Official Clevo NH55JNNQ / NH5xJNxx 12th gen Socket 1700 LGA laptop thread

[win32asmguy]

3 hours ago, electrosoft said:

 

Problem is right now Intel is hoarding a lot of decent bins for 12900ks but most 12900k's will be ~SP84'ish in silicon quality which is not great. If you had an Asus board that supported SP rating (Strix D4, Prime A, Hero) you would probably see the ratings are ho hum at best.

 

I binned 5 and really it was random luck I landed a chip that ran ridiculously low and cool. Most 12900ks chips will blow it out the water for overclocking though. Testing it on that B660M board with that small air cooler and seeing it only run ~5c hoter than a 12100f while pulling 60w more was astounding. I didn't even expect those types of results.

 

Is it a thermal issue or a PL issue? If it is thermal issues, a delid and relid with a good IHS will work wonders. I know everybody is doing Rockitcool, but I would recommend BartX. Not only can you get a nickel plated IHS (no soak issues with LM), but he will custom make an IHS for you with additional height if needed for a tighter pairing.

 

I've purchased a few custom made IHS's from him and they worked great. One for 9900ks and another for 10900k.

Maybe the low pull 12900k are also going to 12900hx BGA packaged chips, too? Shame Intel doesn't have an LGA SKU targeting that type of bin for HTPC type purposes.

 

I could exchange the 12900k for a 12900ks at Micro Center. They have them for $180 more.

 

Also what is considered the best non-LM paste for IHS to heatsink applications? I am running low on Thermalright TFX so looking at ordering more of something. I have heard Kryonaut degrades above 80C so not considering that even though Micro Center carries it.

[jaybee83]

On 5/23/2022 at 5:32 PM, win32asmguy said:

I take it you are trying to flash via FPTW? Or using a SPI programmer directly? FPTW would need Config lock disabled, and it sometimes takes multiple applications + reboots to get the setting to stick. Signing certificates are probably for secure boot.

double post 🙄

[jaybee83]

20 minutes ago, win32asmguy said:

Maybe the low pull 12900k are also going to 12900hx BGA packaged chips, too? Shame Intel doesn't have an LGA SKU targeting that type of bin for HTPC type purposes.

 

I could exchange the 12900k for a 12900ks at Micro Center. They have them for $180 more.

 

Also what is considered the best non-LM paste for IHS to heatsink applications? I am running low on Thermalright TFX so looking at ordering more of something. I have heard Kryonaut degrades above 80C so not considering that even though Micro Center carries it.

havent had any issues with kryonaut, maybe u could give kryonaut extreme a try 🙂 otherwise, MX5 is supposed to be good, as well as noctua NT-H2 and coolermaster master gel pro v2

[electrosoft]

1 hour ago, win32asmguy said:

Maybe the low pull 12900k are also going to 12900hx BGA packaged chips, too? Shame Intel doesn't have an LGA SKU targeting that type of bin for HTPC type purposes.

 

I could exchange the 12900k for a 12900ks at Micro Center. They have them for $180 more.

 

Also what is considered the best non-LM paste for IHS to heatsink applications? I am running low on Thermalright TFX so looking at ordering more of something. I have heard Kryonaut degrades above 80C so not considering that even though Micro Center carries it.

 

1 hour ago, jaybee83 said:

havent had any issues with kryonaut, maybe u could give kryonaut extreme a try 🙂 otherwise, MX5 is supposed to be good, as well as noctua NT-H2 and coolermaster master gel pro v2

 

I would not be surprised if they are binning for 12900hx specs too as they will need a much tighter and lower V/F curve for the lower range and really don't need to hit full 12900k specs. It could go both ways.

 

You could also bin 12900kf as they can never be binned for 12900hx or 12900ks chips as they don't have iGFX.

 

If you have any type of potential pairing/pressure issues, nanogrease is the best as it is the thickest and helps compensate for any pairing issues while being within spitting distance of the slightly better thermal compounds.

