#overclocking
1 messages · Page 31 of 1
Ocbase is the home of OCCT, the most popular all-in-one stability / stress testing / benchmarking / monitoring tool available for PC
GIMPS has free software available for Windows, Linux, FreeBSD, Mac OSX. Contribute to the effort by using your computer's spare processing power.
Prime95 small fft is attempted CPU homicide so be careful lol
P95 is known to actively kill Intel CPUs now lol
Thanks to the whole "unlimited power" debacle
Just gotta set proper current limits
i dont think current limiting cpu really helps, it still blasts current
only thing u can do is like downclock cpu so current low and then do ur mem stress tests and whatnot
pretty sure the CPUs weren't ignoring mobo current limits, just the mobo current limits defaulted to really high?
afaik, limits didnt really help
i love writing a paragraph only for the s word to get it deleted
The "new" Intel defaults might not be low enough but afaik CPU still follows the mobo limits and you can just limit it below the defaults
Although have mobo manufacturers put into place the leaked lowered defaults yet?
bios says its vcore is proportionate to the amount of load its using, so it seems like it is increasing in voltage as the load rises
There's 2 main profiles, Intel Baseline and Intel Extreme. Most boards are now defaulting to Extreme, which is a 250W short limit iirc.
iirc it's a 2600G
I'm aware
oh sorry I'm jumping into the wrong conversation lol
excuse me let me just plow in no big deal
okay I'm on the same page now derp. iirc just run SSE instruction for Intel cuz ya Prime95 lol
The way that conversation went was bouncing off how you said P95 was CPU homicide, and I remarked how it was 100% that on high end Intel, leading to the other commentary
actually i might just do a light oc since I don't have my tower cooler yet so
You should def run AVX512 frfr ||(it's not supported)||
can't you run AVX2 or something tho? I've always hated Prime95
Yes the older AVX instruction sets work
I don't actually want to know my OC isn't stable I just want it to lie to me lol
jk jk
by lvl 3 or 2 you mean mode right?
Yes
okay all set
Same thing
I wasn't sure if 1 was the largest or the highest droop so I went with middle
weird though i don't see any description on em
which board is this? I'll do some homework real quick
cpu voltage decreases proportionally according to cou load
ah my bad, so it's lower v when higher load
right but iirc LLC is setting a certain resistance (ohm thing)
based on the resistance set, the voltage will either drop more or less
Auto should set AMD's limits which I'd follow unless you want to set them even lower lol
just following what you guys said
cpu should be running normally more or less during idle right, then it starts doing the thing when under load
MSI so higher modes = less droop
Set it too high and I've seen it do inverse droop, higher voltage under load than idle 💀
Yay bugs
usually MSI has a picture that looks something like this
i like that band
telling you how voltage behaves relative to the red line (flat one)
I shouldn't be too aggressive on this then so perhaps mode 4 or 5
I mean, if you really wanna know, set the Core multiplier to like 10x or something really low. Set vCore to like 1.1v then go to windows, open HWiNFO then run something to load the CPU
see how far from 1.1v it droops
write it down or something then go to BIOS again & change LLC by like 1-2 modes & do it again
For reference when I had my MSI board I never went above mode 4 and I was getting some of the top scores when overclocking
i actually can't set the multiplier for some reason, it's locked as auto
so you can see which mode keeps voltage higher or lower
hm so i'm getting that mode 2-3 is a lighter oc
lol my Taichi has 3 LLCs & they're like 70 degrees, 45 degrees or flat
For a daily OC I do mode 3
gonna be running occt
@sudden torrent knows better than I do here. I've only had MSI boards for Intel so I dunno how they behave with AMD generally
my cinebench score was like 160 and it's the highest temp i ever seen on my cpu lmao
well then you need to drop some heckin voltage lol
Have you enabled "advanced" overclocking mode? It should be at the top of the overclocking menu
well that was base default, have yet to try it on this mode
yup
wow occt ui looks neat
Hmm. Then it should let you adjust the multipliers unless you have A320, A520, or A620.
Which multiplier exactly was it not letting you change
strange I can't find a BIOS manual... usually their help strings are decent
79 is 100% in spec so you should be good to go higher
Still safe
damn really lol
They're made to run at tjmax 24/7
really is haha, quite daunting at first for me
everything has normalized it seems 81 at 5 mins
System fans caught up then, good
so if it's all fine i should be going for mode 4 now(?)
Only if you get stability issues when you push clocks higher
oh so since it's stable so 2?
that's cool, where do you get that tested?
alr i'm booting into bios
I can't change any values here
cpu ratio just set at auto regardless if advanced or ez
Dang I'm going through my old 2700 submissions and someone went and dropped me to second since I posted the scores
guess that's what you'll be doing later then lol
nice
Only nice in that it incentivized me to upgrade to AM5
any thoughts on the pics? if the boost i have now is all fine I might stick with it unless i can tweak some more while I'm in bios
Not so nice that a critical bios update decided to shove 1.5v into my cpu and kill everything
i was tempted to upgrade to am5 too with ddr5 and all that from an a320 but everything got too expensive
On CPU ratio try setting it to 35x
That will lock it to 3500 MHz.
