Idles at 60

>idles at 60

nothing personel gweilo

Attached: DAM208411723-0-8ba45a32-0909-4ac2-bb21-62bc5b33cbba.jpg (480x480, 42K)

Other urls found in this thread:

reddit.com/r/Amd/comments/cb8u9f/ryzen_3000_voltages_are_far_too_high/
youtube.com/watch?v=M5pHUHGZ7hU
en.wikipedia.org/wiki/Instructions_per_cycle#Factors_governing_IPC
twitter.com/NSFWRedditVideo

PBO?

My 2700x idle at 40°
Be sure that you are looking at the correct temperature. The CPU usually show 10° more than the real one to avoid overheating and to keep the fan spinning.

ouch

Attached: 1549918825663.png (500x530, 39K)

My 3700x idles at 35, load is 68. Be quiet dark rock pro 4 :)

AIOs apparently don't work very well with the new ryzen CPUs due to the chiplets not being centered. High end air coolers actually seem to get much better performance.

my 3600x is idling at 55-60 with an arctic 34 duo push-pull aircooler, I fucked up so bad.

Manual OC or PBO? People are reporting PBO more stable than manual OC.

I didn't OC it, just plugged the thing in and ran it. Ryzenmaster is saying the clocks for each core are running at 4.2ghz, which I guess means it's boosting itself on its own. Is it normal that it should be running this warm with boosted clock speeds?

>12 core CPU has higher temp under load than 6 core CPU
AMD BANKRUPT AND FINISHED FELLOW INTARDS!

cope

Check your thermal paste.

Has nothing to do with cooling Pajeet. Because Ryzen was designed by the brightest street shitters in India it parks single core voltages at 1.45V - 1.5V (instead of advertised spikes) which cannot be cooled down unless you delid it. It's basically ticking time bomb.

Wasn't the 9900k supposed to be a house fire?

>9900k at 57 degrees when literally every reviewer spoke of how it melts under anything that isnt a top tier water cooler and even then it's still hot
discarded

Yes it hurts.

Attached: Screenshot_20190718-120552.jpg (1080x1920, 489K)

Try undervolting around 3.8 - 4.1GHz. Stock settings and PBO are always going to use a high voltage.

>4.6ghz 12core
>only 79C on load
Thats betty good.

Thanks user I will go into bios and disalbe PBO and report back

Not him and sorry for reddit cancer but this seems relevant:

reddit.com/r/Amd/comments/cb8u9f/ryzen_3000_voltages_are_far_too_high/

Basically just 100mV voltage offset instead of tedious underclocking. PLS report results, my 3600 will come in the mail soon.

Attached: IMG_1899.png (1136x640, 1M)

disabling PBO didn't seem to put a dent in temps, it's still idling at 60-low 70's.
I will now go and undervolt and report back

Try the offset thing first, disabling PBO will force ALL the cores to maintain a specific frequency at a specified voltage, it's going to be shit. The negative vcore offset will just simply tell PBO to fucking chill with the voltage and possibly gain you better single AND multi-core.

>new build
>swapped this to amd bracket
>fans rarely need to spin even under load
best pc related purchase I ever made

Attached: be_quiet!_dark_rock_pro_4_06.jpg (1920x1280, 331K)

no positive results, it's still idling in the 60's.
Disabling PBO + adding an offset of -.100v had the most noticeable results, it idled in the high 50's instead.
Should I undervolt it even more? it's sitting at 1.36v right now

You're getting 1.36 at 4.2 GHz?

Yep, but it's a bit jumpy and skyrockets to 1.375 and around on some instances.

and apparently it's idling in the high 50's again, now. so I guess there was some kind of improvement. I will replace the thermal paste with something alot better osmetime soon.
This is extremely annoying

I'd do at least 5 minutes of prime95 before cranking the voltage down any further, sounds way too good to be true desu. Stock PBO says it does like 1.45v for single core.

Attached: cpu.png (400x396, 23K)

Not him but this is i7-8700K stock single threaded performance after taking 12% IPC uplift over zen+, right? Can you do 1T cinabench??

Attached: Screenshot_20190718-201319(1).jpg (720x1021, 56K)

stay mad Rakesh

Attached: file.png (1920x1080, 696K)

Mine idles a 1.45v at 38c. Sounds like you just don't know how to install a cpu cooler.

Heck, I ran my 1600x at 1.4v just fine for 2 years.

Also has a 4ghz base clock I rarely see it go below
It barely tops out at 60-75 either
I had that cooler it was shit
D15 or clc

Is this cpu inside that top part?

