if you're aiming for pure performance go for other options.

It's a dual X62 for CPU & GPU, I'll search them! thanks