CPU Core Voltage set to manual, 1.35v
Once you see if that boots then you run a stress test and watch temps again
cpu ratio is locked, really dunno why
it's in 1st image
there's this one though
Core control would be disabling cores or threads, don't touch that one
But you can use the AMD overclocking menu to set the ratio even if the main overclock menu won't let you
as for cpu core voltage it's at auto at 1.356-1.359v but I'll set it to what you said
OH WAIT
YOU CAN TYPE IN THE AUTO
lmao i feel like one of those who had to call cs bc my laptop wouldn't open (opened backwards)
okay my bad
Nah not nearly that bad
yea just funny
I've seen people call my company because their monitor wouldn't turn off. I unplugged it and plugged it back in. That'll be $150.
done 35 on cpu ratio as well as 1.3500
it used to be 3600 freq and slightly higher voltage though
what's the default clock speed?
what i just said
so you lowered it? lol
I was running 3523 during all core occt so that's our starting point
It was also at 1.31v during that same test
ah so you're referring to the test?
yep right here
well I'm sure it was because it was on auto so this current manual setting should be a good start
Odds are you can get it to 3800 locked on all cores or higher
Yep I like to start my manual tuning where the auto settings take it
is 2600G like 5600G with monolithic die & what not?
i didn't touch pss support btw idk what that is
It's a 2400G, Zen 1 which was always monolithic
cuz 5600G ran pretty high memory frequencies with synced FCLK
FCLK on zen 1 you were lucky to get 1600, same on zen+.
And you don't have the option to desync
btw since cll lowers my v when i run higher loads how does that help with my oc?
I was gonna give it to my dad like this but ended up swapping the 2x8gb for some 2x32gb that was stinky MJR
or is that just like a failsafe before i started changing voltage and frequency manually?
Normally when you put the CPU under load, the voltage drops pretty significantly. Having a higher mode than the default (mode 1) means less voltage drop. Higher voltage under load means more stability under load at higher clock speeds.
and you don't want flat cuz then transient spikes can damage CPU or something iirc
With mode 3 I'd expect the set 1.35v to droop to about 1.31-1.32v under full load
ohh i thought about it completely differently, thanks. didn't know it would drop voltages by default
It can be confusing, since higher number means lower number lol
yeah and iirc some mobos do it opposite
more droop is safer & better for thermals but harder to stabilize unless you set some stupid high vCore to begin with--which is also not a good idea lol
Buildzoid joins to define load-line calibration in motherboard BIOS, then talks about how to ensure you don't kill your CPU with LLC.
Ad: EVGA 1080 Ti SC2 w/ Destiny 2 bundle https://goo.gl/NuWWGU (on Amazon)
We have a new GN store: https://store.gamersnexus.net/
Find Buildzoid's channel here: https://www.youtube.com/channel/UCrwObTfqv8u1KO7Fg...
what a nice thing llc is
lol damn this video is so old it has a 1080ti commercial in it from EVGA
I'll wait till 5 minutes then bump my voltage by what amount? as well as frequency multiplier
eh even .1 is a pretty strong increase
Great that makes it even more relevant to this because their recommended voltages should be about right lol
it's avging 2830mhz now after 5min at 1.34v, pretty fine temps
1.3 to 1.4 may not sound like much but lol
2830? didn't you set the multiplier manually?
yup i remember reading through just bumping by extremely small amounts like 0.025
Increase frequency until it's unstable, then increase voltage by one bump and retest. Don't go above 1.4v core, especially not without a good cooler.
which OCCT test are you using btw
well i did, and initially it was moving between 2900-3200mhz
cpu only, default settings
If you highlight the voltage and press the "+" key it'll go up one step
wow if this is cpu ocing I don't know how long ram ocing would take to learn and test
Honestly, with your memory kit and CPU there's not a whole lot you can do
SCLs and tRFC maybe
pretty much yeah, i wouldn't be expecting much either as long as i have a bit if boost
do CPU>Mode: Extreme>Load Type: Steady>Start at Cycle: 3> Instruction Set: AVX512 or AVX2>Thread Count: what it puts for you (number of cores x2 cuz SMT)
that'll stress it pretty heckin hard
bumped freq +25 so 3525. or is that too low rn
AVX2 since 512 is only for Ryzen 7000
oh right. Wasn't sure if 2600G has AVX512... none of Zen 3 does? never noticed lol
You can go in 100MHz steps until you find a limit, then step it down until it's stable
I'll set it to 3650
Not surprising since there's not any REAL app that uses it except stress tests and PS3 emulation
lol ycruncher... that's all I know
dude my G7400 has AVX512. I never knew until like a week ago
I've always liked the stock ryzen heatsink for what its able to handle, and the looks too
It did originally, unless you got a later model or updated your bios in which case it's disabled by hardware
works on Unify-X lol
Unify-X sets microcode for non-K OC among others
it has "AVX512 Trial Mode" to see if it has AVX512 even lol
this is like one of my best scores cuz of AVX512 lol https://hwbot.org/submission/5591888_ggood003_y_cruncher___pi_1b_pentium_gold_g7400_57sec_686ms
I'll keep stress settings at default?
Extreme, steady, start cycle 3
btw I found a gem on archive.org
There's lots of gems there, you'll need to be specific lol
too bad none of the 3d benchmarks that run best on Win8.1 don't have globals. sad https://archive.org/details/nvcleanstall-474.30-win-8-signed
how are temps?
99C??
this would be overclocking and undervolting right basically for lower temps, higher speeds?
ya like the goal is to get highest frequency at lowest voltage
cool cool
while staying within safe temps tho
which would be like 90?
hiting 99C is counterproductive because the CPU is going to slow down to protect itself
eh 85C is probably safer long term
never hit 85 for cpu while gaming tbh
what instruction set are you using for OCCT test?
instruction set is auto but it was on extreme, steady, and 3rd cycle
hella hard on my cpu
ig auto just bounces around with different instruction sets but I'm not sure
I'm setting volts to 1.3450 from 1.35
but ya, I'd try to get voltage down
i thought you wanted me to lower by that lol
(I'll leave now haha)
ah okay that's fine
seems pretty low for a clock boost
alright I'll keep as is
if temps run fine I'll try 3700mhz
may or may not stop at that lol
ya try to see where voltage doesn't lead to crazy high temps then increase frequency until it won't run anymore
then lower frequency back down a bit & retest
yeah pretty much
3650mhz 94°
great so far
I'm defo asking much for the stock cooler but I'll have a new tower cooler in a few days
save those BIOS settings or if there's a profile in Ryzen Master or something
sorry I haven't used Ryzen Master in a very long time so I don't remember how but iirc you can save the profile
which cooler are you getting? are you getting thermal paste too or using the one that comes with it?