 

If you're confident in your pairing pressure, Kingpin is best I've found but it has a ~6mo shelf life before it starts to degrade followed by good ole gelid extreme and then EVGA Frostbite is a sleeper hit which is on par with Gelid extreme.

 

Personally, I run Nanogrease Extreme on my laptops because I've yet to find a laptop that pairs anywhere near as exacting and tight as a desktop and it is close enough and tends to give me better temps overall. I don't use it on desktops at all.

 

I lost confidence in Kryonaut and Conductonaut due to some bad batches so I avoid both for TIM and LM for YMMV.

 

 

 

 

[GT3000]

Kryonaut Extreme works great personally. I think the issue with it is people have said that over a long period of time it has degradation effect. I don't think it's been available long enough to test that. I know regular kryonaut suffered with breakdown over 80C. Kingpin KPX is also quite good but not sure of the long term viability.

[Clamibot]

6 hours ago, win32asmguy said:

Maybe the low pull 12900k are also going to 12900hx BGA packaged chips, too? Shame Intel doesn't have an LGA SKU targeting that type of bin for HTPC type purposes.

 

I could exchange the 12900k for a 12900ks at Micro Center. They have them for $180 more.

 

Also what is considered the best non-LM paste for IHS to heatsink applications? I am running low on Thermalright TFX so looking at ordering more of something. I have heard Kryonaut degrades above 80C so not considering that even though Micro Center carries it.

 

In my experience, Phobya Nanogrease Extreme is the best thermal paste. It doesn't degrade and doesn't pump out. Additionally, it's more thermally conductive than almost all other thermal pastes. Pretty much the only better TIM is liquid metal.

 

I hear Thermalright TFX is similar to Nanogrease, but I'm sticking to Nanogrease since I've used it and it has a proven track record (for me personally) of yielding excellent temps. I can't confirm if TFX is like Nanogrease since I've never used it.

 

I've used Arctic MX-4, Thermaly Grizzly Kryonaut, and some Cooler Master paste (I'm not sure if it was MasterGel or something else). Nanogrease significantly outperforms all of them.

 

I'm not sure why Arctic MX-4 is recommended by many others because it sucks. Temps are horrible with it.

 

Kryonaut pumps out if you're doing direct die with it. If not, then it's fine, but there's still the problem of it burning up if subjected to temperatures above 80°C.

 

Cooler Master's pastes seem good for longevity, but as I mentioned before Nanogrease is the best. I've never had to repaste Nanogrease either because it doesn't degrade or pump out.

 

So in conclusion, go with Thermalright TFX or Phobya Nanogrease if you want to avoid liquid metal. @Mr. Fox can confirm the quality of Nanogrease as he's had very good experiences with Nanogrease as well.

[GT3000]

Here's a good video on the most recent pastes that should be helpful.

 

[Mr. Fox]

The problem with most thermal paste reviews or "showdowns" is they are difficult to draw any meaningful conclusions from no matter how diligent the entity is that is performing the comparison. In the simplest of terms, thermal compounds are nothing more than thermally conductive gap fillers.

 

The first problem is that such reviews are almost always done using desktops with heat sinks or water blocks that fit extremely well. No conclusions can be drawn in terms of their application on something as sloppy as a notebook computer. There is tremendous variance in fit between models and brands, and the wide variety in quality control, (or the evident lack of it,) makes it likely for two identical laptops to behave differently. A thermal compound that works fantastic on the laptop for Bobby might totally suck for Jimmy even though they both own the same exact product. 

 

The second problem is a showdown only gives you an understanding of how well a thermal compound performs on one machine at the time of fresh application. You have no clue, and neither does the tester, if the temperatures are going to be the same, better, or worse in a few days or a couple of months. The only way he would understand that or have any way to demonstrate it is if the exercise took several years to complete because you allowed a couple of months of data collection per product. Nonetheless, results are going to be all over the board among individual end users.

