Overclock.net banner
1 - 20 of 22 Posts

shanenc14

· Registered
Joined
·
27 Posts
Discussion starter · #1 ·
Hi everyone,

I’ve been tinkering around with OC’ing my 12600KF CPU.
I've tried to configure with some "safegaurds" to prevent any long-term degredation issues, as well as have vCore high enough to keep the 2 P-cores stable @ 5.5 GHz.... The last CPU I OC'd was an AthlonXP w/ WinXP... Showing my age here...lol

I would like some input from others with much more experience on OC'ing the modern ADL/RPL hybrid chips.

  • Is vCore at 1.38-1.39v under light loads (1-2 P-Cores @ 5.5GHz) harmful, or too high?
  • What are safe voltages for light loads, as well as voltages when under load like Cinebench R23 or Y-Cruncher Component Stress Tester?
  • What is the best way to test stability of 1-2 P-cores boosting to 5.5GHz?

My Current System Specs

CPU: i5-12600KF w/ Thermalright LGA1700 Contact Frame
Cooler: Thermalright Frozen Notte 240mm AIO (Top-Mounted exhaust)
MB: MSI MAG Z790 Tomahawk WiFi DDR4
RAM: TeamGroup T-Force VulcanZ DDR4-4000 CL18 2x16GB kit (XMP= 4000 CL18-24-24-46 @ 1.4v)
Currently running at: 3800 MT/s CL17-21-21-21 Gear 1, CR1 @ 1.39v (Haven't tried lower) SA voltage @ 1.205v
GPU: XFX AMD RX 6600 8GB
PSU: Thermaltake BM2 550 Watt 80+ Gold​

Imgur Link to BIOS settings & HWinfo64 after Windows fully loading - showing max vCore under light load

Current OC

1-2 P-Cores: 5.5 GHz
3-4 P-Cores: 5.4 GHz
5-6 P-Cores: 5.3 GHz
E-Cores: 4.2 GHz all-core
Ring: 4.7 GHz
vCore: 1.345v Adaptive Mode

Peak vCore under light load (i.e. Windows startup = 1.386v)
vCore under load w/ Cinebench R23 = 1.292v - 1.294v
LLC Mode: 6
MSI Lite Load: Auto (Mode 1)
PL1= 220 / PL2=241 / IccMAX (Max Current): 240A
System Agent= 1.205v

Any questions, please feel free, and TYVM for all comments/suggestions!
 
Hi everyone,

I’ve been tinkering around with OC’ing my 12600KF CPU.
I've tried to configure with some "safegaurds" to prevent any long-term degredation issues, as well as have vCore high enough to keep the 2 P-cores stable @ 5.5 GHz.... The last CPU I OC'd was an AthlonXP w/ WinXP... Showing my age here...lol

I would like some input from others with much more experience on OC'ing the modern ADL/RPL hybrid chips.

  • Is vCore at 1.38-1.39v under light loads (1-2 P-Cores @ 5.5GHz) harmful, or too high?
  • What are safe voltages for light loads, as well as voltages when under load like Cinebench R23 or Y-Cruncher Component Stress Tester?
  • What is the best way to test stability of 1-2 P-cores boosting to 5.5GHz?

My Current System Specs

CPU: i5-12600KF w/ Thermalright LGA1700 Contact Frame​
Cooler: Thermalright Frozen Notte 240mm AIO (Top-Mounted exhaust)​
MB: MSI MAG Z790 Tomahawk WiFi DDR4​
RAM: TeamGroup T-Force VulcanZ DDR4-4000 CL18 2x16GB kit (XMP= 4000 CL18-24-24-46 @ 1.4v)
Currently running at: 3800 MT/s CL17-21-21-21 Gear 1, CR1 @ 1.39v (Haven't tried lower) SA voltage @ 1.205v​
GPU: XFX AMD RX 6600 8GB​
PSU: Thermaltake BM2 550 Watt 80+ Gold​

Imgur Link to BIOS settings & HWinfo64 after Windows fully loading - showing max vCore under light load

Current OC

1-2 P-Cores: 5.5 GHz
3-4 P-Cores: 5.4 GHz
5-6 P-Cores: 5.3 GHz
E-Cores: 4.2 GHz all-core
Ring: 4.7 GHz
vCore: 1.345v Adaptive Mode

Peak vCore under light load (i.e. Windows startup = 1.386v)
vCore under load w/ Cinebench R23 = 1.292v - 1.294v
LLC Mode: 6
MSI Lite Load: Auto (Mode 1)
PL1= 220 / PL2=241 / IccMAX (Max Current): 240A
System Agent= 1.205v

Any questions, please feel free, and TYVM for all comments/suggestions!
Others may disagree but i think just syncing all performance cores to 5.3 would be best for lower top end voltage. Im not really sure if there is any benefit for that 200mhz that happens for a split second here or there. Just set intel recommended power limits and amp max limit for the 12600k.
realbench is suppose to be good for variable cpu loading. Ycruncher 2.5b seems more effective then cinebench r23 for all core work load. It requires more voltage to pass.
what do you score in cinebench r23?
 