It's not even that the CPU generates an extraordinary amount of heat so much that the heat is concentrated in such a small area that it cannot be effectively cooled. It's a flaw with the chiplet design.

I guess you'd know about small areas generating heat

I'm a well endowed guy who generates no heat because I browse Jow Forums - technology.

1700x 3.7GHz 1.35v

Attached: 3.7ghz.png (641x790, 48K)

the cpu contacts the top part

Attached: 20190719_111517.jpg (487x420, 52K)

Don't be afraid of them volts you faggots.

Attached: 1563481708675_0.jpg (1536x2048, 582K)

You guys are retards, der8aur measured that ryzen idles at .20v

This is me
I just installed that shit two days ago. Its fucking perfect.

yes

Yikes I'm glad I didn't get a 1700X.
My new $150 2700 is crunching in the next room at 3.8GHz 1.22V.

Has anyone tested out a graphite pad on these things? It's like the thermal paste can't get the heat out of the corners fast enough but those graphite pads have really good horizontal heat transfer.

cool story bro
youtube.com/watch?v=M5pHUHGZ7hU

They must've fucked up their test. There's no fucking way a 8700k is 4c warmer than a 9900k under load.

Depends on the motherboard and memory speed. Some guy posted that anytime he ran his memory at 3600 MHz or above on his Gigabyte board that the idle voltage would jump up super high.

i tried 3 different aio coolers.
all 3 died the longest living one lasted 7 months. went back to the evo212 and still running my fx9590 house fire on it

It also idles at a pretty high wattage.

Attached: power-idle[1].png (500x730, 44K)

wait, so what's the wattage under full load? is the 65W sticker bullshit?

>is the 65W sticker bullshit?
Yeah. The only real restriction the 3600 seems to have is that it will absolutely not go past 4.2 GHz.

Attached: power-multithread[1].png (500x730, 49K)

Interesting, so does this mean wattage calculators are all bunk as well? I just want to make sure a 3600+5700XT can work on my 520W PSU. I'd prefer not to upgrade if possible.

AMD measures TDP by the base clock.

Weird for everything on there, my 3400G is idling at 10W maximum right now, at least according to HWinfo

>Total system power
65W is for chip-only consumption

My 3900x idles at 26c.

>AMD measures TDP by the base clock.
False. AMD's sustained boost uses up to 65W.

HES POSTING WHOLE SYSTEM POWER. READ.

>Interesting, so does this mean wattage calculators are all bunk as well?
Possibly.
>I just want to make sure a 3600+5700XT can work on my 520W PSU.
It'd definitely work, 5700 XT should only be capable of drawing 225W. So you'd probably only use 300-325W under absolute max load.
>AMD measures TDP by the base clock.
That's a really weird choice since the base clock is never used unless you've got dogshit cooling and/or VRM.
>Weird for everything on there, my 3400G is idling at 10W maximum right now, at least according to HWinfo
I think the high idle might have something to do with both the X570 chipset and the I/O die. So 3400G is probably ideal in its idle compared to Zen 2.

Attached: boost-clock-analysis[1].jpg (976x504, 86K)

>I think the high idle might have something to do with both the X570 chipset and the I/O die. So 3400G is probably ideal in its idle compared to Zen 2.
Also, maybe Zen 2 just has fucked BIOS' right now. I remember the original Zen launch had really shitty power defaults when it first came out.

Pretty sure the high idle power consumption is a result of user error. Zen 2 requires a chipset driver installed power plan to clock correctly until windows implements the new changes themselves. I imagine most people never read the manual or look into the topic, nor do they check for a bios update.

Chipset driver, direct from AMD's own site is basically required on zen 2.

>zen 2
>reaching 4.6
audible kek

My 3900x is hitting 4650mhz regularly. Read the manual, install the correct drivers.

Attached: 752bc65e963660dc7cbc86d671592ddf.png (1255x397, 20K)

>Pretty sure the high idle power consumption is a result of user error.
It might be but then that means literally every review out there has followed the wrong instructions.
Another thing I've seen done in quite a few reviews is people setting memory frequencies to 3000+ MHz and having the IF freq getting fucked by the BIOS automatically setting the IF to a 1:2 ratio.
Yeah, I noticed none of the R9 reviews showed it hitting boost clocks.

The reviewers also used x570 which is exceptionally broken for the time being. x470 is the best platform to extract the most from zen 2, at the moment.

Aren't safe max temps at a lower value now?

You'd have a hard time achieving them unless you do something wrong. As far as I know they're at 95c this time, the previous generation was 85.

delet

Why even though Ryzen is on 7nm it clocks so low compared to Intel?

