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

[1610ftw]

10 hours ago, jaybee83 said:

haha i can tell you right now, ill likely never get a 12900K, im just waiting for Raptor Lake to come out and then jump to 13900K right away 😛 

 

I knew it - not nice of you to be so cruel but you can make up for it by convincing Prema to release his work at some point so that others can get more out or their 12900K 😄

[jaybee83]

3 hours ago, 1610ftw said:

 

I knew it - not nice of you to be so cruel but you can make up for it by convincing Prema to release his work at some point so that others can get more out or their 12900K 😄

dude, you have noooo idea what u just said 😂 so lemme be a tease once more: IF we are successful with this mod and it brings ACTUAL performance and functionality advantages to the table, then there MIGHT be some vendors who would partner up with the unicorn rainbow master on this for our machines 🤪 all of this purely hypothetical and a total mind construct, naturally 😉

[1610ftw]

2 hours ago, jaybee83 said:

dude, you have noooo idea what u just said 😂 so lemme be a tease once more: IF we are successful with this mod and it brings ACTUAL performance and functionality advantages to the table, then there MIGHT be some vendors who would partner up with the unicorn rainbow master on this for our machines 🤪 all of this purely hypothetical and a total mind construct, naturally 😉

Completely understood, very hypothetically speaking of course and not that I would have any specific vendors in mind 😄

 

[electrosoft]

18 hours ago, jaybee83 said:

haha i can tell you right now, ill likely never get a 12900K, im just waiting for Raptor Lake to come out and then jump to 13900K right away 😛 

 

LOL, my tentative plan is when RPL drops to pick one up and drop it in whatever is my current MB and track down a used NH55 and drop my 12900k in it especially after much of the issues have been fleshed out / documented.

 

 

[win32asmguy]

On 7/11/2022 at 5:16 PM, electrosoft said:

Good to hear you just opened a case and wash your hands of that mess. The more I think of it, unless the box/package was absolutely smashed to oblivion, there is no way the IHS would be removed AND smd's sheered off. I sense shenanigans on the seller's part.

Got another chip on the way now. This one is a 12900ks, SP91 overall, P-core SP100, e-core SP75. Already delidded with the copper IHS glued down.

[electrosoft]

52 minutes ago, win32asmguy said:

Got another chip on the way now. This one is a 12900ks, SP91 overall, P-core SP100, e-core SP75. Already delidded with the copper IHS glued down.

 

Nice! Hopefully it will pop right in and be better from the jump especially being delidded and relidded with a copper IHS in place.

 

Here's some data from my 12900ks I sold to a buddy. It was an SP94 P102 E79 and it was clearly better than my 12900k on the high end about a bin better on an AIO overall but when I was going for lowest pull and temps it bottomed out at ~169w and 1.119v under load CB23 10min run for testing with a .050uv.

 

For a single run it bottomed out at 164w with a 0.75uv but crashed 2-3 runs on CB23.

 

If it had been better on the low end, I would have held onto it but it's in his loop now so all is well that ends well.

 

 

 

 

 

 

[win32asmguy]

52 minutes ago, electrosoft said:

Nice! Hopefully it will pop right in and be better from the jump especially being delidded and relidded with a copper IHS in place.

 

Here's some data from my 12900ks I sold to a buddy. It was an SP94 P102 E79 and it was clearly better than my 12900k on the high end about a bin better on an AIO overall but when I was going for lowest pull and temps it bottomed out at ~169w and 1.119v under load CB23 10min run for testing with a .050uv.

 

For a single run it bottomed out at 164w with a 0.75uv but crashed 2-3 runs on CB23.

 

If it had been better on the low end, I would have held onto it but it's in his loop now so all is well that ends well.

Yep. Did you mention if there was a specific way to test how leaky a chip is? I think I found that certain desktop boards have dedicated voltage regular sensors that more accurately report how much voltage and current is being supplied. So the more leaky chip would draw higher current at the same voltage and load, correct?

[electrosoft]

2 minutes ago, win32asmguy said:

Yep. Did you mention if there was a specific way to test how leaky a chip is? I think I found that certain desktop boards have dedicated voltage regular sensors that more accurately report how much voltage and current is being supplied. So the more leaky chip would draw higher current at the same voltage and load, correct?

