#overclocking

1 messages Β· Page 63 of 1

quick rose
#

VPP, VDDP, VDDG, etc?

sudden torrent
#

I bumped the SoC so fclk is stable, should be 1.1

quick rose
#

Yeah but that's the only one?

sudden torrent
#

Besides that and the DRAM everything else is auto

#

Load line calibration to level 3

#

That's it for power

quick rose
#

This is from Fitz for an issue on X570 but for a 5000 series CPU. Might be worth tinkering with these, except the high SoC voltage for now.

#

1.15 soc
1.0 vddp
1.05 vddg iod
1.05 vddg ccd

sudden torrent
#

I'll give it a go

#

Hoping I don't have to reset BIOS again πŸ€¦β€β™‚οΈ

quick rose
#

I'm coming up with a lot of blanks with regards to both XMP profiles being the same

#

Even with different CPU's or same CPU the whole time?

upper rock
#

Is fitz some professional overclocker or something?

quick rose
#

Yes, he competes with Forks on HWBOT

sudden torrent
#

2 different cpu, 2 different sets of ram

#

Blue screen

quick rose
#

I wouldn't call it professional though, they don't get paid

#

Now, I HAVE read a lot of XMP 2.0 profiles BSOD'ing on that board

sudden torrent
#

Post fail x2

quick rose
#

This might be a conversation you should have with someone like Forks

upper rock
quick rose
#

They're on a team that competes there

upper rock
#

I wonder how high my 3070 will be able to go when I put pads on it

sudden torrent
#

Well time to reset bios again... It won't boot even after reverting changes

quick rose
upper rock
#

How does one join the team?

tall pelican
upper rock
#

Aight

#

cool

#

im there

tall pelican
#

if you go to the benchmarks tab, you can see the supported benchmarks, and each one has a set of rules that need to be followed

upper rock
#

nothing against doing a bench at stock settings just for the hell of it, right?

#

I dont have to be overclocking on HWBot now, right?

tall pelican
#

you dont have to overclock, just follow the rules for each benchmark

upper rock
#

alright cool

#

Ill run timespy or something like that within the next few days when im on my PC

proven canopy
#

Please do read the rules and feel free to ask questions here

upper rock
#

thanks forks

#

will do

tall pelican
#

read the rules twice, bench once BigBrain

#

then realize there's bench variance, and you'll end up running it 5+ times to get the best run

forest radish
#

read the rules twice, bench once, open two cpu tabs of cpu z

#

then don't have the willpower to bench again

sterile flame
#

How to overclock 760?

vernal sail
#

@sterile flame the nvidia card or the i5?

#

There were 2 760s you have to be more specific lol

flint quiver
#

It’s time to finally punish/push my 3090 strix.

uncut shore
#

what program do i use to benchmark if i wanna oc gpu

#

im new to all that

clever epoch
#

unigine heaven probably

quick rose
#

That or 3DMark yes

uncut shore
#

Also can i keep my fans on auto if i already have a good fan curve in the adrenaline driver?

clever epoch
#

auto fan curves are usually fine. If you start hitting temps of 80+C and your fans aren't at full blast, you should take a look at it then

clever epoch
#

I feel like I can undervolt my 5800X, not just using curve optimizer. Might try negative voltage offset

#

that's +200 on core and +1100 on memory on the 3070 btw. EVGA XC3 Ultra

clever epoch
#

Ok voltage offset just reduces my clocks. My temperatures as well, but still. No bueno. Curve optimizer only it is

sterile flame
#

@tall pelican what ram do you use for that 5800x/5900x whatever?

#

want a good kit for 5900x when i get it

tall pelican
#

4000 16-19-19-39 1.4v 2x16

#

if you're doing just xmp, get the corsair 4000 16-16-16 2x8

sterile flame
#

thanks

vernal sail
#

We taking this puppy to 2.6

#

Thinking of doing some serious Overclocking on a qx9650

#

A lot of cheap beefy lga775 mobos

#

Some people can get 4.6 4core on those old core 2 quads

vague gull
#

i get some intel stickers from ebay/ali

torpid siren
#

Hello everyone again.

#

Is 1.5 voltage safe for DOCP/XMP RAM?

sudden torrent
#

DDR4?

torpid siren
#

Yes

sudden torrent
#

For Samsung B-Die I'd call it safe but it will probably degrade quicker

#

Any other die 1.45 is the max I'd do

torpid siren
#

Okay.

#

I'd just got some crazy memory and I just found out that the controller for the RAM is on the CPU

#

Personally, I have a 3950x now, planning for 5900x/5950x later

sudden torrent
#

Correct, but the quality of the die and its rated speed determines the RAM capability much more than the controller

torpid siren
#

Would you like see what I got?

sudden torrent
#

Sure, lets see

torpid siren
#

And if you want the full list of parts, the whole rig. I can as well.

sudden torrent
#

Nice set, definitely B-Die
CL15 you could probably get to 14

#

Keep in mind that your 3000 series CPU is going to need the fclk set manually because it won't work at auto for that memory

torpid siren
#

Maybe, am just skeptical. Am like an overclocking noob.

#

Was like a stable f clock most likely no faster than 1800, at least with zen 2

#

???

sudden torrent
#

With AMD it rounds the CAS latency up to the nearest even number, so it would run at CL16 without any modifications

#

1800 is definitely attainable on Zen 2, 2000 is likely with Zen 3

#

You can probably run the XMP profile for this set and manually set the CAS latency to 14 and see if it works

torpid siren
#

So like 14 16 16 36

sudden torrent
#

Right, because AMD gear down mode will change a CL 15 to 16

torpid siren
#

Huh

sudden torrent
#

AMD had issues with odd numbered CAS latencies, so they started doing the gear down that automatically changed the CAS to an even number

torpid siren
#

Even for zen 3?

sudden torrent
#

Yes, even on my 5800X

torpid siren
#

Strange

#

Lastly, how can you tell what die it is on the ram sticks?

sudden torrent
#

Samsung B-Die is probably the highest quality die, and the only one that will ever have 1.5v on a XMP profile

#

There's lower binned B-Die too but with those ratings there's no way it's low bin

torpid siren
#

It is shame that you can't tell 100% unless you get it and go to like xmp profile to see what die it is

sudden torrent
#

Yeah, you can guess with the obviously higher quality sets like this one pretty accurately but mid and low tier sets it's much harder to tell, and they can even use different dies on the same model

torpid siren
#

That's cool to note.

#

How long you been doing this overclocking? Ram? CPU? GPU?

sudden torrent
#

Well lets see, how long ago was AM2? My first chip I overclocked was a Phenom

torpid siren
#

2006, 07? Welp. 🀣

vague gull
#

falcie old

torpid siren
#

Again, am new to this. At least more of the idea.

sudden torrent
#

Well I was a bratty teenager at the time but yep I'm an old fart

torpid siren
#

I love my first built rig last year with the stimmy checks to help and no regrats so far.

sudden torrent
#

Nice! I know a lot of people used those to break into PC gaming

torpid siren
#

And lately, i felt addicted to upgrade my rig

sudden torrent
#

Definitely, once you get the building bug it itches every time a new series comes out πŸ˜…

tall pelican
torpid siren
#

Well honestly, that's the thing. At least with sockets, you're gotta stop at few years at best and may/hopefully stick with it for like 6+ years.