one with it
generic
but my friend has a big syringe of it and I'm mostly sure it's also generic
btw 3700mhz 1.3250v
nice nice
see if it'll run with voltage dropped one more step down
okay on it
oh I meant which CPU tower cooler
ah my bad it's the idcooler se 214 xt (think title is right)
just a budget single tower, GN also recommended it as the value cooler
i was gonna go for the pretty decent one which was thermalright something something
for twice the price
figured i wouldn't need to invest much for a more expensive cooler as I'll only be getting a 5600 and be using that for at least 4-6 years
if you could get 5700X3D or 5800X3D, you'd be set for awhile
yeah very possible, but as a student with not a stable income 5600 would suffice
how much is 5600?
6-7k php online
5700x3d is $233 on newegg but it's out of stock
oh heck I'd have to convert that hold on lol
i bought the 2400g at 7-8k at a physical store during the pandemic 
ah okay so it's like $100 cheaper nvm that's a lot lol
yeah exactly
honestly more happy with my ram, the hardest process i do is video editing so if i can work editing with a 2400g 1050ti, should be a big marginal difference to now
$192
nice price
7600x $171 and 12700kf is around the same price too
what's ram frequency? ram overclock is even more fun
96° max 1.32v 3700mhz stable
12600kf $140
RUAAHHHH
I'll do more research on that in the future lol
Are you replacing mobo or nah
i was boutta cop a550 that's how tight my budget was
gonna try 3750 at same voltages
this is with stock cooler, I'm pretty impressed myself
DDR4 overclocking bible lol https://github.com/integralfx/MemTestHelper/blob/oc-guide/DDR4 OC Guide.md
thanks thanks
gonna go for a jog if 3.8ghz fails
3800 freq is same as 3.8ghz right?
yes
it's just metric measurements
Tera Giga Mega Kilo Hecto Deca iirc
Tera Giga Mega are the important ones
i mightve won the silicon lottery w this
1st time oc'er it feels like it
lol
wait i might actually have a pretty stable fps now with esports games
i just want above refresh rate
I'd still try to drop voltage if possible. 98C is hot. Don't forget, water boils at 100C lol
so it's either i boost clocks or lower voltage for temps
yes I'm happy with 3.8
i keep forgetting it's a stock cooler lmao
very happy though i had a good time doing this
what can i expect to see if voltage is too low for my current clock?
if it's a little too low, the stress test will stop
if it's a lot too low, computer will BSOD
BSOD will probably say WATCHDOG TIMEOUT
what temps should be fine? 90° under load?
or Kernel something
I'm at 1.3v still at 3.8ghz so far max is 91
ehh I'd aim more for like 80-85C but for now ig aim for 85C or so then see if temps improve with the new cooler
ah gotcha
would it be okay if it's 90 something? i wouldn't really be running anything remotely close to that intensive
tbh I don't know 2600G well enough to say yes
well run valorant with hwinfo open in the background & see how temps look after like 20-30 minutes
wouldn'tve thought of that, good idea
stress tests are good for stability but often don't tell you much about real world performance
like how often are you running every core at 100% like that otherwise? lol
it's also good to test with things you do all the time to see how temps, etc. look when other components are in play as well. For example, will the CPU end up hotter because of the heat coming from the GPU at the same time
wow people can run 3.8ghz at 1.25v i gotta humble myself loool
lol well you haven't found out that you can't yet so stay positive
for these cases does it mostly come down to silicon luck at that point?
mobo vrms so and so
keep in mind, ambient air temps matter a lot too. If you're in the Arctic & outside temps are -40F, it's a lot easier to cool than somewhere it's 80F of course
silicon lottery variation can be pretty crazy
m that's p interesting
AMD or Intel only care that the CPU can hit its specs
they don't care if it barely can or it can go way past them
it's yes or no lol
1.26v 3.8ghz, if temps are below 90 I'll stop for now and get ready for some meeting
ya & try using the computer the way you usually would & just leave hwinfo open so you can see how temps look
thanks a lot ggood and @sudden torrent really learned a lot
later after the meeting I mean lol
I'll actually leave it running while i get ready
yeah for sure
glad to help. Sorry our schedules are so out of sync. Ping me whenever & I'll get back to you when I can
gotcha
good luck lol
lmk then haha
If I start overclocking again should I update my bios?
7950x3D and same set of Hynix die ram.
oh g0d you are giving me flashbacks with kernel....
Not necessarily but you should update your chipset drivers.
This latest version helped a lot with the multi CCD stutter issues in games
Oh okay.
Integrals guide is honestly pretty bad and outdated at this point with a bunch of meme rules and 💩
Eden’s guide is a lot better imo
@modern walrus how u doing
huzzah
3.9ghz 1.27v
going to test it in a bit
idle temps 30-40 c
heck I'll check it out
hanging in there lol
how's it going?
nice cooler
i figured i'd stay 3.9Ghz clock @ 1.30v
i really don't know if 100 more mhz is gonna have much difference and if it's worth pushing it to slightly higher voltages
the cooling is pretty much not a problem anymore
OCCT: Extreme, Stable, 3rd cycle 70c under load
what do u think tho
try running programs, games, etc. you use often & see how it does thermally. Also, did you try either further dropping voltage or see if 4.0ghz can run?
i have not, no
i only finalized on hitting 3.9Ghz while upping voltages until stable
did 1.27v initially, but now at 1.3v stable
oh so you got it down to 1.27? how do thermals look?
ah okay so you can't do lower than 1.3
well just keep in mind, lower temps should help with longevity of the CPU.
that's why I suggested running games, etc. you usually play. OCCT won't tell you if you're gonna be hitting high temps for daily use
i'm running firefox discord and valorant on high
it's the most taxing thing i might realistically do
how do temps look doing that?