 

Now, these product comparisons do provide value to the extent they demonstrate how well a thermal compound works when first applied in controlled circumstances where there are fewer opportunities for errors caused by variables such as fit and application.

 

The ultimate test of which thermal solution(s) work best rests on the shoulders of the computer owner using a specific machine under unique circumstances.

 

TL;DR: YMMV.  The best one is the one that works best. If your temperatures are truly good, be happy. Don't over-think the situation or start a new religious technology cult based on it. Most notebook owners can't say that because theirs are not. Whether they even know or care how their system is or isn't performing is another topic.

[GT3000]

Yep, agree. YMMV. They're roughly the same so pick a few and see what fits your situation the best. Anything beats stock paste usuAllly.

[jaybee83]

On 5/25/2022 at 8:18 AM, Mr. Fox said:

The problem with most thermal paste reviews or "showdowns" is they are difficult to draw any meaningful conclusions from no matter how diligent the entity is that is performing the comparison. In the simplest of terms, thermal compounds are nothing more than thermally conductive gap fillers.

 

The first problem is that such reviews are almost always done using desktops with heat sinks or water blocks that fit extremely well. No conclusions can be drawn in terms of their application on something as sloppy as a notebook computer. There is tremendous variance in fit between models and brands, and the wide variety in quality control, (or the evident lack of it,) makes it likely for two identical laptops to behave differently. A thermal compound that works fantastic on the laptop for Bobby might totally suck for Jimmy even though they both own the same exact product. 

 

The second problem is a showdown only gives you an understanding of how well a thermal compound reforms on one machine at the time of fresh application. You have no clue, and neither does the tester, if the temperatures are going to be the same, better, or worse in a few days or a couple of months. The only way he would understand that or have any way to demonstrate it is if the exercise took several years to complete because you allowed a couple of months of data collection per product. Nonetheless, results are going to be all over the board among individual end users.

 

Now, these product comparisons do provide value to the extent they demonstrate how well a thermal compound works when first applied in controlled circumstances where there are fewer opportunities for errors caused by variables such as fit and application.

 

The ultimate test of which thermal solution(s) work best rests on the shoulders of the computer owner using a specific machine under unique circumstances.

 

TL;DR: YMMV.  The best one is the one that works best. If your temperatures are truly good, be happy. Don't over-think the situation or start a new religious technology cult based on it. Most notebook owners can't say that because theirs are not. Whether they even know or care how their system is or isn't performing is another topic.

nice conclusion, laptops just SUCK too much to do any proper paste comparisons 😂 i love it haha

[win32asmguy]

On 5/24/2022 at 3:15 PM, electrosoft said:

I would not be surprised if they are binning for 12900hx specs too as they will need a much tighter and lower V/F curve for the lower range and really don't need to hit full 12900k specs. It could go both ways.

 

You could also bin 12900kf as they can never be binned for 12900hx or 12900ks chips as they don't have iGFX.

 

If you have any type of potential pairing/pressure issues, nanogrease is the best as it is the thickest and helps compensate for any pairing issues while being within spitting distance of the slightly better thermal compounds.

 

If you're confident in your pairing pressure, Kingpin is best I've found but it has a ~6mo shelf life before it starts to degrade followed by good ole gelid extreme and then EVGA Frostbite is a sleeper hit which is on par with Gelid extreme.

 

Personally, I run Nanogrease Extreme on my laptops because I've yet to find a laptop that pairs anywhere near as exacting and tight as a desktop and it is close enough and tends to give me better temps overall. I don't use it on desktops at all.

 

I lost confidence in Kryonaut and Conductonaut due to some bad batches so I avoid both for TIM and LM for YMMV.

Unfortunately a 12900kf is a no go as the laptop screen needs the iGFX.

 

I got a Newegg 12900k in, this one is slightly faster than the others. CBR23 multi 21891 without an undervolt. This one is the first I have tested that is an older batch, V141 as opposed to all of the others which are X202+. With this one wattage peaks at 129.1, so a bit higher than the others. I guess that means it has a higher VF curve if the EC has a 90A limit.