#

But who knows

tall pelican
#

you can run cl 13, 2t very easily, but its worse performance than cl14 gdm on

sudden torrent
#

Really? I was misinformed then but you're definitely an expert over me

torpid siren
#

With am5 and ddr5 around the corner and that itch could get me hard right there

sudden torrent
#

Yep we're going to need whole system upgrades then once those are released

torpid siren
#

But we see. For me and my rig, Get the new parts, and sell the old ones, and not really losing alot of money. At least now, and if you are lucky to get it at MSRP

sudden torrent
#

I think I'll be ok for at least a year or two, maybe an incremental upgrade. I was just happy to finally upgrade my Ryzen 2700 to something that doesn't suck. Should be set for a while.

torpid siren
#

We just hope the shuffle gods get us the 5xxx CPUs. At least for me

#

And RX 6900 XT

#

Thank you, @sudden torrent

#

Again, I just hope the memory controller on the CPU goes nuts at 1.5V

sudden torrent
#

The DRAM controller can definitely handle 1.5v, it's the SoC you have to worry about. Set that to 1.1 if you have problems.

torpid siren
#

Note

#

Gn

sudden torrent
#

G'night

ember rock
#

guys i need to know if i overclock my cpu or gpu what are the pros and cons pls tell me

#

im new to this stuff

left bladeBOT
#
π•΅π–†π–œπ–˜ β«·β«Έ#1755 has been warned

Reason: Bad word usage

ember rock
#

wtf

tall pelican
#

Im certain google will tell you all the info you need

dark ferry
#

10700K 4.90GHz 4.80GHz 4C+100MHz 1.190V 210W 100% <--- can someone break down the 4c+100mhz and 1.19 die sense for me from the historical bin chart? I know the die sense is the VR VOUT so I'm assuming that means I have to test till I find a proper vdroop that falls within that number under load, but what in the hell does the 4c+100mhz mean?

dull ginkgo
#

I'm guessing it's how well the chip scales with lower temps, aka. for every 4c you drop, you can get another 100mhz out of it

dark ferry
#

Interesting. I thought it was referring to specific cores.

vernal sail
vague gull
vernal sail
#

I have a collection of those stickers mostly from dumpster diveing

#

Well over 50 i think!

#

Probably my most valuable is a clear red pentium pro sticker

vague gull
#

oooh

vernal sail
#

Have a couple black core extreme ones too

#

Aah the i7 980

steady current
#

overclocking a gt 710

#

20fps call of duty

#

woo hoo!

proven mango
#

so x1 precision says gpu +131 mem +200 when i did the auto overclock check, will if affect this lifespan of my 1650?

quick rose
#

No

#

Probs can go further manually

proven mango
#

aight

#

im mainly going for getting more performance without it dieng faster

quick rose
#

That's what overclocking does

proven mango
#

aight

#

so this is basically a small OC?

quick rose
#

Yep

proven mango
#

aight ima try and see how stable it is

proven mango
#

so does overvolting and overheating kill gpus? i just got into the oc scene

quick rose
#

Heat and voltage do, not overclocking itself

proven mango
#

aight so what does overvolting do when done right?

quick rose
#

allows higher clocks.

proven mango
#

makes sense aight thanks dude

#

i dont think it changed the gpu clock XD

quick rose
#

VOltage won't change the clocks

proven mango
#

where can i put a pic of it?

torpid siren
#

Morning all! πŸ‘‹

proven mango
torpid siren
#

Or simple... good day! 😜

proven mango
#

good day

#

yeah so the memory clock straight away changed to 4201 when i applyed it but the gpu clock stayed at 1485 and didnt change

proven mango
quick rose
#

VBIOS limit probably

proven mango
#

aight so only mem clock will change prolly or is there a way to fix it?

quick rose
#

MSI afterburner, all sliders to the right. That's all you can do without a custom VBIOS and it's not worth it for that card

proven mango
#

kk

#

would +400 still be good?

#

on ram

proven mango
#

ok its appearing now all i needed to do was load into a game

cold whale
#

Can overclocking hurt your cpu?

dull ginkgo
#

Yes

#

High voltage and high heat can deteriorate chip slowly

cold whale
#

Ouch

#

Is there a good cpu cooler for overclocking

dull ginkgo
#

any decent cooler, depending on your chip

#

a 30$ se-224-xt will keep anything below like a 10850k cool

upper rock
#

Dark Rock Pro 4

cold whale
upper rock
#

you could honestly just go stock with that thing tbh

#

Ryzen 3 comes with stock coolers, right?

severe forum
#

bc vintage said no

quick rose
#

I wouldn't

cobalt rapids
#

it'll work okay for a 3700x if you don't do any like extended multi core stuff I think

#

but yeah I wouldn't get one for that purpose

#

worth trying/testing if you have one though

dull ginkgo
#

@severe forum stock/pbo, I'd say yes, actual all core oc, or smth all core synthetic, no

torpid siren
#

@sudden torrent

#

Like what does it mean on SoC?

#

Like I was just looking around the BIOS since someone suggest to default the settings except XMP, and I thought the SoC voltage would be affected along with the DRAMS since the controller is on the CPU, but at least current ones right now before my set comes here, doesn't do that.

#

DRAM voltage is normal since that is what it needs in order to work with this kit.

sudden torrent
#

The SoC is part of the voltage that goes into the cpu and is partly used the infinity fabric. The f(abric)clk increases along with memory speed, and requires more voltage to work at those higher frequencies. Insufficient voltage will cause system errors and crashing.
1.1v is good for higher frequencies and is well within the safe range.

torpid siren
sudden torrent
#

Correct

torpid siren
#

kk

#

I swear, with voltages. researching can be a real headache for a first timer.

#

We will see tomorrow for the new set of RAM

sterile flame
dark ferry
#

Which two OCCT tests am I supposed to run?

#

Was it Large AVX2 then Small SSE?

dark ferry
#

Nevermind.

rocky apex
#

so im trying to overclock my ryzen 5 3600 any suggestion for core voltages I did some calibration already

sudden torrent
#

You'd increase all of those when overclocking to 2000 fclk

sterile flame
#

Weird because I have a 1.1 SOC and managed to find 2000 FCLK stable

#

Unsure if adding .05 onto that would really make me go up a whole other level

quick rose
#

if it's stable, leave it

#

You got lucky

errant mirage
#

By what increments should I OC ram? OCing ram on Ryzen is still new to me

proven canopy
#

Looks like 16 gbit rev.b, btw you can right click on the "start benchmark" and test memory only

#

I would start with aiming for 3800 16 18 18 38 1.4v vdimm , 1.15v soc, keeping 1:1 with fclk

torpid siren
#

Got me RAM in the mail!

proven canopy
#

Lol wow, nice kit, what are your plans with it?

#

Do you hwbot?

torpid siren
#

Hwbot?

#

To do with it? Just leave it outside in the melting snow. 😝

#

Oh

proven canopy
#

Those are some high end sticks, was wondering if you bench a lot

torpid siren
#

Opposite.

#

Just would like better RAM, at least more logical for gaming.

sterile flame
torpid siren
#

Once again, If you're wondering, 4 x 8 GBs, 4000 MHZ, C15, 1.5V

sterile flame
#

And what CPU are you running that on? lol

torpid siren
#

3950x. Am somewhat aware about the f clock won't run on 2000 MHz on Zen 2 unless you increase the voltage.

#

I will be getting 5900x/5950x, if I can find one at MSRP

quick rose
#

Probs won't do it period

#

To be completely honest

torpid siren
#

F clock?

quick rose
#

Yeah on 3000 series

#

Especially the 3950, probs better luck with a 3900

torpid siren
#

I can at least enable xmp

#

And check

#

Kk getting them in now

quick rose
#

Ye

#

Honestly, people are lucky to hit 1900 FCLK with 3000 series so best of luck!!!

torpid siren
#

Kk. Done.