164 fps capped and constant which is nice with the settings
let me run it for a bit
56C maximum lol
oh ya that's fine nvm lol
flabbergasted rn
hovers around 54-56 with 35-40 idle
so should i take it as a sign to boost to 4.0Ghz
that 4.0 sounds great from a 2400g ngl
and judging from what i'm seeing, the cpu doesn't use all 4.0Ghz when it's not in use, only it does when i run a stress test
i usually see it idle at 300mhz or smth
yeah i think i get it, it basically is like - core usage 20%; so it uses 800mhz is that correct?
@steady lance how’s the OC doin nowadays
@modern walrus do I go back to OCing my 4090?
I’m bored and have a spare 1200W PSU to power it
That's never a bad way to pass some time
sounds like a good reason lol
do itt
lol 4400c15 1T rn at 1.7v DRAM
ah DDR4 I missed you
When in doubt, 12V
That sucks
I'm dropping to 1.65v to try again. I think 1.7 was probably too much anyway
it ran it when I was at 2T
heck I should loosen timings, get 1T running then tighten
trying flat 16s at Real 1N
10900K is meh
this one sucks
I think raising SA & IO might've upset it
Kinda regret getting rid of my 10700k
the problem with 10700K for me is it's pretty worthless on hwbot
Then again, it’s the reason I got the 13900k
I'm so over 13th gen
I mean, it's still the best CPU I have but the inconsistency is maddening
hopefully yours is more predictable if it's like later batch or something
but I dunno anything about the batches so nvm
Oh it ain’t predictable at all IMC wise
I got 13900KS the day it came out & it's worse than many non KSs I see on hwbot apparently
One day I’m able to run 8000 MT/s without any issues at all, next day I can’t even post
that's the inconsistency
like it's fun to bench but trying to get daily stable seems impossible
so much so I didn't even really try
As a CPU mine has been pretty good but no idea where its limit is because never delidded it yet
it's like, it'll bench 8600 sometimes & then 8k stable seems impossible
Daily stable for me is at 7800
which board?
Z790 Aorus Master
CTLvRefUp helps iirc
Don’t have it on my board iirc
All I have are RTTs
Don’t got skews iirc, it’s been a minute since I’ve been in the bios
Yeah I’ve figured out the limits of my setup at this point, put at least 400 hours into messing with it
sounds like my 12900K experience
7800 is the daily stable limit, 8200-8400 for benching
that's the last time I put full effort into daily stable for post-11th gen Intel
7800 isn't bad I'd be happy with that
sure starting with an 8 is cool & all
Yeah for a 2DPC board I’m really happy with it
but difference is marginal I imagine
ya that's not bad at all
I was having a heck of a time trying to get even 7200 working on 1DPC Unify-X before I gave up
VT3 is no fun
It’s not fun at all but once it passes it’s so satisfying
I hate it when it makes it through like 4 hours then derps
it's like nooooooooo
heck I just realized I'm on old BIOS
2022 bios lol
I’m afraid to update bios in case it ruins my stability
oh ya that part sucks too
I used 37 different BIOS versions on Z790 Apex
I'm so sick of throwing spaghetti lol
I just settled on the newest one at the time and then stuck with it
Apex had tons of betas
I think after like the 7-8th one there was little to no change frankly
1.8 is the highest you'll go without changing VPP, which can make a lot of other things unstable. 1.7 is just a "safe" zone to account for droop.
Currently just running XMP and PBO. I found out that my bios was bugging out and wasn't always applying the proper settings so I cleared the cmos and went to default.
Once I have some time probably this weekend I will update my bios and every thing then try ocing again.
sounds not worth the mission lol
There's definitely very little gain for a lot of work at that point
Only worth it if you're going for LN2 records really
10900K & DDR4
my waterblock is heckin trash
B-Die?
yes
I'm at 4400c15-16-16-36 presently but my core temps are awful
Thought you were shoving 1.7v into your 7900x3D
I'm gonna put another waterblock on cuz this thing is awful
heckin EK
I even cleaned it really well
but the frame on it is slightly bent so it's probably not sitting evenly
EK really has gone down in quality, even before the big controversy
but it only has to cover the die because the 10900K is delidded & using the direct die frame
Exactly
it's an old EK Velocity block
Their mounting system is meh
lol I bought the direct die for 12/13/14th
after that disaster, I'll never buy another EK product again
Freaking convex plate for a DD kit
well that disaster coupled with two waterblocks with internal leaks or the brand new pump that lasted ~8 months (they replaced it but that doesn't un-melt the tubing mess it caused)
Whose idea was that
I sanded it
& it was about as good as the Heatkiller IV WITH THE DAMN IHS ON
the LF2 cooled better on this 10900K wth
ive had lots of success with corsairs newer blocks
ive used i think xc5, xc7 and the newer xc7 as well
cooled 7980xe, 12900k, 13900k
@modern walrus final is 3950Mhz 1.35V OC
i didn't like going over 1.35v after 4Ghz
70c max stress test
60c max under gaming load
prolly not needing to oc even more, i find no better reason to
except for ram i might but my mobo only supports 3200mhz
I have two XC7s I think but they're the older ones. They're not bad. I dunno why but this Thermaltake W7+ is killing it rn on direct die 10900K. 342w ycruncher only hit 71C
always awesome when you finally get a good mount
jesus christ.
How much wattage is going through that cpu???
Doesn't that cpu have a stock of like 150w?
Wait why am I even surprised... Intel cpus have been power hungry for generations at this point..
.