 

I am unsure about measuring the pressure. It does seem to be less than what I have seen with the NH-D15S on my desktop. The max temp range has generally been no more than 5C on the P-cores in situations when its not thermal throttling (max fans, P+e cores enabled, on a laptop stand, ambient temp 65F). I have used nanogrease extreme before but it was pretty viscous. When you apply it do you just do a bead on the IHS, or try to spread with the spatula?

[jaybee83]

52 minutes ago, win32asmguy said:

Unfortunately a 12900kf is a no go as the laptop screen needs the iGFX.

 

I got a Newegg 12900k in, this one is slightly faster than the others. CBR23 multi 21891 without an undervolt. This one is the first I have tested that is an older batch, V141 as opposed to all of the others which are X202+. With this one wattage peaks at 129.1, so a bit higher than the others. I guess that means it has a higher VF curve if the EC has a 90A limit.

 

I am unsure about measuring the pressure. It does seem to be less than what I have seen with the NH-D15S on my desktop. The max temp range has generally been no more than 5C on the P-cores in situations when its not thermal throttling (max fans, P+e cores enabled, on a laptop stand, ambient temp 65F). I have used nanogrease extreme before but it was pretty viscous. When you apply it do you just do a bead on the IHS, or try to spread with the spatula?

same here with core to core temp spread, im seeing roughly 5C difference. ive tried both manual spread with spatula and "cross" method, i think the latter is more suitable for the low mounting pressure in a laptop. the manual spatula spread is ideal on desktop because it ensures that every corner is truly and completely covered. the rest is done by the high mounting pressure and mostly even surface fit between the cooler and ihs. in the laptop, however, the manual spread cannot ensure balancing out any potential (but likely, in this context) uneven fit between cold plate and ihs, so the cross method i think provides the most ideal approach in filling all gaps. blob / pea method is not bad, per se, but based on experience leaves a lot to be desired in terms of reaching the last bit of the ihs corners. the only downside to the cross is that u might end up with a bit of paste spillage over the edges, but if its non conductive no worries, just a bit more cleaning up to do during next repaste 😋

[electrosoft]

2 hours ago, win32asmguy said:

Unfortunately a 12900kf is a no go as the laptop screen needs the iGFX.

 

I got a Newegg 12900k in, this one is slightly faster than the others. CBR23 multi 21891 without an undervolt. This one is the first I have tested that is an older batch, V141 as opposed to all of the others which are X202+. With this one wattage peaks at 129.1, so a bit higher than the others. I guess that means it has a higher VF curve if the EC has a 90A limit.

 

I am unsure about measuring the pressure. It does seem to be less than what I have seen with the NH-D15S on my desktop. The max temp range has generally been no more than 5C on the P-cores in situations when its not thermal throttling (max fans, P+e cores enabled, on a laptop stand, ambient temp 65F). I have used nanogrease extreme before but it was pretty viscous. When you apply it do you just do a bead on the IHS, or try to spread with the spatula?

 

Most likely the first two were in the same "average" mediocre range and you finally got a chip that is a better silicon. V variants are from China while X variants are from Vietnam. When I say 12th gen variance is extreme that is an understatement. Even at stock you're looking at end tweaking potential pull variances of ~70-75w (~162w-241w). Throw in leaky vs non-leaky chips and when you're dealing with a limited cooling solution it can make a difference.

 

If you settle on that chip, a delid and milled, flat IHS will probably work wonders.

 

You could also do some shim tests to see if extra height gives better temps and performance. I know on my P870TM ordering an IHS that was 4.3mm vs stock 3.3mm made a world of difference. Doing shim tests on my X170SM showed it didn't have nearly the same effect but my X170SM-G is modded from Ztec so a lot of the ground work was done already.