#

All modules appears good from BIOS

#

Can't POST once XMP is enabled.

quick rose
#

At 2000 fclk?

torpid siren
#

Just 1800 for now

quick rose
#

oh wow

#

Won't boot even with that?

torpid siren
#

Let me check

#

Is set to auto

#

Maybe set to 1800 to start?

torpid siren
quick rose
#

Ah, I thought it wouldn't boot at 1800flck

#

that's not normal lol

torpid siren
#

It was set to auto and maybe kept trying 2000 since they are 4000 MHz ram kit

quick rose
#

Yeah, that's it

torpid siren
#

VC?

#

It did POST once I remove XMP

#

@quick rose no? You busy?

quick rose
#

wat?

sterile flame
torpid siren
#

Oh lol

quick rose
#

Yeah, if you were asking about Voice Chat I don't have a mic either lol

torpid siren
#

REEEE

#

I can POST once XMP is off.

quick rose
#

Yeah, like I said, I doubt you'll get the 3900 to 2000 FCLK, even 1900 may be a struggle

#

Even the new 5000 series struggle to hit 2000 FCLK

torpid siren
#

Even forcing the fclk to 1800, no dice

proven canopy
#

Most chips can do 1900, almost none can do more

#

New agesa can help

tall pelican
#

With zen2, see how high IF you can post with auto ram (2133)

torpid siren
#

@proven canopy @tall pelican you guys mind VC?

proven canopy
#

can't at the moment

bright thorn
#

looking to overclock my 9600k what should i use to stress test, and how do i know at what point to increase the voltage, and whats a temp thats to high to that i should reduce voltage and speed

#

ping/dm me with answer

quick rose
bright thorn
#

ok can you at least tell me what to use to stress test

quick rose
torpid siren
bright thorn
quick rose
#

You didn't even read it

bright thorn
#

?

bright thorn
quick rose
bright thorn
#

im trying to see the silicon lottery there

quick rose
#

You're not trying to overclock

bright thorn
#

but there isnt a 9600k

bright thorn
quick rose
#

Then read the Wiki

bright thorn
#

im confused

quick rose
#

It literally answers all the questions you just asked in one place. How and when to change voltages, what to do first, what software to use, etc

torpid siren
#

@quick rose Sorry for the back and forth.

bright thorn
#

still why no 9600k?

quick rose
#

WDYM? Overclocking isn't just plug in numbers and win lol

bright thorn
quick rose
#

IDK why

sterile flame
#

because they likely didn't test it or didn't have a big enough sample size?

proven canopy
#

SL buys chips to bin and resell to enthusiasts with lots of money - 9600k doesn't really fit in there

torpid siren
#

@sudden torrent Welp. I think 3600 at 14-16-16-36 will do for now.

#

Hopefully the full 4000 when I get Zen 3 when I have the chance.

#

Love you all for helping this journey on this.

#

πŸ™‚

hushed charm
#

can anyone help me oc my cpu? πŸ™‚

quick rose
hushed charm
#

yeah im looking over the pins rn

#

do u know how to use gigabyte's ez oc thing?

#

wat are binning stats?

#

should I use the bios or the aorus easytune?

hushed charm
#

I don't understand how the host bclk relates to frequency

clever epoch
#

wouldn't recommend automatic OC tools

#

binning stats are statistical trends for silicon quality. EG 100% of a particular model should be able to reach 4 GHz at 1.2 Volts.
Then only 75% can reach 4.1 GHz at 1.2 Volts.
Then only 10% can reach 4.2 GHz at 1.2 Volts.
etc. etc.

#

the numbers I used are pretty much completely random, but the idea is to figure out how likely it is that your chip can reach a certain speed with a certain amount of voltage, which is a measure of how good your quality is on the chip

#

I would also recommend doing all your OC in BIOS

#

idk much about BCLK either, but I understand it as a sort of multiplier for given core ratios. EG with a multiplier of 40x at a BCLK of 100 MHz, you will run at ||40 * 100 =|| 4000 MHz. If you increase BCLK to 101 MHz at the same (40x) multiplier, you will run at ||40 * 101 =|| 4040 MHz

dark ferry
#

The binning stats page helped me tremendously with finding a starting point.

sterile flame
compact delta
#

should i push my ryzen 7 3700x oc to 4.3MHz at 1350mV or 4.0MHz at 1250Mv

tall pelican
#

1250mv

clever epoch
#

1350 is not safe for daily

native wind
#

whats a good core voltage for R5 5600x to hit 5ghz

tall pelican
#

whatever pbo uses

severe forum
#

lmao

sterile flame
#

Lmao.

#

Thats what I mean.

#

Ryzen master boosts not super hard.

#

and people usually use PBO

upper rock
#

whats the max yall have ever gotten on a 3800x

#

or yall have ever seen at the very least

severe forum
#

no one has a 3800x, sara

#

besides you

sterile flame
#

^

proven canopy
#

mr1111 does

sterile flame
#

I'd assume 3800x is better boost

#

Then 3700x

upper rock
#

only a hundred megahertz or so

#

iirc

tall pelican
#

I have a 3900xt, and stock gave better single core boosting, pbo gives better all core

dull ginkgo
#

@upper rock 4.4 all core max

vernal sail
#

is this puppy enough for some serious core 2 extreme 9770 overclocking

#

like 4ghz+

#

hoping to get 4.2 on a xfx motherboard

#

if I had one would use a rampage 2

clever epoch
#

I know it's the opposite of the channel intent, but I'm looking to underclock my GPU so it's mining more efficiently when I'm not gaming.
Now I've been using MSI Afterburner, but it bottoms out at -502 on my core clock, resulting in a core clock of 1170 MHz on my EVGA 3070 XC3. However, I have a friend with a 3070 TUF OC and he can get his clocks down to ~950.

I'd like to reduce clocks even more, but I'm not sure how I could do that, because afterburner does not allow me to go any lower

#

and I can't seem to figure out how to enable voltage control (it is enabled in settings), otherwise I'd mess with that

sterile flame
#

undervolt

#

do ctrl + f

#

flatten the entire thing then drag it down

clever epoch
#

I see, thanks

tall pelican
sterile flame
#

i have to stay up for another 6 hours waiting for best buy restock

#

want 5900x sadge

tall pelican
#

oh, I was posting it because memory go brrrrrr

#

I've had the cpu for almost 2 months now

sterile flame
#

yeah i was wondering why the memory looked odd

tall pelican
#

4733 on an oem stick LeoKek

sterile flame
#

how low do you think you could get your latency on your good ram

#

at 2133mhz

#

kek

tall pelican
#

no idea

#

I got 48.3ns at 4200c14

sterile flame
#

damn

manic helm
#

rev e on an oem stick?

Thats a waste. how many volts.

tall pelican
#

1.38 since I dont have vtt control

hollow edge
#

Top noob tip I learned recently. Just because you have 3000Mhz ram, doesn't mean it's running at that. Turns out I'd be running it at 2133 the whole time. Oops xD

sudden torrent
#

Common mistake, not turning on XMP

ionic socket
#

I think these are just cheap rams with not that much MHz there not my rams this is from a pc was given to me I’m still building my pc.