How do I update my chipset drivers? It's through AMD Adrenaline I think but I don't see a option to install/update my drivers.
Or is the AMD Software Adrenaline Edition the same as Chipset Drivers? Because it did give me the option to update that and I did.
I have noticed that AMD seems to be bundling the chipset drivers with the GPU drivers, but here's the standalone package if you want it. https://www.amd.com/en/support/downloads/drivers.html/chipsets/am5/b650.html
Thank you.
Also I am slightly confused as it says B650?
Aren't chipset drivers for the CPU?
Read back what you wrote here
Chipset drivers... are for the chipset
Which interacts with the CPU
🤦♂️ gosh I am out of touch.
I have a MSI X670-P board so I assume I should find the chipset drivers for a X670?
It's the same driver
Chipset drivers, Management Engine, Serial IO, GNA Scoring Accelerator, I really have no clue what any of those do but I install Chipset & ME lol
Really? Why do they even separate it then?
ME being an Intel only thing, with the AMD equivalent being the agesa which is built into the BIOS
iirc those 4 are Intel
my 5900x is on some ancient Aorus BIOS cuz I remember the AGESA reduced all-core
muh frequencies bruh
The others in that list either have no AMD equivalent or they're included in the AMD chipset drivers
damn I am trying to push this 10900K so hard rn lol
Funny enough Intel doesn't distribute chipset drivers any more, they rely on the motherboard support pages to update it 💀
But does anyone know why they separate the chipset drivers into categories like X670, X670E, etc if they are all the same?
that seems sketchy....
I'm at 53x multiplier; 100.65mhz BCLK; bdie at 4429c15-15-15-35@1.74v
SA voltage is 1.49; IO is 1.39
It's mostly just indexing on the website, you can have multiple pages link to the same download
I dunno what's up with this thing but even at like 1.6 SA & 1.55 IO, 4533 doesn't wanna run with 2x8 bdie
I never messed with OC on 10th gen so idk if those are high or not but I am gonna assume those are some very impressive and slightly extreme settings.
these heckin 4000c14 G.Skill Royal Elite things are overrated
look cool though.
Oh yeah royals are such random bins
Speaking of royals any updates on the Trident Z5 Royals for DDR5 yet?
BIOS doesn't seem to approve
I remember eye balling them like 6 months ago when they teased it.
I had some Trident Z5s that I broke
they looked nice
I took the heatspreaders off nonetheless

but then when I tried to remove them to reuse the heckin modules on different memory, an SMD came off with one of the thermal pads
they were only 7200 but heck they were so expensive when I got them
now 7200 is like $130 lol
after I get myself a new graphics tablet I am probably gonna save up and purchase a lower end kit just to get the head spreaders.
Maybe.
How are the royal heatspreaders this gen?
I just noticed, VPP on this board is auto 2.5v
could've sworn 1.8 was typical
I'm throwing spaghetti & setting PLL voltages cuz heck it
Are you sure it's not 2.5 max
max is 3.770 apparently lol
wat
(DDR4 don't forget)
I wonder what that DRAM VREF Voltage Control setting does
the only options are "Auto," "by CPU," & "by HW"
Ouch...... $200 for 2x16GB Z5 Royals.
Rated 7200c34 @ 1.4v
cheapest I could find on amazon.com.mx
on 7950x, you don't really need 7200
get some of those cheapy Klevs I use & have 64gb lol
I just want the heat spreader lmao
9950X can use it supposedly
$80/stick rn https://www.amazon.com/gp/product/B0CGWFYBXB/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
Overwhelming Performance Improved DDR5 Memory: KLEVV DDR5 memory module delivers starting speed at 5600MT/s, 75% faster than DDR4. Lower Voltage, Highly Improved Performance Built-in PMIC: While achieving 20% lower power consumption compared to DDR4 memory, DDR5 memory can reduce dynamic voltage ...
We'll find out in under 2 weeks
I already got myself a 2x32 Hynix A die kit from Kingston Fury.
oh heck why so fancy? lol
maybe I should grab two more & see what I can OC 128gb to
lol 6000c26 1:1 but 128gb? almost sounds worth it
When I was purchasing it was on sale making it the cheapest kit that hit my criteria which was 6000cl32 at 2x32GB
I got these cuz they were cheapest I could find & I didn't have to remove any stinky heat spreaders
Surprisingly kits without heat spreaders cost more at the time.
I got these babies for let's see.
hmm I wonder if I should be running DR for 7zip
I just got a 5600c40 kit that was guaranteed hynix since I was going to be tuning anyway
Currently 6200c30
heatspreaders take time to remove; time is money; so, heatspreaders take money to remove? lol
these klev are same 5600 but c46 iirc
5600c46-46-46-84 is a-die afik
Yikes, $220 about.
Yeah it was weird.
Shocking thing was that this kit was $60 cheaper then the next closest kit at the time of purchasing.
heck, cheap A-die is cheap as heck now. $75 for 2x16 5600c46-46-46-84 https://www.amazon.com/gp/product/B0BPHRVYX9/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
Yep that's my kit
I purchased teamgroup originally but during shipping something happened to my kit that they deemed it unshipable and returned it.
I considered getting 24GB sticks but decided against it. I'll get those in a few gens maybe.
the 5600cl46 A die kit that every one kept recommending me
lol I got those
I got the 7200 m-die 2x24 & the 2x16 5600
I'm basically a creator now according to teamgroup
A or M die is better? I keep forgetting.
So A is better on AMD then.
It's called M die but it's different from the 16Gb die
Z690 Dark ran 7000c32 1T iirc
oh really? heck
the Kingstons I have were a struggle to stabilize at 7k once upon a time
seems so long ago now. strange
4429c14
think it'll POST?