 

Best way to apply nanogrease extreme is a dollop in the middle and let the natural pairing pressure spread it out where it needs to be and compensate for any gaps. Due to the rectangular shape of the IHS you may need to apply a micro sized dot above and below the central dollop. You can always do a few test applications with cheaper stuff to make sure you're getting the ideal spread.

 

 

[jaybee83]

issue is that all cpu and gpu heatsink screws are equipped with springs, thus u cant "overtighten" them but u can also not tighten them properly 😛 there have been some previous user tests where removing and replacing those springed screws with regular ones can significantly increase the cold plate pressure on the IHS.

naturally, ull then have to be cautious not to overtighten / strip those replacement screws! 😅

[electrosoft]

On 5/26/2022 at 3:03 PM, jaybee83 said:

issue is that all cpu and gpu heatsink screws are equipped with springs, thus u cant "overtighten" them but u can also not tighten them properly 😛 there have been some previous user tests where removing and replacing those springed screws with regular ones can significantly increase the cold plate pressure on the IHS.

naturally, ull then have to be cautious not to overtighten / strip those replacement screws! 😅

 

Ztec did this to their X170's to increase pressure. They removed the stock spring loaded screws with normal screws. I ended up replacing even those with even more dialed in screws to get exacting pressure to my standards. The trick as you mention is to not overtighten them. Too much pressure and the system will not boot or it will crash as soon as you introduce any type of load that heats up the CPU and the socket/chip expands ever so slightly. I found the right pressure with a "loose hand" screw then finger tips tightening.

 

Get a power screw driver with controllable torque and you can find the exact setting that works the best and know for future heatsink removals you have an exact setting. I might take this approach in the future. Takes a bit to dial in the exact settings through trial and error but them you know exactly what works the best and at what setting.

 

 

 

[jaybee83]

4 hours ago, electrosoft said:

 

Ztec did this to their X170's to increase pressure. They removed the stock spring loaded screws with normal screws. I ended up replacing even those with even more dialed in screws to get exacting pressure to my standards. The trick as you mention is to not overtighten them. Too much pressure and the system will not boot or it will crash as soon as you introduce any type of load that heats up the CPU and the socket/chip expands ever so slightly. I found the right pressure with a "loose hand" screw then finger tips tightening.

 

Get a power screw driver with controllable torque and you can find the exact setting that works the best and know for future heatsink removals you have an exact setting. I might take this approach in the future. Takes a bit to dial in the exact settings through trial and error but them you know exactly what works the best and at what setting.

 

 

 

i like your style 🤠😁

[win32asmguy]

On 5/26/2022 at 12:36 PM, electrosoft said:

Most likely the first two were in the same "average" mediocre range and you finally got a chip that is a better silicon. V variants are from China while X variants are from Vietnam. When I say 12th gen variance is extreme that is an understatement. Even at stock you're looking at end tweaking potential pull variances of ~70-75w (~162w-241w). Throw in leaky vs non-leaky chips and when you're dealing with a limited cooling solution it can make a difference.

 

If you settle on that chip, a delid and milled, flat IHS will probably work wonders.

 

You could also do some shim tests to see if extra height gives better temps and performance. I know on my P870TM ordering an IHS that was 4.3mm vs stock 3.3mm made a world of difference. Doing shim tests on my X170SM showed it didn't have nearly the same effect but my X170SM-G is modded from Ztec so a lot of the ground work was done already.

 

Best way to apply nanogrease extreme is a dollop in the middle and let the natural pairing pressure spread it out where it needs to be and compensate for any gaps. Due to the rectangular shape of the IHS you may need to apply a micro sized dot above and below the central dollop. You can always do a few test applications with cheaper stuff to make sure you're getting the ideal spread.

What is the best way to determine how leaky a chip is?

 

It would be nice if I could find a 12900k that is better than the new HX BGA chips. Although with the vapor chamber and higher power limits that may be a tall order. Still need to post that WTB for a good chip. If I offered $1000 I could probably find someone willing to part with one.