#

I was just curious

#

is that why or something else ?

sudden torrent
#

Yeah default voltage for DDR4 is 1.2v, 1.35v is well within tolerance for any die, even the bad bin ones.
I'm assuming that it turned red because it's going up an amount above whatever limit the motherboard maker decided. Nothing to worry about.

proven canopy
#

dram voltage is stored in the spd data, the xmp profile sets it higher

left bladeBOT
#
GOTHBOY#2730 has been warned

Reason: Bad word usage

ionic socket
#

Oh ayeee I was using 2133 as the default as is but I turned the xmp it’s now using the 3200mhz. So I guess this ram isn’t bad

#

I thought this ram was in the 2000s

cunning shard
sudden torrent
#

Profile 2 is usually the higher performing one but it depends on a few factors like SPD timings and speed, and the CPU you're using.
Lacking other information I'd recommend starting with 1, test the speed, and then go to 2 and test the speed. Odds are that 2 will be faster, but not by much.

cunning shard
#

Ihave r5 3600

#

Ill put it on profile 2

#

Thanks

sudden torrent
#

Ryzen has a trick to it. Up to 3600 speed memory should work fine with other auto settings, but any faster and you might need to play with the fclk to keep it stable.

cunning shard
#

Iput my ram to 3200

#

Its enough

sudden torrent
#

That'll be perfect on auto settings

cunning shard
#

do you know diffrence btw cpu temp and cpu temp Tctl ? @sudden torrent

#

it shows defrnt results

sudden torrent
#

Looking at which program?

cunning shard
#

quick cpu gives me 65 and game booster gives me 48

proven canopy
#

Use hwinfo64 or ryzen master

#

I like coretemp

#

Very lightweight

sudden torrent
#

The first is the actual temperature and the second is accounting for an offset

cunning shard
#

aight thanks guys

balmy kelp
#

I guess I did a great job overall

icy crag
#

Hey guys, i have a ryzen 1700x and a b450m aorus and uhmn i tried to overclock it and uhh i did this

#

Cpu clock ratio 40, core performance boost disabled, amd cool&quiet function disabled, svm mode disabled, global c-state control disabled, psu idle control auto, opcache control auto, downcore control auto, smt mode auto.

#

Dynamic vcore(dvid) +0.102v
Dynamic Vcore SOC Auto
DRAM Voltage 1.370

#

CPU:
Frequency 4008.80MhHz
BCLK 100.22MHz
Temperature 16.0Β°C
Voltage 1.416V - 1.428V - 1.440V

tall pelican
#

voltage is a bit on the high side, what's the hwinfo64 svi2 tfn voltage reading?

tall pelican
#

max and mins

#

avg doesnt matter

icy crag
tall pelican
#

then you're fine

icy crag
#

So i dont have to change anything right?

tall pelican
#

nope, you're perfect

icy crag
tall pelican
#

god I wish there was a way to disable reply @ s

finite frigate
#

me too

dull ginkgo
unkempt temple
#

i overclock with liquid nitrogen

#

got 7 ghz

#

feeling good

copper fable
#

Getting a full 30l next week for some fun! Was supposed to be this week....vehicle issues

vernal sail
#

Unpopular opinion, I liked the old blue on grey

#

Where you could overclock without a mouse

forest radish
#

can still do it without a mouse

#

I never touch the mouse in the bios, works great

#

little clunkier than it used to be, you're right

thin rapids
#

Should I run my ram higher or equal speed to what I can set the infinity fabric clock to?

sudden torrent
#

There's a penalty in latency running out of sync from the RAM, so unless you can get the RAM significantly higher than fclk it's best to run equal to fclk

thin rapids
#

so 3866 is not worth if my infinity fabric can only go to 1900

#

i am using the 5600x if that matters

sudden torrent
#

It would still be worth it, you can tighten the latencies a bit more running RAM slower than rated speed. I knocked my CAS down 6 notches just running 3866 instead of 4133

thin rapids
#

Assuming I can't tighten the timings any further. Would the 3866 be worth or should I just run 3800 to match infinity clock

sudden torrent
#

1933 fclk should be possible unless you're just plain unlucky. I say get the 3866 RAM and try to run fclk higher by bumping voltages, but if it's not stable enough you can always run it at 3800. You absolutely can tighten the timings doing that, even if it means just going from 16-19-19-39 to 16-18-18-38 it's worth it.

thin rapids
#

What voltages would I have to change for higher fclk?

dull ginkgo
#

iirc SoC? don't quote me on that though

thin rapids
#

Would HWINFO be the way to check what my voltages are at right now?

proven canopy
#

yes

robust aurora
#

Is intel extreme tuning utility good for overclocking or should i do it in bios?

proven canopy
#

Use motherboard specific software like turbov or eleet if anything. Bios always works as well.

robust aurora
#

Ok thanks

#

Which software is for gigabyte motherboards?

#

Is it easy tune?

upper rock
#

could just use BIOS, when push comes to shove

robust aurora
#

I just never oc my cpu before so i dont want to mess something up in the bios

zenith palm
sudden torrent
robust aurora
#

do you have any good guides for the 10700k?

#

Thanks for helping

thin rapids
#

Would I want to push soc or core voltage or some toher voltage to push my fclk?

sudden torrent
sudden torrent
thin rapids
#

Would 1.18 V on SOC be fine since that is what my pc was running at on auto?

sudden torrent
#

If that's the auto it should be ok but when you've tested if it's stable, see if it's stable at lower voltages; .01 volt steps down, test each step down

tall pelican
#

1.18 is fine, but not because that's auto

sudden torrent
#

Really? I've always been warned about going over 1.15
You're the best OC expert we've got here though so I'm inclined to believe you

tall pelican
#

its fine to 1.25v-ish, but generally over 1.2 you see it getting worse

sudden torrent
#

Well I'm happy that my chip is perfectly stable 1933@1.12v and I can sacrifice a hair of stability to push it to 2000

tall pelican
#

||its probably not, just error correction in the IF doing its thing||

thin rapids
#

Maybe I am just unlucky, but my fclk does not want to go to 1933 at all

sudden torrent
#

Sounds like you might just be unlucky, BIOS is updated? Agesa update 1.2.0.0 helped with stability

thin rapids
#

yeah everything is updated might just run ram at 3800 and tighten some timings some more

#

so far I tuned my timings from 16-18-18-38 to 15-17-15-36 at 3800

sterile flame
#

Can you over clock an i9-7900x

proven canopy
#

yes

sterile flame
#

how

quick rose
#

Read the pinned messages

upper rock
#

Forgot about the Intel i() (....)x CPUs

thin rapids
#

so far I got the timings to 15-17-14-34 at 3800. I think I am happy enough

#

first two timings won't go down any further from what I tested

upper rock
#

Think I could get 3800mhz on B-Die tridentZ RAM with a 3800x with 1:1 FCLK?

sudden torrent
#

It'll be tricky on zen 2, if not impossible depending on binning
You could always ask Fitz but he'll probably just say something like "read the pins"

upper rock
#

Heh

#

Got a 360mm AIO coming for it too

#

So I will be able to keep it cool

sterile flame
#

where do i look to find out how to overclock my intel cpu

tall pelican
upper rock
#

Yeah ik

#

I tried this once before with a R5 3600

#

Didn’t work great

#

Had to reinstall windows

tall pelican
#

1.1-1.2 soc
.9-1.0 vddp
.95-1.1 vddg iod/ccd

sterile flame
#

what link do i click on to learn how to overclock intel cpu’s

upper rock
#

Check pins KEK

sterile flame
#

which link in pins

upper rock
#

Discord pins

#

You know what pins are, right?