Discussion starter · #3 · (Edited)
I do have a OC profile saved in BIOS with 5.3 all p-core, 4.2 all E cores, and 4.7 ring. Multi core CBr23 is a bit over 20k. Based on the tinkering I've done, I agree with settling on 5.3/4.2 all core.

Edit: I've tried setting iccmax to 175A according to Intel specs, and even with the all core 5.3/4.2 overclock, hwinfo64 says I'm being held back by current limit, however, the current actually showing VR VCC Current field doesn't hit 175a even during y-cruncher component stress testing with ICCMax set to 240A in BIOS, so I'm a bit confused with that.
 
I do have a OC profile saved in BIOS with 5.3 all p-core, 4.2 all E cores, and 4.7 ring. Multi core CBr23 is a bit over 20k. Based on the tinkering I've done, I agree with settling on 5.3/4.2 all core.

Edit: I've tried setting iccmax to 175A according to Intel specs, and even with the all core 5.3/4.2 overclock, hwinfo64 says I'm being held back by current limit, however, the current actually showing VR VCC Current field doesn't hit 175a even during y-cruncher component stress testing with ICCMax set to 240A in BIOS, so I'm a bit confused with that.
From point of view the new bioses really gimp performance. I would run unlimited amps and keep your power limits set. See what you get now for perf.
 
Discussion starter · #5 ·
From point of view the new bioses really gimp performance. I would run unlimited amps and keep your power limits set. See what you get now for perf.
Right now I have the 5.3/4.2 all core, ring 4.7 profile loaded, doing some stability testing with OCCT and y-cruncher. I have the following "safeguard" settings in place. Max current: 240A, PL1=220 watts, PL2=241, tau 56 sec. Would you say this is safe? Or, should I set PL1=PL2= 220 watts, or something similar? I could post BIOS screenshots of my 5.3/4.2 all core OC to Imgur if it would help. And thanks a million for helping me out here. I've been out of the OC loop since AMD Athlon XP says.... Many moons ago, indeed!
 
Right now I have the 5.3/4.2 all core, ring 4.7 profile loaded, doing some stability testing with OCCT and y-cruncher. I have the following "safeguard" settings in place. Max current: 240A, PL1=220 watts, PL2=241, tau 56 sec. Would you say this is safe? Or, should I set PL1=PL2= 220 watts, or something similar? I could post BIOS screenshots of my 5.3/4.2 all core OC to Imgur if it would help. And thanks a million for helping me out here. I've been out of the OC loop since AMD Athlon XP says.... Many moons ago, indeed!
I do not think i have helped much at all. But amps is limiting you. Try unlimted amps or set it to 300 and see how perfs is then

edit: i keep thinking your processor is 13600k for some weird reason. If you achieve 20k on a 12600kf in cinebench r23 that is working good. You could have 12900k performance cores in that 12600kf but just two failed and your cpu was demoted to a 12600 sku. I would say if your cpu is stable and not exceeding 90c with that amps and watts good cpu you have im my opinion.
 
Discussion starter · #7 ·
I do not think i have helped much at all. But amps is limiting you. Try unlimted amps or set it to 300 and see how perfs is then

edit: i keep thinking your processor is 13600k for some weird reason. If you achieve 20k on a 12600kf in cinebench r23 that is working good. You could have 12900k performance cores in that 12600kf but just two failed and your cpu was demoted to a 12600 sku. I would say if your cpu is stable and not exceeding 90c with that amps and watts good cpu you have im my opinion.
Yep, it's a 12600KF. I've kinda done away with the OC profile in my orig. post, and went with the 5.3 All P-core, 4.2 All E-Core, and 4.7 Ring. I just ran a quick CineBench r23 Single & Multi (link below). My BIOS Settings are as follows during this run.
1.295v Adaptive, LLC Mode 6
MSI Lite Load: Intel Default
PL1=PL2=220 Watts
Max Current: 240A (Intel Spec for 12600k is 175A, 12700K is 240A, so I just went with that...