I don't think it's gonna POST
lol
nvm it POSTed easy wth
this 10900K always hated 1T I dunno why
even on SR kits
I wonder if it would POST XMP on a 2x8gb 5000c28 kit
Are you still synced or desynced?
it ran 4533 on a DR DJR kit for like almost 2 years when it was in my son's computer
Intel doesn't do synced/desynced
there's an FCLK I think but it doesn't matter really
it's not like AMD
when UCLK:MCLK is 1:1, is that the same as running gear 1 on Intel? I never understood that whole thing... maybe trying to draw parallels is an apples & oranges type thing
heck
decoding error at c14
If you're considering DDR5 for that question then it's like Gear 2 iirc
I wonder if maxmem is a good idea for 7zip
ya sorry ddr5 only question I meant
Oh damn. Not sure if this is placebo but I think I already see a improvement with the stuttering when switching between apps.
I usually feel a weird stutter when first opening my pc and when switching apps.
So far not feeling it.
someone was telling me yesterday that they run 6600c26 at some insane voltages with SOC voltage at like 1.37 or something
On AMD?
Afaik yes.
Though was that under load or?
Yeah that might be a fire hazard if they're unlucky enough
I found it
Would it be beneficial if even a little to update my bios?
so jk on the tCL but still look at that SOC voltage
It's been a good few months since I updated.
1.38 soc 
BIOS update allowed me to run 6400 1:1 on Taichi but it could also be a total waste of time & you lose your OC profiles
High soc is fine if the cpu is not under load right?
& even if you manually put the exact same numbers in, it's possible it won't work
SOC drops a decent amount under load
Well my profiles I found out are janky because my current bios doesn't always apply the settings when saving for some dumb reason.
I set SOC LLC for that lol
So I am not to concerned with that part..
but I dunno wth this board is up to
It's current that kills but you can't boot without putting load on the CPU
Wait, isn't 6400 1:1 normal on AM5?
You have to be quite lucky
Gotcha. Thanks for the tip.
Really???
I thought I was unlucky when I saw my CPU can do 6400/cl32 synced which I think means 1:1
so what do you think the odds are that this is some sort of SPD bug? cuz I'm starting to think this board takes some serious liberties here since I'm set at 1.55/1.45 respectively
I never got around to primaries so idk if my current cpu can do better then 6400 at tighter timings..
is using almost 6w per DIMM normal? that doesn't sound normal lol
My kit used around 8W average per dimm.
For that to be synced your UCLK would be at 3200
6W is normal for a memory stick
that voltage tho
Your GPU memory uses that for every memory die lol
1.755/1.725v VDD/VDDQ? lol
yeah that seems off.
It's the load voltage that matters but the good news is 1.75v won't kill your sticks, at least not right away
Mine run around that with the stock spreaders and a fan
My UCLK is set = to MCLK so I think so.
So does that mean I actually got lucky with my cpu?
at 6400, I can't run FCLK at 2100 without exceeding 1.3v SOC... would 6200 be better performance since 2067 would be synced FCLK or something? AM5 confuses me
With the memory controller at least yes, assuming it's actually stable and not clock stretching
Test and find out. The answer will vary by what benchmark you're using.
heck I suppose I should
I never figured out core offset magnitudes
they're at like -15 or something I don't even remember
it seems like there's no effective way to test them anymore
hours of core cycler pass but then crash on R23
like uh
The thing with the FCLK is there's a very slight improvement in latency running it 3:2 with UCLK:FCLK, some benchmarks like the latency more than the speed and vice versa for others
I could just do all-core testing but then I have the tedious mission of doing each core at a time
how did you test stability for COs?
iirc you did per CCD
What options did you use for corecycler?
what is clock stretching?
where you put 0 for the size so it keeps switching it up or something
It's difficult to explain, but the result is obvious when it happens because you'll get worse scores at higher frequencies. It's a very common issue on Intel chips.
How many hours because for it to effectively test those settings you'd need at least 2 days
SHA3 10 rounds, 4 iterations of Core Cycler set to whatever the max setting was called on P95 and then 1 round of OCCT got me extremely stable core offsets on the cores I did using that method.
eh probably like 8-10
hours
It didn't go through all the cycles then
I gave up tho & just left them where they're at
I think I set like uh, 45 minutes per core or something
that got me extremely stable results where I did that..
the problem is it's on my daily & I get impatient
I just left Core Cycler ruynning in the background 24/7
I do one core a day roughly was how I went about it.
Unless I am at the computer when it passes then I go back and then rinse and repeat
Do SHA3 first as that's the most sensitive to CO values
Unfortunately my Aida trial ran out and I couldn't finish my method on the final 3 cores or so.
If you use the complete uninstall method in revo uninstaller that usually lets you reset trial timers
Since it scrubs the registry as well as file structures
is revo uninstaller a 3rd party software or?
Aida license is like what $40?
60
Any idea what they mean by for 3 pcs?
And it's only for 1 year
You can use the same key on 3 different PCs at once
On a different topic has there been any news on next gen gpus and maybe, hopefully a 8950x3D?
.....
Thank you, Fal.
9950X3D is due in a few months, expected to launch at the same time as the new chipsets with USB 4 standard
9950X and other next gen non-3D comes at the end of this month
oh awesome.
iirc MSI boards come with it or something
I never actually redeemed it lol
I had an engineering key for awhile but when I update AIDA it's like omg need new key
Gotcha.
I read some article mention that a multi-gpu graphics card might be releasing?
Nvidia might paper launch their 5090/5080 in December but most of their production is focused on $20k AI cards right now
It's one possible configuration. Consider it a proof of concept or engineering sample. It may or may not make it to market.
oh crap I just remembered 10th gen RTL/IOL tuning
lol maybe sell one 7950x?
My father had a medical problem and later I got sick which ate up every thing I saved.