sterile flame
#

ya but which link in the pins

#

there are multiple links

#

@upper rock

sudden torrent
upper rock
#

Any of them lol

#

But especially the Reddit one

iron quail
#

I HAVE A GOLDEN 9900K

#

LIKE PERRRRRRRFFFEEECCCT

#

I'm running an all core 5gh at 1.2v

robust aurora
#

Should i try intel auto oc and see what i get before trying it my self in the bios?

vernal sail
#

@robust aurora i just like reading up and seeing what the normal range is

robust aurora
#

ok

#

Will do that

iron quail
#

This is amazing

#

I got 5.1ghz at less than 1.2v right now lol

dark ferry
#

What are you using to test it?

dim plank
#

hello. I would like to ask what does voltage offset specifically do? Does it switch two fixed voltages or does it change voltage dynamically according to the required voltage?

robust aurora
#

hi so i did the auto oc and its says its 5 ghz

#

but when i run cinabench the clock speed are at 4800

sudden torrent
#

Cinebench single or multi core test?
It's pretty normal for it to not hit max boost under heavy, constant load.

robust aurora
#

Multi core

dim plank
#

Hello guys. Does voltage offset change voltage dynamically or does it only change according to the neagtive/positive offset number set?

sudden torrent
# robust aurora Multi core

That's a "worst case scenario" kind of test, every core is being hit 100%. It won't be hitting the max for long, if at all.
Other multi core benchmarks that don't do 100% load like time spy would be good to watch boost behaviors.

sudden torrent
#

Stock voltages do tend to be aggressive and far more than necessary

dim plank
#

@sudden torrent Thanks for the info.

I would also like to ask, Is the dynamic voltage offset set as a range where the voltage can change any value around the original voltage and the set negative/positive offset? Or does it only change into a fixed value set using the negative/positive offset?

I am trying to understand if the dynamic offset is not just a switch between two voltage values. Say orginal voltage and one specific voltage offset value.

sudden torrent
#

Normal motherboard and cpu behavior is the board will read what the cpu wants and is rated for in a given load. Once that value is calculated, the offset is added/subtracted.
My cpu, for example, might hover around 1.05v when not under load, and go up to 1.38v in heavy load. Any value between the two would be valid for the motherboard to set. The offset would change the calculated value, wherever in the stock range it falls.

dim plank
#

I see. So the voltage offset doesn't only set the value typed. But it can also assign a voltage value in between the orginal value and the offset value.

Ex: 1.45v with a negative offset of .5v ; result will not only be 1.4v but can also be with values in between 1.45v - 1.4v?

robust aurora
sudden torrent
#

If 1.45 is the absolute max voltage, the offset -0.05 will make the max 1.4v

sudden torrent
robust aurora
#

so thats too high i guess

dim plank
#

"If 1.45 is the absolute max voltage, the offset -0.05 will make the max 1.4v"

@sudden torrent So no in between right?

#

How can I achieve a votalge with in between values? Not just a fix offset? Is there a possible way?

#

I guess I will just set it to auto?

sudden torrent
#

There's only 3 modes, auto, offset, and override. Override will set a constant voltage.
Auto will do fine. It won't usually ask for its max voltage anyway.

robust aurora
#

i dont understand what is all core die sense vcore, is it core voltage ?

dim plank
#

I see. For some reasons I thought , the voltage offset treats the value as a range but from what I am learning there is only two sets of values to swtich.

sudden torrent
# robust aurora so thats too high i guess

There's no definite answer. If that's what it's stable at and won't work lower, that's the limit of your bin. The silicon lottery page is more of a guideline of what to test.

robust aurora
#

i just did the intel auto oc and that is what i got from it . so now i should try and low the voltage till my pc crash?

sudden torrent
#

Basically, yes. I like to run prime95 and gradually lower the voltage while it's running. Once one or more of the threads crash, increase a hair again and test longer.

robust aurora
#

πŸ‘

robust aurora
#

@sudden torrent i dont understand all the settings here like what is avx2 ratio offset and core voltage offset. can you please explain to me what do they mean

#

and if their values rn are ok

dull ginkgo
robust aurora
#

Now i did all cores x51 and its working on 4900

#

So its always 200 less mhz

#

Might be avx2 ratio offset setting?

#

Its set for me for x2.0

upper rock
#

tbh I cant take you seriously with the gravity falls pig in your pfp

robust aurora
#

πŸ˜‚

dim plank
#

generally what is the safest min cpu core voltage? 1.25v?

hushed charm
#

Can someone help me actually change my frequency? I can't get my bios to let me do anything

upper rock
#

what board, what CPU

hushed charm
#

Aorus gaming 7 z370 i3 8350k

#

F4 bios

upper rock
#

is it possible for you to go to F7 bios?

hushed charm
#

Yeah I can try to update it

#

Gotta look for it tho

dim plank
#

What voltage does some pc usually run in idle? Is it around 1.25v? Currently I am running around 1.35v

hushed charm
#

Wait why f7, there's an f15a available

upper rock
#

F7 is the only one I know for sure works well for overclocking with that board

hushed charm
#

Ok

upper rock
#

wait

#

actually dont do that

#

wait for someone else

#

I gtg

hushed charm
#

Ok? Lol

#

I'm updating to f7...

#

Dang, I thought I was actually gonna get some help

#

I blue screened at 4.5, rip

upper rock
#

So is it working?

hushed charm
#

Yeah, I had to use page up and down to adjust stuff

#

I increased my voltage to 1.3

#

4.5ghz

#

Got into windows!

#

It's working so far let's gooo

upper rock
#

yay

tall pelican
#

what load line are you setting?

robust aurora
#

should i use thermal velocity boost?

hushed charm
#

load line?

#

didn't bluescreen in valorant, should be good i think

#

oh i left load line on auto

#

should i put it on turbo or something?

tall pelican
#

on auto, 1.3v is dropping down to 1.1-1.15ish, high will keep it in the low 1.2s, and should let you clock higher

hushed charm
#

ok sick, thanks, if I stay at 4.5 u think it's fine to keep it on auto?

tall pelican
#

if you want to

gentle perch
#

Hello new to overclocking here

#

I have a 5800x running 4.6 all cores at 1.20625v on ryzen master and have tested it for stability by using cinebench r23 for an hour (two 30 minute tests)

#

My questions are: is that a good overclock or should I go higher? Max temp it reaches is 79 C

#

And also would I see more of a difference if I did the overclocking in the bios instead of applying the overclock on ryzen master as soon as I start up my pc

#

Is there a big difference to overclocking through the app versus doing it through the bios aside from not having to set the overclock manually every time

sudden torrent
# gentle perch My questions are: is that a good overclock or should I go higher? Max temp it re...

A good practice when overclocking Ryzen is to use precision boost overdrive instead of an all core constant overclock. My 5800X will boost to 5 GHz, which will give a good bit of extra performance when gaming since games tend to like faster single cores, and it never breaks 75C (with a decent 240mm liquid cooler).
A good place to look for tips and how-to is the reddit page that's pinned, https://www.reddit.com/r/overclocking/wiki/index

vernal sail
#

Am getting 4.5 on my 5600x

#

Honestly don’t want to push it too far 4.7 at bace volts is unstable and this is kind of mission critical

#

Really fun cpu to overclock tho

#

I mean its no opteron...AMD archthonk

sudden torrent
#

Mission critical stuff you don't want to do much overclocking at all tbh
Just enable PBO in BIOS and call it good.

