my Settings: only offset -0.100 p1 253 p2 270 no changing any clocks on -0.120 a was getting a windows-crash after cinebench. when you will see the 6,2 Ghz than you must change pl2 to 280 watt max and do for security max temp trothling to 89° i have to testing this more when i lover pl1 and pl2 one of you have this option testet!? and how many can we go down with watts to see 6,2 ghz (+ with -0.100 undervolting) we all need a point for max Underwatting and undervolting together and max 6,2 ghz POWER 😂 please help need it cooooool ❤
This is a pretty bad way to undervolt it, while it does work, it's not the optimal way at all. You should learn how to use V/F Curve points, it's much better in every way. Also why BY CORE USAGE and not SYNC ALL CORES? I've seen weird things happen when using by core usage, even when set to to the same multiplier, just use sync all cores instead. There's no reason to set Min. Cache Ratio to 50, you're just disabling your rings ability to idle and drawing power more power for no reason at all. It looks like you're using Hynix, you can easily set tRRD 8/4 and tfaw 16, tfaw 12 does nothing because it can only be trrd * 4. Can also set tRCDWR to 16. The tertiaries are also pretty bad. You could easily do 16, 8, 12, 8, 60, 50 ezpz
@@Yahfz By core usage lets the CPU down clock under certain loads, never while gaming. It’s funny, I have never experienced anything weird. Sounds like a you problem. Also, I see zero point in doing anything with the ram timings you mentioned as they have zero impact on gaming. Also, a 42K in r23 after a 30 minute loop with an AIO? Yeah I see no issues here.
@@BrettNerd SYNC CORES downclocks too.. wtf are you saying? Also the timings have zero impact in gaming? lol Thanks for reminding me why i don't try to help people online anymore :)
@@BrettNerd What makes you think mine isn’t stable? Also karhu is just one test, you still gotta do ycruncher vt3/tm5 on top of it to be truly stable. You're obviously new to overclocking based on the timings/bios settings which is ok, i was just trying to help. I don't feel like providing proof for the timings because this is such an obvious thing literally anywhere online will tell you that tuning secondaries/tertiaries improves performance. If you don't wanna do it, that's also fine, i'm not gonna go out of my way to change your mind, it's your loss. Also you seem to have a really weird idea about this, just because you're scoring 42K or whatever it is you think is good doesn't mean it couldn't be better.
@@Yahfz karhu is the only benchmark that properly tests ddr5. Ive seen people run ycruncher, anta777 profiles and pass and fail Karhu in 30 seconds. New to overclocking? Ive been overclocking on x58 before you even knew what an i7 was. You disable ht, which already tells me you know a lot less than you think you do. As a buildzoid fanboy, he would be the first to tell you the timings you keep yapping about have very minimal difference. Der8auer, who I trust a lot more than random zero smt guy in youtube seconds that.
@@BrettNerd Actually not true, the bug is in the Microcode where the VID request can spike 1.6-1.7v but these short spikes won't always register in hwinfo. No matter how you optimise your VID table it won't fix the "bug".
@@aurorakai Why are you spreading misinformation? Buildzoid has hooked several 13/14th gen cpus under the osciloscope several times on much older microcode and there's no such spikes. These only happen if you're running the fused voltage unaltered which is what he said. TLDR: High VID requests can only happen if you let the CPU do whatever it wants on microcodes pre-129. If you tune your voltages you don't need to update your mcode at all.
@@Yahfz Sorry, I went on buildzoids videos too. I may have misunderstood but honestly why not use the latest microcode? It sets a limit of 1.55v. I am using the latest Microcode but I also tuned the voltages myself, my VID table is around 1.32v max and on load its around 1.170v
@@GuilhermeVieira-92 You will have to play around with the undervolt offset for stability. I would start with a 0.06V offset and see what your VID table looks like. If 0.06v is stable, try 0.08V and so on until you are no longer stable. Anything above 1.375V is on your vid table is getting a bit uncomfortable in my eyes.
what cooling config are you using this screams custom water loop....IF so thats kinda irrelevant to the vast majority of users who are most likely using 360mm AIOs......I have a 420mm aio and my voltages are bouncing up and down all the time. Thats with IA ceps enabled cstates enabled cores locked (multicore enhancement off though) and power limits to 253 pl1 and pl2. LLC 6 AC/DC @ .5/.5 HOWEVER I am on an MSI z790 ACE mobo and MSI will NOT unlock the IA CURRENT max limit settings for their boards which is why I will NEVER EVER EVER buy MSI again. It will be ASUS 2 dimm slot only boards from this point forward. MSI is garbage kinda like corsair is now.
ALSO you cannot adjust or even view your VID table on and MSI board without setting the voltage mode to Adaptive VFcurve....Its kind of ridiculous how dumb the MSI technicians/engineers are when it comes to mobo programing developing etc.
Thank you for posting a guide.