I had seriously been considering selling this chip and getting a 13700k, BUT with all the horrendous stuff I've been reading online, and vids I've watched on YT with regards to 13/14 Gen RPL and RPL Refresh, I'm not so sure I want to do that at all. Since those degredation issues don't seem to be an issue with ADL, I'm considering just getting a 12900k. My purpose for upgrading from my current 12600KF is as a last hurrah for my LGA1700 platform and hopefully extend the life of the system for another 5+ years, and at least 1 GPU upgrade from my current RX6600 when I eventually move from 1080p to 1440p gaming.

Cinebench R23 - P: 5.3 GHz / E: 4.2 GHz / R: 4.7 GHz
 
Just keep under the max Intel Turbo Spec wattage for your particular chip. It's more or less accurate.
 
  • Rep+
Reactions: shanenc14
Are you saying to limit the chip to 150 watts (Intel spec PL2)? Or are you saying to keep it under the max current according to Intel specs which is 175A for 12600k/kf?
Regardless of what others keep claiming, in my own personal experience, the total wattage has been the most important factor, not the current.
After all, wattage is a function of current and voltage.

Don't overcomplicate things and just keep it simple.
 
Yep, it's a 12600KF. I've kinda done away with the OC profile in my orig. post, and went with the 5.3 All P-core, 4.2 All E-Core, and 4.7 Ring. I just ran a quick CineBench r23 Single & Multi (link below). My BIOS Settings are as follows during this run.
1.295v Adaptive, LLC Mode 6
MSI Lite Load: Intel Default
PL1=PL2=220 Watts
Max Current: 240A (Intel Spec for 12600k is 175A, 12700K is 240A, so I just went with that...

I had seriously been considering selling this chip and getting a 13700k, BUT with all the horrendous stuff I've been reading online, and vids I've watched on YT with regards to 13/14 Gen RPL and RPL Refresh, I'm not so sure I want to do that at all. Since those degredation issues don't seem to be an issue with ADL, I'm considering just getting a 12900k. My purpose for upgrading from my current 12600KF is as a last hurrah for my LGA1700 platform and hopefully extend the life of the system for another 5+ years, and at least 1 GPU upgrade from my current RX6600 when I eventually move from 1080p to 1440p gaming.

Cinebench R23 - P: 5.3 GHz / E: 4.2 GHz / R: 4.7 GHz
Well there is rumors of a pure 10core and 12core cpus coming for lga1700 yet. No ecores so that might be the last hurrah for this socket. Depends i suppose if they are for consumer end but have to wait and see.

edit: i just seen your imgur picture of your benchmark run. Looks pretty good for a 12600kf. I believe a 12700k stock is around 22,000 points.
 
Discussion starter · #12 ·
Well there is rumors of a pure 10core and 12core cpus coming for lga1700 yet. No ecores so that might be the last hurrah for this socket. Depends i suppose if they are for consumer end but have to wait and see.

edit: i just seen your imgur picture of your benchmark run. Looks pretty good for a 12600kf. I believe a 12700k stock is around 22,000 points.
Well, I had thought about Bartlett Lake... that is until I saw THIS. Are the 13/14 gen chips really this bad, or are people blowing things out of proportion?
 
Discussion starter · #13 ·
Regardless of what others keep claiming, in my own personal experience, the total wattage has been the most important factor, not the current.
After all, wattage is a function of current and voltage.

Don't overcomplicate things and just keep it simple.
Right now I'm running PL1=PL2=220 Watts, Max Current 240A. Intel spec for 12600k/kf is 175A, and with these settings (240A), even running y-cruncher component stress tests, I have yet to see the "VR VCC Current (SVID IOUT)" field in Hwinfo64 reach 175A, so I'm assuming these settings are safe long-term... Would you agree? When I set a limit of 175A in BIOS, Hwinfo64 claims I'm being limited by ICCMax even when Windows is loading, and the P-Cores will not boost above 4.9GHz. BIOS vCore is 1.295v, MSI LLC Mode 6, MSI Lite Load: "Intel Default", so I raised the limit from spec 175A to 240A.
 
Right now I'm running PL1=PL2=220 Watts, Max Current 240A. Intel spec for 12600k/kf is 175A, and with these settings (240A), even running y-cruncher component stress tests, I have yet to see the "VR VCC Current (SVID IOUT)" field in Hwinfo64 reach 175A, so I'm assuming these settings are safe long-term... Would you agree? When I set a limit of 175A in BIOS, Hwinfo64 claims I'm being limited by ICCMax even when Windows is loading, and the P-Cores will not boost above 4.9GHz. BIOS vCore is 1.295v, MSI LLC Mode 6, MSI Lite Load: "Intel Default", so I raised the limit from spec 175A to 240A.
Every kind of limiter is flawed because of how PCs work. The power/current demanded by the CPU at any given millisecond can fluctuate from one extreme to the next.
It is for that reason that most overclockers opt to not use any limiters at all, and properly overclock/undervolt. Limiters are like a breaker trip.