#

That should let it boost to 4.7 or more

thin rapids
#

Is there any way to check if PBO is enabled outside of bios?

gentle perch
#

I’ll check out pbo @sudden torrent

#

Need to consume more information

zenith palm
#

I only boost to around 3.9 or sometimes peeking 4ghz on a Ryzen 5 3600 does that sound right? It has stock cooler btw which i will be upgrading soon with a id cooling sse 224

sudden torrent
sudden torrent
zenith palm
#

Okay thanks

dim plank
#

Are the last tile renders in Cinebench more heavier compared to the earlier tiles? I noticed the temp going high in the last tiles.

proven canopy
#

Ya, the intensity and difficulty of the workload for each tile varies.

gentle perch
#

so if i pass an hour of r23 i'm stable probably right?

tall pelican
#

until you give it something more stressful than r23, sure

dull ginkgo
#

Kinda want to say p95 time, but maybe occt's cpu test?

gentle perch
#

I’ll try an hour of prime95 later today

dull ginkgo
#

smallfft is harsh

#

very harsh

proven canopy
#

Ya, try occt avx2

sudden torrent
#

Brutal. OCCT showed errors not a single other software could find. It's very stable in everything else, even other stress tests.
Currently I'm at 3866/1933 synced. When I was testing my RAM I was able to push it to 4333 (from XMP 4133) pretty easily, I could go higher loosening timings most likely.
At what point does the latency penalty for having desynced fclk not matter any more? Should I push higher or leave it synced lower?

dull ginkgo
#

OCCT is great for errors, P95 is great for space heater and crashing

sudden torrent
#

Yeah I tested a lot more than just p95 but that was one test I did. I find it better for CPU overclock testing than RAM.

dull ginkgo
#

dram calc can be useful

sudden torrent
#

Problem I'm having with it is it doesn't like to go over 3866, but seems good for below that

dull ginkgo
#

@proven canopy looks like when I hit it with a load, it'll sustain the 2133mhz, both cpuz and aida64 reports the c16

proven canopy
#

good ram I guess

dull ginkgo
#

no load, both reports the c10

proven canopy
#

idk, I'm unfamiliar with the platform

proven canopy
dull ginkgo
#

fair

#

just an interesting phenomenon

sudden torrent
#

Uh oh, is that the membench thing they got in trouble for?

dull ginkgo
#

@proven canopy by the way, do you know if it's possible that this ram is physically switching freq/timings? Not just a reporting error?

proven canopy
#

I don't

sterile flame
#

is MFR M die?

tall pelican
#

Its mfr

#

Mfr = hynix
M die = samsung

robust aurora
#

how much memory clock should i increase when increasing core clock? what is the ratio?

robust aurora
#

is msi afterburner oc scanner good? should i use it

zenith palm
#

Id that makes sense

zenith palm
zenith palm
# robust aurora is msi afterburner oc scanner good? should i use it

Here's a tutorial of sorts https://youtu.be/cqshaJTI8FQ

Play World of Warships for free: http://bit.ly/2ldpyOX New players will receive 1 Million Credits, the USS Langley Aircraft Carrier, 3 days premium time, and more by using my code PLAYLANGLEY2019. Big thanks to Wargaming for sponsoring this video.

β—‹β—‹β—‹β—‹β—‹β—‹ Items featured in this video available at Amazon β—‹β—‹β—‹β—‹β—‹β—‹

β–Ί Amazon US - http://bit.ly/1meybO...

β–Ά Play video
robust aurora
#

Thanks will help alot@zenith palm

zenith palm
#

Np

robust aurora
#

Manged to get 1970mhz avg on timespy now adding some memory clock so far 500 is stable

zenith palm
#

Nice

#

What gpu you running

robust aurora
zenith palm
#

Nice, you could probs go further on clock tbf

robust aurora
#

I know but i dont want to push it too far rn, maybe later.

sterile flame
#

Managed to manually clock em to 3733 however using DRAM calc hasn't given me good results

forest radish
#

probably because it's dram calc

#

also the colloquial naming schemes for memory are dumb, but usually x-die means samsung, three letter codes like MFR mean hynix, rev x means micron, etc

#

ideally we would include the density, like instead of saying b die, you would say S8B

proven canopy
#

I think the naming convention works well

dim plank
#

Why motherboard Vcore always start at 1.45v by default? Is there a way to change the base Vcore so that it will make it easier to use a positive offset?
Using negative offset just makes the idle voltage higher than the load voltage.

tall pelican
#

which cpu? Idle voltage should always be higher than load, but ryzen's idle vs load is more drastically different than intel's

bright thorn
#

What software should I use to stress test CPU and how long should I run it to be effective

dull ginkgo
#

@bright thorn an hour of p95 smallfft for really heavy stress, an hour occt avx2 test for heavy stress

bright thorn
#

What difference between really heavy and not really heavy

dull ginkgo
#

P95 is crazy

#

Occt heavy, not as crazy

#

P95 smallfft made my CPU spike to 180w power draw, occt spikes only around 150w

#

And also, p95 is lots of heat and stress, occt is better for finding errors

bright thorn
#

an hour is alot tho

#

can i do other stuff like watch youtube while its running

dull ginkgo
sudden torrent
#

Oh, the new version (7.0) of hwinfo has tooltips for all the sensors. Nice! Maybe we'll get less "what does X mean" questions.

vapid berry
#

What does x mean

zenith palm
bright thorn
#

i turned on prime95 how exactly do i turn it off

dull ginkgo
#

Test, stop, stop all workers

bright thorn
dull ginkgo
#

it what

bright thorn
#

?

bright thorn
#

well temps are dropping...

#

was at 69c somewhere around that now im dropping so

dull ginkgo
bright thorn
#

i did

dull ginkgo
#

it takes a moment to stop, but should stop in like 10 sec

bright thorn
#

its been longer than that

#

should i restart

dull ginkgo
#

check task manager rq

#

it should've stopped

bright thorn
#

cpu is 100%

#

but it doesnt say what is using it

dull ginkgo
#

task manager should say what is eating what

bright thorn
dull ginkgo
#

click the 100%

bright thorn
#

My screen shots on my pc just dont work

dull ginkgo
#

that'll sort it

bright thorn
#

Oh I see

#

Its there

#

So can I close the task right

dull ginkgo
#

sure

bright thorn
#

Ok cool weird

#

That's was annoying

#

Well the more you know

dull ginkgo
#

ye

#

weird things

sudden torrent
#

Task manager sometimes doesn't account for idle time, it's a known bug, and happens frequently with Oculus headsets. Sometimes it will just report 100% usage constantly. If that's happening to you, run these commands:
PowerCfg /SETACVALUEINDEX SCHEME_CURRENT SUB_PROCESSOR IDLEDISABLE 000
PowerCfg /SETACTIVE SCHEME_CURRENT

#

I had this bug on two separate installations of Windows, and this always works for me.

bright thorn
sudden torrent
#

Well it never hurts to reboot anyway, sometimes Windows is just weird. If you check later and it still says 100% try those commands.

sudden torrent
#

Confirmed, I can get my 5800X to boost past 5GHz on at least 3 cores

novel cobalt
#

I posted all details there so I don't have to spam discord repeatedly

rough cairn
royal quiver
#

How do I turn off the turbo timer

sleek garnet
#

hello just wanted to ask what the software is called that tells me what kind of a sample of my cpu I got. The one for Ryzen

quick rose
#

CPU-Z?