7nm has higher thermal density than 14nm. If Intel was on 7nm I highly doubt they could get 5 GHz.

My ryzen 3 1200 idles at 25 and doesn't go over 50. Run everything ultra.

You have your reasoning backwards. The smaller, denser, and more advanced the node is, the harder it is to achieve higher clocks.

Next gen intel processors will clock lower than 9th gen, without a single doubt.

If you think clockspeed is even relevant, you need to remember the disaster that was bulldozer, 5ghz means nothing.

All the ryzen temps seem good to me. My old 2500k and 4690k systems are oced to 4.6 and generally sit around 32 idle. That's four cores granted old chips at this point.

Go home intel you are drunk again

Attached: 191156565.jpg (512x512, 12K)

So 3nm and beyond will be even lower clocked and cost even more?

Absolutely.

The deal is, once we hit 3nm, a 1ghz chip will perform many times higher than a current 5ghz chip.

What if both of these chips had all its cores disabled except one. Which core would perform better?

I'm talking about a single core. Yes, core for core a 1ghz core on such an advanced node will beat a current 5ghz core.

People don't really understand this stuff.

Zen 2's core outperform's intel's 9th gen core at the same clockspeeds, by a significant margin. The reason intel wins in some tests is due to architecture and optimization differences, not the core being faster.

Generally new nodes are clocked lower and cost more. Pentium 4s were almost at 4 GHz in 2004.

>I'm talking about a single core. Yes, core for core a 1ghz core on such an advanced node will beat a current 5ghz core.
How so? Could you go into detail, I really want to know more.

Think of each tick of the processor in the frequency as a tick on a clock. So you have a bucket worth of calculations you can perform per tick of the clock. The more advanced the core architecture is, the more calculations it can perform per tick of the clock.

Basically the bucket will be so large that even at a massive clock speed deficit, more calculations will be performed in the same time period.

It can with CCX specific overclock.

They were also dogshit slow allegedly due to the pipeline meme.

You can also just have a better architecture with way more IPC at or near the same node.
en.wikipedia.org/wiki/Instructions_per_cycle#Factors_governing_IPC
Pic related, 16nm Chinese CPU getting dominated by 14nm Zen.

Attached: Untitled.png (854x4013, 119K)

Thank you. How will prices shape up to be exactly? I thought we'd get if not 16 core at least 12 core chips replacing 8 cores at the 300$ price point. Will it now always stay 8 cores for 300$?

My seemingly primitive R7 1700X idle between 32C and 41C with water cooling.

Goes to around 50C-60C when playing games (not maximum load, could get hotter if encoding videos or running a benchmark test, etc).

Cores will keep exploding as technology pushes forward. There's certainly a limit on how many cores are practical, and it's pretty firmly around 8000, and that's for the most perfectly parallelized workloads on earth. This is why you'll never see a gpu with more than 8000 cores/sps.

Prices will keep going up as R&D, time, and effort keep ratcheting up as we march into the ridiculously small nodes. Just remember we went from 8 cores being $2000+ to less than $500 when ryzen first launched. Now we have 12 cores for $500. Where do you think this leads? I think it's obvious. We're getting 64 cores on HEDT likely before year end.

When will we see 16 cores for 300$ or so? I know it's kind dumb question.

1950X will be $300 in a year or two.

Well the 12 core 1920x was being sold for $300 throughout the week.

I imagine the 16 core 1950x will shortly follow.

Is there a possibility for AMD to launch a 300$ 16 core chip?

The intel cope here is real

my 6600k idles at 23c, room temperature and i've NEVER seen it go higher than 45c all with some low end $30 cooler master air cooler.
these high temps are a big concern to me for the 3000 series. does the 9700k also really get that hot or are you guys just memeing. i've never owned an intel cpu that had bad temps so it surprises me.

You're on a 4 core low power CPU, of course it runs cold. OP doesn't know what he's doing with his bios and broke his power states.

you sure looking at tdie not package?
tdie is 20C lower.

chipset 20w tops, CPU 150w tops, GPU 250w tops,RAM what is it now 8w?
count.
I'd also count monitor which is around 30-50w depending how bright you like it
Running PCs was never cheaper than now.

whu retards insist on touching pwoer states on ryzen?
1. it makes things only worse
2.it doesn't increase performance, especially for zen+/2 where they perfected internal control

at 5nm maybe, not on 7nm

My 2700X idles at 32°.

Noctua D14

The TDP at full boost is only 65 watts on that model.
It's physically impossible for it to be hotter than your 160 watt base clocked, 280Watt boost Intel CPUs.

Attached: image0-119.jpg (3264x2448, 771K)