 

Once I get ahold of at least three chips I start drawing correlations between pull and temps. More chips equal better data visualization due to increased sample size.

 

A leaky chip set to the same frequency would draw higher current and produce more heat at the same load and may require higher Vcore.

 

For me, leaky means higher power consumption at the same frequency and load setting which tends to also correlate to higher temps. For example two 10900k chips set to 5.3ghz. One of them draws ~264w and the other 330w+. Exact same conditions and settings. That is an extremely leaky and hungry chip which is also a furnace. The byproduct in a constrained setting thermally (like a laptop) is an instant disaster. For overclocking, that leak does sometimes let those chips handle higher voltage to push even further (as long as you have the cooling), but for our purposes leaky chips are not a good thing as you want less leakage, less draw and subsequently less voltage and heat to maximize dissipation via the smaller heatsink as much as possible.

 

When I bin on the desktop I always monitor voltage, package power and heat under a fixed load (CB23 in this case) to draw baselines and after 3+ chips you usually can start to differentiate which chips run hotter, which are hungrier and which are both (usually bottom of the barrel chips for targeted purposes).

 

You've tested enough 12900k/ks on your laptop now to see one of them float to the top. On a desktop board you would have much more control to more accurately differentiate between them but in the end the ultimate test is setting like conditions with your laptop and testing them and seeing which one(s) are clearly better and give you more thermal headroom to boost higher.

 

 

[win32asmguy]

22 hours ago, electrosoft said:

Once I get ahold of at least three chips I start drawing correlations between pull and temps. More chips equal better data visualization due to increased sample size.

 

A leaky chip set to the same frequency would draw higher current and produce more heat at the same load and may require higher Vcore.

 

For me, leaky means higher power consumption at the same frequency and load setting which tends to also correlate to higher temps. For example two 10900k chips set to 5.3ghz. One of them draws ~264w and the other 330w+. Exact same conditions and settings. That is an extremely leaky and hungry chip which is also a furnace. The byproduct in a constrained setting thermally (like a laptop) is an instant disaster. For overclocking, that leak does sometimes let those chips handle higher voltage to push even further (as long as you have the cooling), but for our purposes leaky chips are not a good thing as you want less leakage, less draw and subsequently less voltage and heat to maximize dissipation via the smaller heatsink as much as possible.

 

When I bin on the desktop I always monitor voltage, package power and heat under a fixed load (CB23 in this case) to draw baselines and after 3+ chips you usually can start to differentiate which chips run hotter, which are hungrier and which are both (usually bottom of the barrel chips for targeted purposes).

 

You've tested enough 12900k/ks on your laptop now to see one of them float to the top. On a desktop board you would have much more control to more accurately differentiate between them but in the end the ultimate test is setting like conditions with your laptop and testing them and seeing which one(s) are clearly better and give you more thermal headroom to boost higher.

Yep that is a big difference between those 10900k chips. Its a shame that this metric tends to go unnoticed by most of the Clevo DTR resellers, and they force you to buy some kind of throw away processor regardless (unless it was sold by RJTech as a barebone). Maybe some people get lucky and are randomly shipped a great chip, but I otherwise feel its wasteful to pair a DTR with poor silicon, especially since its one of the few machines on the market where you CAN do something about it.

 

I went ahead and ordered a Strix Z690 D4, LGA1700 bracket for my NH-D15S, and a second 120mm fan. I figure that way I can better compare things without having the NH55 specific limitations skewing results.

 

Not sure if microcode injection is going to be possible for raptor lake if modded bioses cannot be flashed on the NH55, but maybe if Jaybee and Prema keep working at it they may find a way. Sadly I doubt I could ever get a modded bios from Prema for mine, as I bought it from Eurocom. It is stupid that they are the only North America reseller for the NH55J... their support also has stopped responding to emails about my open ticket for the wrong keyboard they sent. Not exactly happy with their service.