#

That will tell you what CPU you have as well as Motherboard model and also RAM

dull ginkgo
#

I think it's a question more about the sample for silicon lottery

quick rose
#

Well, AFAIK, there is no software to tell that

dull ginkgo
#

Like Clock Tuner for Ryzen

#

CTR

#

the kinda auto-oc thing

#

that kind of tells you how good a chip you have

sleek garnet
#

thank you!

zenith palm
#

You also need ryzen master with it

sleek garnet
#

yes yes already have that

#

thank you:)

dull ginkgo
#

:)

upper rock
#

i have it open rn

dull ginkgo
#

pretty sure it uses RM to edit your clocks

zenith palm
upper rock
#

where does it say how good of a chip you got in ryzen master

dull ginkgo
#

no, CTR (clock tuner for ryzen) uses RM to auto-oc, and by the end, it'll tell you the best OC it can get that's cinebench stable

upper rock
#

oh

#

fml

quick rose
#

What is this black magic

upper rock
#

says the wizard with no cables

quick rose
#

Of course, wiring is my thing

#

You should see the custom wiring I do for all my rigs and bikes, never see it and can't distinguish from factory

dull ginkgo
#

vintage op racer

quick rose
#

Having OCD sucks

quick rose
#

It's official, my 3700X sucks lol

#

CTR says Bronze sample

upper rock
#

rip

quick rose
#

Then again, I could probably tweak some settings in the BIOS

dull ginkgo
#

my 3800x is prob going to be like bottom 1% sample

upper rock
#

3800x is literally just a slightly higher binned 3700x

dull ginkgo
#

not for early sample against current 3700x's

quick rose
#

My 3700x is very early

#

Launch day early

#

NOt surprised it sucks

upper rock
#

kek

#

hey Fal

dull ginkgo
#

Mine is month after launch early

#

prob within weeks

quick rose
#

Ye, bad early bins

#

I might be able to get it better with some LL settings in the BIOS

#

need to play with it

sudden torrent
# rough cairn hey, sorry for a lot of questions in advance. Considering amping my 5600x to 5gh...

First off, not every chip is going to be able to do 5 GHz, it will take a little luck. My chip can on a few cores, the rest get close but not quite.
I used PBO auto voltage but load line calibration level 4. Max offset is +200. Mine can do +150. The rest is just a minor adjustment with curve optimizer, a couple steps on negative but a little more on my "best" cores because they can go a hair higher.
I'm on a MSI X470 gaming plus, latest bios.
Trial and error are going to be needed to see how far your chip can go.

rough cairn
# sudden torrent First off, not every chip is going to be able to do 5 GHz, it will take a little...

Thanks, yeah I've oc'ed my 5600x to 4.85 with no prob, with negative 15 curve optimizer and pbo +200. Also on auto voltage but it never actually goes above 1.3 volts even on high loads. Gives 12230/1603 score in cinebench23.

Haven't touched the load line calibration this time, but used level 3 when had all cores oc'ed to 4.85. Then had it locked at 1.3125 volts and had 12481/1603 score in cinebench23.

I'm on asus tuf x570 3005 bios. Heard a new bios update will allow offsetting the pbo higher than 200 (as some of the msi boards allow). Wondering if anyone managed to oc it to 5ghz but haven't found anyone yet.

bright thorn
#

to update hwinfo do i have to delte it and reinstall the new version?
i have the same question but with occt

sudden torrent
bright thorn
#

would unistalling it and reinstalling the new one work

#

cuz like i dont know what you just said so that'd be easier

sudden torrent
# rough cairn Thanks, yeah I've oc'ed my 5600x to 4.85 with no prob, with negative 15 curve op...

I'm pretty sure I got a really good sample for my chip. If I leave everything at its default I still get 4.75 single core boost, and enabling PBO auto gets me 4.85 without any changes.
Ryzen chips seem to tend towards higher voltages at low load, and lower voltage at high load. That's what LLC is for, it keeps the voltage a little higher under load.
I imagine your CPU likes 1.3 max because it's a 5600X, which is a 5800X with 1-2 failed cores that got cut off or didn't meet boost spec, so it needs a little less power.

sudden torrent
bright thorn
#

ok

#

why cinebench gotta run fo 10min

#

like it just runs the same thing 10 times

sudden torrent
#

For stability test? It needs to run that long to see if it crashes under sustained load

bright thorn
#

5753

#

multi core

sudden torrent
#

Well it's good to get an average score to help account for any background tasks

bright thorn
#

i do have youtube and discord open...

sudden torrent
#

If it runs the same thing 10 times, you'll get a slightly different score each time depending on what's running in the background.
You'll get the most accurate score with as few things running as possible.

bright thorn
#

yeah but i mean its prolly mostly fine that i have youtube open

#

dont wanna just sit here for 10min

#

does it save the score when you exit

sudden torrent
#

If a youtube video is running it will use probably 5-10% of your processor time, reducing your score.
I think it saves? Not completely sure, I just screenshot it every time I want to keep a result.

bright thorn
#

Yeah ok next time I'll run it with nothing

random cave
#

i bought a 2x8gb kit of 3200mhz cl16 corsair vengeance LPX memory. i currently have XMP enabled, would it be risky to go up to 3600mhz?

tall pelican
#

you generally cant just increase frequency and call it a day

upper rock
random cave
#

r5 5600x

#

oh ok

bright thorn
dull ginkgo
#

LPX dies be bad

upper rock
#

corsair RAM has horrible die

bright thorn
#

I have lpxNM_PensiveSad

random cave
#

oh kk

dull ginkgo
#

generally bargain bin

bright thorn
#

Buts it's fine cuz its 3200 and I have intel so no need to oc

tall pelican
bright thorn
#

?

tall pelican
#

lpx = heatspreader

#

nothing about the die is related to the heatspreader

bright thorn
#

Could you take the heatspreader off

#

Like Is that possible

upper rock
#

yeah but why would you?

tall pelican
#

yeah, why would it not be possible?

bright thorn
proven canopy
#

It can be pretty hard, it'll rip the chips off the pcb if you're too rough

tall pelican
bright thorn
upper rock
#

thats...

upper rock
#

not how that works

bright thorn
#

Why

proven canopy
#

?

upper rock
#

well

#

its possible

#

but useless

tall pelican
dull ginkgo
#

@bright thorn it's the die inside that's bad

bright thorn
dull ginkgo
#

lpx is their budget ram

#

they don't put anything good in lpx :P

proven canopy
tall pelican
#

Mr111, the heatspreader has nothing to do with the ram inside

bright thorn
proven canopy
dull ginkgo
#

@proven canopy I think so, yes

bright thorn
tall pelican
dull ginkgo
#

but that's considering common sense

#

I don't think anyone will buy expensive lpx

#

:P

bright thorn
#

Does crucial use samsung dies

dull ginkgo
#

@tall pelican also, I thought this was about 3200c16 lpx?

tall pelican
#

LPX IS JUST THE HEATSPREADER, IT HAS NOTHING TO DO WITH THE ACTUAL MEMORY

random cave
proven canopy
#

Is lpx usually a lower bin or lower end IC? sure - but don't make that blanket statement

tall pelican
#

4.32 = cdie, so dont bother trying to overclock it

random cave
#

ok

bright thorn
#

Does crucial use Samsung dies

dull ginkgo
#

what

#

Micron owns Crucial

bright thorn
#

So thats a no

upper rock
#

kek

dull ginkgo
#

Crucial uses Micron dies

proven canopy
#

Modern crucial memory is almost always micron inside

bright thorn
#

Ok

#

Who does use samsung

dull ginkgo
#

I mean, who buys high end lpx kits :(

proven canopy
#

fitz

dull ginkgo
#

damn it fitz

proven canopy
#

189 for that kit isn't bad at all

dull ginkgo
#

For the kit it's not bad

bright thorn
dull ginkgo
#

Yes, just because they exist doesn't mean you should buy them

bright thorn
#

Who said I would

dull ginkgo
#

read the message you replied to

#

:P

bright thorn
#

I guess I shouldn'y have replied to that then

dull ginkgo
#

there was like a 5,000$ 5400mhz ballistix max kit on amazon at one point

upper rock
#

good luck sustaining that

bright thorn
dull ginkgo
#

tho it seems more like a bryce thing

#

just splurging on hardware

tall pelican
#

oi, I buy good ram, not stupid ram

proven canopy
#

You don't daily 5400mhz in your gamer rig? might as well buy an xbox

dull ginkgo
#

lol

tall pelican
#

bryce buys good ram to run it at xmp

upper rock
#

kek

dull ginkgo
#

@proven canopy what's the point where running decouppled starts giving you benefits in benches?