 

It would be even better if Eurocom would publish a bios update with the MUX and undervolting available. Although at the moment they don't even have the Raptor X15 listed on their site so maybe stock of all configurations has dried up.

[jaybee83]

2 hours ago, win32asmguy said:

What is the best way to determine how leaky a chip is?

 

It would be nice if I could find a 12900k that is better than the new HX BGA chips. Although with the vapor chamber and higher power limits that may be a tall order. Still need to post that WTB for a good chip. If I offered $1000 I could probably find someone willing to part with one.

 

It would be even better if Eurocom would publish a bios update with the MUX and undervolting available. Although at the moment they don't even have the Raptor X15 listed on their site so maybe stock of all configurations has dried up.

oh wow that was fast with stock drying up....hope this is due to the machine being so popular and not because their stock was super low to begin with....

no updates yet from my side btw, but will be sure to let u guys know once smth is up 🙂

[win32asmguy]

12 hours ago, jaybee83 said:

oh wow that was fast with stock drying up....hope this is due to the machine being so popular and not because their stock was super low to begin with....

no updates yet from my side btw, but will be sure to let u guys know once smth is up 🙂

Maybe they just had a website issue and had to restore from a backup. It is back again today.

 

On that note, I should tell Eurocom the Raptor X15 can boot the 12900ks just fine.

[electrosoft]

16 hours ago, win32asmguy said:

What is the best way to determine how leaky a chip is?

 

It would be nice if I could find a 12900k that is better than the new HX BGA chips. Although with the vapor chamber and higher power limits that may be a tall order. Still need to post that WTB for a good chip. If I offered $1000 I could probably find someone willing to part with one.

 

It would be even better if Eurocom would publish a bios update with the MUX and undervolting available. Although at the moment they don't even have the Raptor X15 listed on their site so maybe stock of all configurations has dried up.

 

By temps usually especially with the X15 as you are hitting thermal throttling in some instances. That last 12900k you picked up was able to pull a little more possibly because it is a better silicon sample and can pull a bit more because it isn't throttling at 125w but >125w instead. Same reason @jaybee83 might be able to pull 140w with his 12600 since it isn't as thermally limited at 125w.

 

1 hour ago, win32asmguy said:

Maybe they just had a website issue and had to restore from a backup. It is back again today.

 

On that note, I should tell Eurocom the Raptor X15 can boot the 12900ks just fine.

 

Nice! How is the 12900ks running in there versus the other 12900k's you've tested? I'm glad you have decently deep pockets to extend these tests to this level.

 

Keep everything at stock and start setting some baselines and make a chart for UVs and see how they fair on a CB23 run (temp and pull wise) at -0.00, -0.50, -0.75, -0.100 (anything past this is gravy). You can also do some Cb23 10min runs to see how they handle sustained runs.

 

 

[jaybee83]

12 hours ago, electrosoft said:

 

By temps usually especially with the X15 as you are hitting thermal throttling in some instances. That last 12900k you picked up was able to pull a little more possibly because it is a better silicon sample and can pull a bit more because it isn't throttling at 125w but >125w instead. Same reason @jaybee83 might be able to pull 140w with his 12600 since it isn't as thermally limited at 125w.

 

 

Nice! How is the 12900ks running in there versus the other 12900k's you've tested? I'm glad you have decently deep pockets to extend these tests to this level.

 

Keep everything at stock and start setting some baselines and make a chart for UVs and see how they fair on a CB23 run (temp and pull wise) at -0.00, -0.50, -0.75, -0.100 (anything past this is gravy). You can also do some Cb23 10min runs to see how they handle sustained runs.

 

 

rockit cool kit is still in the US, but making slow/steady progress towards me 😋 reaaaally curious as to what the max sustained wattage will/can be with sufficiently lowered temps. ideally with UV enabled via unicorn magic 🤠

 

doesnt surprise me that the 12900ks also works, since were not talking a different cpu here, just a different label haha.