There are lots of approaches to find the ideal power setup for your BIOS, but a power limit with unlimited voltage and current is the quickest and simplest one.
That way, the PC will use whatever voltage/current is necessary for a program, but if the power limit is hit, it'll throttle down accordingly.

Current limiting is more effective and accurate, but involves a lot more effort to perfect.
Voltage limiting (setting a max Vcore) is more of a meme and not recommended, since you can go up to 1.40-1.50V just fine if the current is super low.

The most ideal setup though? Properly overclock and test the maximum potential of your chip at your max wattage spec, and keep to it.
 
  • Rep+
Reactions: shanenc14
Well, I had thought about Bartlett Lake... that is until I saw THIS. Are the 13/14 gen chips really this bad, or are people blowing things out of proportion?
Oh i do not watch that guy. Most of this rumor stuff and “from my sources” is like a soap opera to me. I think alot of people do not have issues with 13th and 14th gen. There are of course people definitely have issues which is what rma is for. Bound to be faulty samples with the millions of processors that are made.
 
Discussion starter · #16 ·
Oh i do not watch that guy. Most of this rumor stuff and “from my sources” is like a soap opera to me. I think alot of people do not have issues with 13th and 14th gen. There are of course people definitely have issues which is what rma is for. Bound to be faulty samples with the millions of processors that are made.
I'd like to sell my 12600kf, and upgrade to something like a 13700k/kf to extend the life of the LGA 1700 platform, but I'm just not sure what to think of a lot of the stuff I've read lately about 13/14 gen. In your personal opinion, would you hesitate at all to buy a 13/14 gen CPU at the moment?
 
I'd like to sell my 12600kf, and upgrade to something like a 13700k/kf to extend the life of the LGA 1700 platform, but I'm just not sure what to think of a lot of the stuff I've read lately about 13/14 gen. In your personal opinion, would you hesitate at all to buy a 13/14 gen CPU at the moment?
I have a 13900ks. Purchased it when it first released. I first ran it stock and under single thread load in cinebench r23 it was showing vcore of 1.412-1.450 at 6.0ghz. Then multicore was using around 340watts at 100c. So after a little fiddling in bios and delidded cpu it runs in high 60c range at 275watts multicore workload. Voltage is around 1.19 or so with performance cores locked at 5.6 and ecores at 4.5. It works very well so far.
 
12th gen is not affected by oxidation/degradation. Unlock it's full potential and do not be worried. I've been running one of those at 5.3/4.1/4.7/4200 ram 1.35/1.55V ram since day one and haven't noticed any degradation.
Well, I had thought about Bartlett Lake... that is until I saw THIS. Are the 13/14 gen chips really this bad, or are people blowing things out of proportion?
Source 1 says:
Source 2 says:
Yeah and my dad works at Nintendo. That guy is usually full of ****. My advise to you good sir is to wait and see.
 
I'd like to sell my 12600kf, and upgrade to something like a 13700k/kf to extend the life of the LGA 1700 platform, but I'm just not sure what to think of a lot of the stuff I've read lately about 13/14 gen. In your personal opinion, would you hesitate at all to buy a 13/14 gen CPU at the moment?
No, I wouldn't hesitate.

I would wait to see what Bartlett Lake brings about before upgrading, especially if you're willing to spend more. I don't anticipate Bartlett Lake to be high value, and I wouldn't see a point in buying less than the 12core/24thread highest SKU.
 
Discussion starter · #20 ·
No, I wouldn't hesitate.

I would wait to see what Bartlett Lake brings about before upgrading, especially if you're willing to spend more. I don't anticipate Bartlett Lake to be high value, and I wouldn't see a point in buying less than the 12core/24thread highest SKU.
Waiting for Bartlett Lake is kinda what I'm debating on in my head... I'm on a fixed income these days, so I have to spend wisely and try to get the most bang for my buck... That's why I'd like to extend the useful life of this platform for as long as possible. My next platfom upgrade will have to include CPU/MB as well as RAM, because I'm still on DDR4. I'd say for CPU upgrade alone, somewhere around $325 USD give/take a bit would be about what I'd be willing to spend, then turn around and sell my 12600KF to recoup some of the upgrade costs. Eventually, further down the line, I'd like to upgrade my PSU, GPU, and monitor to a 1440p capable rig, but I'd say that would be in the next 2-3 years or so... between my next CPU and full platform upgrade. After upgrading CPU I'd like to shoot for 5+ years out of my current platform.
 
1 - 20 of 22 Posts