Well done 👍 👌
Use voltage caps job done
my Settings: only offset -0.100
p1 253 p2 270
no changing any clocks
on -0.120 a was getting a windows-crash after cinebench.
when you will see the 6,2 Ghz than you must change pl2 to 280 watt max and do for security max temp trothling to 89°
i have to testing this more when i lover pl1 and pl2
one of you have this option testet!?
and how many can we go down with watts to see 6,2 ghz (+ with -0.100 undervolting)
we all need a point for max Underwatting and undervolting together and max 6,2 ghz POWER 😂 please help
need it cooooool ❤
This is a pretty bad way to undervolt it, while it does work, it's not the optimal way at all. You should learn how to use V/F Curve points, it's much better in every way. Also why BY CORE USAGE and not SYNC ALL CORES? I've seen weird things happen when using by core usage, even when set to to the same multiplier, just use sync all cores instead.
There's no reason to set Min. Cache Ratio to 50, you're just disabling your rings ability to idle and drawing power more power for no reason at all.
It looks like you're using Hynix, you can easily set tRRD 8/4 and tfaw 16, tfaw 12 does nothing because it can only be trrd * 4. Can also set tRCDWR to 16. The tertiaries are also pretty bad. You could easily do 16, 8, 12, 8, 60, 50
ezpz
@@Yahfz By core usage lets the CPU down clock under certain loads, never while gaming. It’s funny, I have never experienced anything weird. Sounds like a you problem. Also, I see zero point in doing anything with the ram timings you mentioned as they have zero impact on gaming. Also, a 42K in r23 after a 30 minute loop with an AIO? Yeah I see no issues here.
@@BrettNerd SYNC CORES downclocks too.. wtf are you saying? Also the timings have zero impact in gaming? lol
Thanks for reminding me why i don't try to help people online anymore :)
@@Yahfz show me proof it does. If it ain’t Karhu stable, it ain’t stable.
@@BrettNerd What makes you think mine isn’t stable? Also karhu is just one test, you still gotta do ycruncher vt3/tm5 on top of it to be truly stable. You're obviously new to overclocking based on the timings/bios settings which is ok, i was just trying to help. I don't feel like providing proof for the timings because this is such an obvious thing literally anywhere online will tell you that tuning secondaries/tertiaries improves performance. If you don't wanna do it, that's also fine, i'm not gonna go out of my way to change your mind, it's your loss. Also you seem to have a really weird idea about this, just because you're scoring 42K or whatever it is you think is good doesn't mean it couldn't be better.
@@Yahfz karhu is the only benchmark that properly tests ddr5. Ive seen people run ycruncher, anta777 profiles and pass and fail Karhu in 30 seconds. New to overclocking? Ive been overclocking on x58 before you even knew what an i7 was. You disable ht, which already tells me you know a lot less than you think you do. As a buildzoid fanboy, he would be the first to tell you the timings you keep yapping about have very minimal difference. Der8auer, who I trust a lot more than random zero smt guy in youtube seconds that.
I got almost an identical bin to yours. But I have a slightly better 56x and 59x point.
Should yield you a decent undervolt I would think
yea same VID table as yours. I also put in 1400mv cap. Hitting 78C on a 360mm aio at 320W.
That’s too pretty. I’m trying to achieve the same with my 14900hx. Even with 65ac and some offset it ain’t happening
Are you on the latest bios with the microcode update?
@@holyearth No. it’s not needed if you manually tune your voltages properly and don’t cook your CPU.
@@BrettNerd Actually not true, the bug is in the Microcode where the VID request can spike 1.6-1.7v but these short spikes won't always register in hwinfo. No matter how you optimise your VID table it won't fix the "bug".
@@aurorakai true
@@aurorakai Why are you spreading misinformation? Buildzoid has hooked several 13/14th gen cpus under the osciloscope several times on much older microcode and there's no such spikes. These only happen if you're running the fused voltage unaltered which is what he said.
TLDR: High VID requests can only happen if you let the CPU do whatever it wants on microcodes pre-129. If you tune your voltages you don't need to update your mcode at all.
@@Yahfz Sorry, I went on buildzoids videos too. I may have misunderstood but honestly why not use the latest microcode? It sets a limit of 1.55v. I am using the latest Microcode but I also tuned the voltages myself, my VID table is around 1.32v max and on load its around 1.170v
I can use this settings on 13900k?
@@GuilhermeVieira-92 You will have to play around with the undervolt offset for stability. I would start with a 0.06V offset and see what your VID table looks like. If 0.06v is stable, try 0.08V and so on until you are no longer stable. Anything above 1.375V is on your vid table is getting a bit uncomfortable in my eyes.
Thanks for the info!!
what cooling config are you using this screams custom water loop....IF so thats kinda irrelevant to the vast majority of users who are most likely using 360mm AIOs......I have a 420mm aio and my voltages are bouncing up and down all the time. Thats with IA ceps enabled cstates enabled cores locked (multicore enhancement off though) and power limits to 253 pl1 and pl2. LLC 6 AC/DC @ .5/.5
HOWEVER I am on an MSI z790 ACE mobo and MSI will NOT unlock the IA CURRENT max limit settings for their boards which is why I will NEVER EVER EVER buy MSI again. It will be ASUS 2 dimm slot only boards from this point forward. MSI is garbage kinda like corsair is now.
ALSO you cannot adjust or even view your VID table on and MSI board without setting the voltage mode to Adaptive VFcurve....Its kind of ridiculous how dumb the MSI technicians/engineers are when it comes to mobo programing developing etc.