 

@win32asmguy dude, youre really taking the fall here to push community findings on this machine, really appreciate it. hope ill also be able to contribute more soon 😊

 

@electrosoft dunno how much time / motivation u have for it, but maybe it would make sense to provide some links in the OP to the most important posts here in the thread with findings on clocks, temps, tweaking headroom, etc?

[win32asmguy]

17 hours ago, electrosoft said:

 

By temps usually especially with the X15 as you are hitting thermal throttling in some instances. That last 12900k you picked up was able to pull a little more possibly because it is a better silicon sample and can pull a bit more because it isn't throttling at 125w but >125w instead. Same reason @jaybee83 might be able to pull 140w with his 12600 since it isn't as thermally limited at 125w.

 

 

Nice! How is the 12900ks running in there versus the other 12900k's you've tested? I'm glad you have decently deep pockets to extend these tests to this level.

 

Keep everything at stock and start setting some baselines and make a chart for UVs and see how they fair on a CB23 run (temp and pull wise) at -0.00, -0.50, -0.75, -0.100 (anything past this is gravy). You can also do some Cb23 10min runs to see how they handle sustained runs.

 

 

The 12900ks is not so great so far in testing.

 

20736 at 0mv

21766 at -50mv

21983 at -75mv

22178 at -100mv

 

All pull around 136W when under load. At -100mv it can make it through a run without thermal throttle. This one was also pasted with TG Hydronaut because that was all they had a Micro Center. Nanogrease should arrive this week which may help.

[jaybee83]

1 hour ago, win32asmguy said:

The 12900ks is not so great so far in testing.

 

20736 at 0mv

21766 at -50mv

21983 at -75mv

22178 at -100mv

 

All pull around 136W when under load. At -100mv it can make it through a run without thermal throttle. This one was also pasted with TG Hydronaut because that was all they had a Micro Center. Nanogrease should arrive this week which may help.

so youre also getting close to that 140W peak that i saw. whats your hwinfo monitoring resolution? i set it to the fastest 50ms polling to get a better idea of what the actual temp/clock/wattage ranges are.

[electrosoft]

1 hour ago, win32asmguy said:

The 12900ks is not so great so far in testing.

 

20736 at 0mv

21766 at -50mv

21983 at -75mv

22178 at -100mv

 

All pull around 136W when under load. At -100mv it can make it through a run without thermal throttle. This one was also pasted with TG Hydronaut because that was all they had a Micro Center. Nanogrease should arrive this week which may help.

 

You can definitely see the uv definitely giving some head room. When you run the same range of uv's on your other 12900k's what range of scores do you achieve? Do the other 12900k's also make it through a run without thermal throttling at -100mv?

 

You're reaching that point where your next step is a delid with a milled, flat (and probably ever so slightly taller) copper IHS. Delids on 12th (especially 12900k) have consistently shown. The idea now is to separate (even if only by the tiniest amount) the best of your group of 12900k's and delid that one and install a proper IHS.

 

I can't recall if you did a spread test (dollop of compound, center application, install heatsink, run a few CB23 runs to heat it up and then remove heatsink to see how it naturally spread)?

 

 

[electrosoft]

7 hours ago, jaybee83 said:

rockit cool kit is still in the US, but making slow/steady progress towards me 😋 reaaaally curious as to what the max sustained wattage will/can be with sufficiently lowered temps. ideally with UV enabled via unicorn magic 🤠

 

doesnt surprise me that the 12900ks also works, since were not talking a different cpu here, just a different label haha.

 

@win32asmguy dude, youre really taking the fall here to push community findings on this machine, really appreciate it. hope ill also be able to contribute more soon 😊

 

@electrosoft dunno how much time / motivation u have for it, but maybe it would make sense to provide some links in the OP to the most important posts here in the thread with findings on clocks, temps, tweaking headroom, etc?

 

I was waiting till we have more substantive and definitive data and links. Right now, we're knee deep in the fact finding portion of the program. 😁