#

on zen 2/3

tall pelican
#

only on a few benches, and you need like 4500+

#

the IF is still the limiting factor

dull ginkgo
#

That sounds about right

proven canopy
dull ginkgo
#

That's what bryce will buy

proven canopy
#

I should do the same, lol

#

I just put in a lazy 4300c16 for 3d benching

#

Still 1 rank/channel so not ideal for time spy

sterile flame
#

For someone who’s never gotten into overclocking before. How do I get started on overclocking cpu?

dull ginkgo
#

Forky, check the pin emoji time

sudden torrent
#

What would you guys call a good, safe SoC voltage max on my 5800X?
I've got my RAM stable 4000@16-18-18-38 and I'm trying to cut down one whea errors. 1.15v isn't cutting it.

proven canopy
#

1.2v, but that can introduce it's own source of instability

#

My guess is that >1900 is just not gonna happen

#

Or you'll see performance regression

sudden torrent
#

Noted. I'll go a little at a time and see what I get. Then I'll save my settings and try lower clocks, maybe I can get 3800c14 at 1.5v (b-die)

proven canopy
#

How's the airflow around the b-die? I'd point a fan at it around that voltage

sudden torrent
#

Good point. I'll run the closest fan higher rpm. My radiator fans directly above it are already running fast because they're rattling at low speed.

proven canopy
#

Some sticks have temp sensors, check hwinfo

sudden torrent
#

Mine don't

#

Also, did you see I got my processor to boost over 5GHz? Must have a good sample

tall pelican
#

this is what happens when you think you have anything higher than 1900 stable

dull ginkgo
#

3866 WHEA errors eh

sudden torrent
#

Yeah when I was running at 3866 I was getting a few whea but it was still stable for gaming and benchmarks

#

I dropped down to 3800. 1.18v SoC I got even more errors

#

0 errors now on 3800

sudden torrent
#

Looks like my final result tuning is 3800, 14-16-16-36, 1.5v
In quite happy with these results. I've got a mini fan I used to use for a router that has a 3 pin plug I can use for the ram.

proven canopy
#

Surprised you can't run straight 14's with 28 tras with 1.5v

#

Have you tuned secondaries?

dull ginkgo
#

28 ras? Really?

sudden torrent
#

I'll give it a go but I don't expect much more from this set tbh

#

It refuses to boot under 1.5v so I'm already on the edge of stability

proven canopy
#

It's b-die, right?

sudden torrent
#

Yes, according to 2 different utilities

proven canopy
#

14 16 16 36 is what I'd expect out of rev.e or something

#

Probably higher rp/rcd, but still

sudden torrent
#

Umm... Running the asrock utility gave me a blue screen with some asrock driver as the cause. I'm on MSI

proven canopy
#

weird, that software works on pretty much any platform - never used it on amd myself though

sudden torrent
#

Asrdrv101.sys def not working on my computer

proven canopy
#

Idk why I thought the asrock tool would work on amd lol

#

But it's the goto for everyone on anything modern-ish intel

sudden torrent
#

I haven't tuned my subtimings since that's about the point it decides to make me clear CMOS

sterile flame
#

Should I stick to POB for the 3600 πŸ˜…

zenith palm
sterile flame
#

Oh its pbo lool oops

tall pelican
#

on a 3600, you have to test pbo vs a manual overclock (manual oc at 1.25v)

sterile flame
#

Im not experienced in manual overclocking πŸ˜…

tall pelican
#

you set vcore to 1.25v, and increase core ratio until its unstable in stress tests

sterile flame
#

Sounds easier than it looks

#

πŸ₯²

tall pelican
#

then overclocking might not be for you πŸ€·β€β™‚οΈ

zenith palm
sterile flame
zenith palm
#

Yeah i think mine was like 1st month after launch so probs not great bin

proven canopy
#

11900k benching on cascade phase change cooling

bleak sable
#

for BTC mining i have more power limit at 60% and mem clock at max, anything else i should change

#

its stable

tall pelican
#

you probably arent bitcoin mining on a gpu

sterile flame
bleak sable
#

I might use firestorm to OC

#

i can have a max mem of +3000

#

and true voltage control not a power limit

sudden torrent
#

You'll need a modified vBIOS for voltage control if your card doesn't support it out of the box...

#

Unless it's doing pseudo-voltage control adjusting power on the fly

bleak sable
#

zotac is weird i have it just only on their software

sterile flame
#

Holy

zenith palm
#

damm cl16?

sudden torrent
#

What do you mean?

zenith palm
#

is it cl16 or like cl20 or something to get 3800mhz stable?

sterile flame
#

3600cl18

#

Lol

zenith palm
#

ahh

sudden torrent
#

I have my RAM and fclk stable at 3800c14 lol

sterile flame
#

I lost my settings lol

#

Really should of left it

#

It was soc 1.1

#

Cldo 0.950

#

3800 1900

#

Vddg and iod 1.05

#

Booted a few times, now wont lol

sudden torrent
#

It took 1.5v to get it stable but B-Die can handle it

faint tangle
#

Just came here to ask a question, sorry for interrupting: I have a 3600 and 3060 ti on a 1080p monitor, bottlenecks 17%. is it worth it overclocking the 3600 a bit?

#

my mobo only has a 105w power delivery

sudden torrent
#

That bottleneck checker thing is garbage

sterile flame
#

^

tall pelican
faint tangle
#

I checked multiple sites

sudden torrent
#

Does not reflect the real world at all

faint tangle
#

all said around 15-20%

sterile flame
#

You will always have a bottleneck lol

faint tangle
#

17% is a bit much tho

sterile flame
#

Never have 0 bottleneck

quick rose
#

Bottleneck calcs are garbage

sterile flame
#

And where did you get the info from?

sudden torrent
quick rose
#

100% of the time

sterile flame
#

Userbenchmark?

faint tangle
#

nah

#

i just googled "3600 3060 ti bottleneck" and checked a couple sites

tall pelican
#

Trfc on bdie is hands down more important than primaries

sterile flame
#

Bruh

faint tangle
#

that used real benchmarks

sterile flame
#

Lol

quick rose
#

3600 just fine for 3060ti.

sterile flame
#

It can handle a 3070 quite fine

faint tangle
#

even at 1080p?

quick rose
#

There are no benchmarks that calculate bottlenecks.

#

Only fps.

#

Reliable ones, anyway

#

Bottleneck ia such a trash term

faint tangle
#

the one im trusting the most checked 3060ti fps with a 10700f, and a 3600

sudden torrent
#

What's getting used more between CPU, GPU, and RAM depends on what you're playing more than what resolution you're playing

faint tangle
#

lost about 20fps in most games

#

cold war, warzone, minecraft still a good bit

#

gta

quick rose
#

10700 vs 3600 is bad comparison