Father had what is basically a brain eating bacteria, thankfully it's almost gone now.
ah heck I was kidding but I hope your father makes a full & speedy recovery if he hasn't already
And I have been coughing and sneezing blood for the past two nearing three weeks now.
and you do too
Ahh gotcha
Makes sense.
Thank you, I appreciate it.
your lungs okay or?
btw if anyone is guilty of hoarding hardware here, it's likely me lol
Not sure but I have been getting better.
The docter didn't explain what I had when he gave me the medicine....
xD
I think we both are.
I still have my 10700F, 5700G, and every thing for them
it works out nice for my son. He just got upgraded to 13600K/Z690 Unify-X/RTX3070 for his 7th birthday
I feel you I have 5900x/B550 Aorus Master; 10700KF/Z590 Unify-X; 10900K/Z590 Unify-X (I'm an idiot & got two cuz I forgot to return one to Amazon); a G7400 with no home; 12900K/Z690 Dark KP; 13900KS/Z790 Apex; a random MSI Z490 Gaming Carbon Wifi; a Z590 Aorus Master I've been meaning to mail for weeks now... I think that's all of it
well not including GPUs
I have way too many GPUs & most are only useful for hwbot
others could still be okay
& others are good for nothing at all lol
wowza, its crazy when u remove limits and oc cpus the power goes up!
intel power hungry 😡
(me moments after setting my PL1 and PL2 to 9999)
Tbf it’s on 14 nm++++++++
True.
y-cruncher almost pushed my cpu's power up to 160w
ring ratio hasn't been configured yet so its running stock
so far im satisfied with the undervolt
oc'd the ring to 4.7ghz, small difference but i'll take it
oc'd the ring to 4.7ghz, small difference but i'll take it
annnnnnnnd it crashed in linpack
lol
well passed linpack but had to sacrifice both e-core + pcore clocks by -100mhz
lel
the cpu peaked at 150w
so not too bad
ill compare my 5.4ghz oc to the 5.2ghz oc/uv later
And still somehow alive smh
Just built different
Real
Too bad one of the likely causes of Intel chips failing is that ring bus running too high of voltage and increasing the clock on it would exacerbate the issue
is it?
i thought it was a current issue?
just the node being blasted with amps, thus why the 2 higher boosting p cores (at stock) degrade first
It's also potentially (likely) an issue with the physical manufacturing
Gn made a video, it mentioned the oxidation method? It's honestly above my pay grade education wise
Gn being gamers nexus
The oxidation is caused by improperly applying the compound meant to stop corrosion, or other contamination during the process of applying it.
What that means is it requires more voltage to overcome the resistance caused by the oxidation, which can contribute to the rapid degradation and would also explain why not every chip seems to be affected.
That was initially what was thought to be a cause, but it turns out that's more of a symptom
Intel's 13900k's and 14900k's are crashing at an alarming rate? Why isn't anyone talking about it and what is Intel's solution?
Forum Thread here: https://forum.level1techs.com/t/intels-got-a-problem-13900k-ks-kf-14900k-ks-kf-crashing/213008?u=autumn
0:00 Intro
1:07 The Data
5:08 Data Breakdown
10:17 What's Causing the Problem?
13:30 What's Be...
cue the "my 13/14th gen i7/i9 isn't experiencing this so it must be fake" comments
They're already flooding Facebook
(Most of the commenters are likely not owners of the affected chips)
Evc mod who?
Nice, what card? Looking to bench for scores? or daily
7900XTX, just daily probably
Doubt I’ll take full advantage of it but it’s just a cool project
First time soldering went pretty well
Yeah, looks good - those stranded wires are hard to solder to small pads like that - but thankfully the SPI dedicated pinout is quite large on most cards
Soldiering looks solid yeah
Why not? Just because you have EVC doesn't mean you need to push voltage
what else r u doing with it lul
monitoring vcore?
changing llc?
increasing vrm switching 
Pushing it when you're benching and fixing voltage stability issues for daily
fair enough
sounds like a hassle to me personally
but i dont oc daily gpus anyway
what pwm controllers do high end amd even use these days? do cards need evc to read the pwm before gpu bios loads still?
or can u read them in OS now?
you spend 5 mins once to set it up
Very easy to setup
I mean, if you just increase the power limit to like 7-800 realistically games will only take up to 600(for me so far)
yeah for sure, just cards i used evc on - i needed to read the pwm controller before gpu bios loaded. so id need a 2nd pc to run the run/control the evc
not sure if newer cards or pwm controllers need that
what oc software do you guys recommend for over clocking a rtx 4090 24g (asus tuf gaming) please ping in response
MSI Afterburner
Lunar has a good guide. Stock voltage is more than generous so like Segue mentioned, could check an UV out (combined w/ an oc or not)
https://github.com/LunarPSD/NvidiaOverclocking/blob/main/Nvidia Overclocking.md
It can plug right into my own PC now
Whole software for it
Super sick I’ll send a picture tmrw
yeah ur controller probably lets it read/write after bios loads
think i had a up9512 that didnt lemme do anything unless i did it pre-bios load
not sure if read write is the correct terms even
my limit was cooling anyway, +150mV offset made clocks worse lol
if i had the means to take it cold, VID mod wouldve probably of been a must
not sure if refin/vref mod would work on the card
@proven canopy hate to ping you, but I'm not versed in stability testing. I believe prime95 is what I would use to test this 12600k, but I'm not sure what settings to use. I'd appreciate your input.
I would just download benchmate, run ycruncher 1b for mem+cpu, then run the bbp tests for cpu, quick and easy.
If you want something more thorough for mem, try memtestpro, ycruncher vst/similar
Gotcha. I'd like to do so, but I'm not in a position for a few days that I can download a 1.5gb installer.
Prime95 not a good stress tester?
I'm just making sure this PC is ready to go. Stock speeds. It was my original PC that I had no issues with, but I sold it to a friend and pretty much from week 1 he had issues. Finally got it back in my possession to troubleshoot. I used TM5 to test the ram which seems to be fine. Now I just need to do the CPU.
p95 is totally fine for testing most aspects of core stability
If I were to sell a gamer rig, I'd do the benchmate stuff because it takes a few minutes and will catch anything obvious, then move on to the rest and run it overnight afk
Give occt a try if you haven't already
I will definitely do this in the future.
Didn't know this existed. Thank you sir
I just finished my pc and got everything up-to-date, is there anything you guys would recommend me doing for overclock or anything?
CPU: 12600k
GPU: Geforce RTX 4070
RAM: G.Skill 32GB 6000-36
Perhaps latest bios if you haven't actually done that. Otherwise, I'd just install windows or do some benches w/ it totally stock 1st.
btw Y cruncher VST got phased out, VT3 replaced it.
no clue if its better or not, just know that it works 
VT3 seems good
yeah it works great, not sure if its faster or better than VST tho.
More demanding than VST but takes slightly longer
@proven canopy if it's easy enough could you tell me what "ddr5 nitro mode" does in a tl;dr
more fast
mclk can go far past that, uclk sync, eh
I have this graphic card:
Gigabyte GV-N108TAORUSX WB-11GD AORUS GeForce GTX 1080 Ti Waterforce WB Xtreme Edition 11G Graphic Cards
And this screenshot is what my MSI afterburner is showing.
I have never done any overclocking before - my friend gave me an advice to look into that. I dont want to mess anything up lol. Is there maybe a website that shows like "idea" overclocking settings and stuff?
the 1080 is a pascal card btw
personally gaming has been my stress tests
3dmark doesnt cut it for real world instabilities
thanks!
the world if good gpu stress tests exist 🏙️
So anyone know any surefire/good OCs for a RTX 4080 Super? (Don't want to break it or anything just got it. Zotac Trinity OC White for my card)
Yeah, just wanted to know if the 4080 Super had headroom or not like the 4070Ti Super (The reason I went with the 4080 Super over it was becuase it did have a notable step over the 4070Ti Super in a number of the games/settings I played, probably due to it's GPU size difference)
you bought a 4070 ti super and 4080 super?
ok
well its bc its bc its significantly better than the 4070 ti super
idk if it has any headroom all i know is that modern gpus arent good at oc
yeah most modern gpus and cpus dont have much OC headroom without some exotic cooling solution
they've gotten good at binning the chips
crank that power limit and increase the fan profile if you dont want to deal with stability testing
I didn't buy both lol. I went with the 4080 Super after looking at how the 70Ti Super faired versus it
What fitz said, crank power limit, and also unlock voltage and increase that
Then whether you get a modest or great result depends on how much you crank core and memory
Stability test? Games are the best thing for that lol
Ada has some pretty great headroom in general
Anything interesting happen lately?
In the world of overclocking? Not really
Ahh rip, what about new CPU and GPU improvements in general?
The only recent launch is the 7600X3D which you can't overclock.
Everyone is scared to overclock their Intel chips now, plus the recent patches kinda kill performance.
Ouch.
More and more I am glad I listened to you, Felix, and Bacca regarding going AMD when I made my purchase.
14600K yet to actually replace my solid 12400. 
well, team cup is ending in a day and 9 hours https://hwbot.org/competition/HWBOT Team CUP 2024/
Level 2 OC-ESPORTS competition running from 2024-07-15T12:00 till 2024-09-15T12:00.
hello
i’m in my bios rn
and idk if i should turn on xmp
do u guys recommend it on or off
What CPU do you have and what memory kit
Usually you want it on but sometimes there's a compatibility issue
5 3600
and a mpg b550 gaming plus motherboard
with ddr4 3200mhz 16gb
i turned it on
my pc is fine
but now my ram is at 3200 instead of 2444 before
or wtv it was before
Yep should be no issues with that. It will keep games running smoother.
bro... what is that vcore voltage im seeing?
That's normal for bios
It's a low load enviroment
Just enough to make it boost but not enough to overheat it
ah ok
5 billion volts
I mean honestly anything up to 1.5v is fine for a short duration, even 1.6v. It sounds crazy but the CPU will just overheat & shut off before any voltage damage can be done. You DO NOT want to run it at 1.4v or anything higher for a prolonged duration like say a month long or even a week long that's bad. But a few hours while benching won't kill it.
I know this from experience 
3+ week response but ill take it
That's how back logged the Techie dept is... lmao
😂
Guys help. I'm getting the itch to try to overclock my memory again. Someone stop me from pushing vSoC up to/above 1.3v to get an extra 200 MHz mclk/uclk. There's so many other things I need to spend my time on today.
WHEA uncorrectable will appear in your dreams
must keep urge away
I tested fclk stable up to 2133, 2166 saw regression but no crashing, so it would still be stable for 3:2 ratio
It's running at 2066 currently so it's 3:2 for 3100 mclk
sounds reasonable enough, you don't need more instability when you can be doing something else
like laundry or something
It's my day off so the "something else" would be working on refurbishing other hardware to sell
make money so you can buy more ram
That reminds me, I need to sell some hardware to get another AM4 CPU to replace my server
I can scratch the overclocking itch once I have that since I still have my b-die
I'd argue that my kit of 5600 jedec Hynix M is more fun to play with because b-die is so predictable
Might have missed it but what’s your latency?
Check this out on @Newegg:Team T-CREATE CLASSIC 32GB (2 x 16GB) 288-Pin PC RAM DDR5 5600 (PC5 44800) Desktop Memory Model CTCCD532G5600HC46DC01 https://www.newegg.com/team-32gb-ddr5-5600/p/N82E16820985008
I currently have it at 6200 c32 with buildzoid easy timings for my daily OC