[jaybee83]

21 minutes ago, win32asmguy said:

Yep that is a big difference between those 10900k chips. Its a shame that this metric tends to go unnoticed by most of the Clevo DTR resellers, and they force you to buy some kind of throw away processor regardless (unless it was sold by RJTech as a barebone). Maybe some people get lucky and are randomly shipped a great chip, but I otherwise feel its wasteful to pair a DTR with poor silicon, especially since its one of the few machines on the market where you CAN do something about it.

 

I went ahead and ordered a Strix Z690 D4, LGA1700 bracket for my NH-D15S, and a second 120mm fan. I figure that way I can better compare things without having the NH55 specific limitations skewing results.

 

Not sure if microcode injection is going to be possible for raptor lake if modded bioses cannot be flashed on the NH55, but maybe if Jaybee and Prema keep working at it they may find a way. Sadly I doubt I could ever get a modded bios from Prema for mine, as I bought it from Eurocom. It is stupid that they are the only North America reseller for the NH55J... their support also has stopped responding to emails about my open ticket for the wrong keyboard they sent. Not exactly happy with their service.

that sounds quite sketchy with eurocom dude....keep spamming them for that ticket, dont give up!

 

well be sure to keep working on this until all options are exhausted. will keep u guys updated, of course.

 

greetz from croatia btw, another vacay *lol* 

[Tenoroon]

@jaybee83Do these models support Prema BIOS from a partner? I have a buddy who is looking into DTR's and I'm sure he would be interested in grabbing one of these if Prema has worked with a partner 🙂

[jaybee83]

17 hours ago, Tenoroon said:

@jaybee83Do these models support Prema BIOS from a partner? I have a buddy who is looking into DTR's and I'm sure he would be interested in grabbing one of these if Prema has worked with a partner 🙂

thanks for showing interest mate! the unicorn magic master and i are currently working on a mod for this machine, well have to figure out what kinda performance and functionality advantages we can achieve. based on how successful we are, there is indeed a tentative plan to release it to partners. "people" are being kept in the loop, but its still early days.

 

in any case, ill keep the guys here updated on the progress 🙂

 

speaking of which! we mightve gotten rid of the bug that kept us from properly signing and flashing modded firmwares, next test built is ready to be flashed 😁👌🏼 "unfortunately" *cough*, i still got a week of vacay here in croatia ahead of me and "Alfred" is waiting at home 😋

[win32asmguy]

Finally got some data for my 12900k that correlates why it can maintain higher clocks in the NH55JNNQ. It seems like even though its a poor SP rating, the VF points are lower at 4200mhz and below. I guess these numbers do not indicate how leaky it is, but it is something.

 

[electrosoft]

8 hours ago, win32asmguy said:

Finally got some data for my 12900k that correlates why it can maintain higher clocks in the NH55JNNQ. It seems like even though its a poor SP rating, the VF points are lower at 4200mhz and below. I guess these numbers do not indicate how leaky it is, but it is something.

 

 

That makes a lot of sense at that V/F point. That voltage is better than almost every 12900k I've seen including some monster SP96-SP100 chips.

 

How did your other 12900k's and 12900ks faire?

 

And yeah, that is one good indicator but if it is a leaky/hot chip a lot of that goes right out the window.

 

 

 

 

[win32asmguy]

7 hours ago, electrosoft said:

 

That makes a lot of sense at that V/F point. That voltage is better than almost every 12900k I've seen including some monster SP96-SP100 chips.

 

How did your other 12900k's and 12900ks faire?

 

And yeah, that is one good indicator but if it is a leaky/hot chip a lot of that goes right out the window.

The 12900ks is much higher at the start but lower at the top end.. Curve is as follows:

 

8x 0.914v

18x 0.914v

40x 0.969v

48x 1.159v

52x 1.354v

53x 1.354v

55x 1.354v

 

47x 1.327v Cache

 

It did not run very well in the NH55 at stock. 135W at 33x P-cores, around 20400 in CBR23.

 

I also tried the locked 12900 non K in the strix z690. It could not read VF curve and the SP rating was crazy high - SP135 or so. Also undervolting had no effect either with the locked chip.

 

The NH-D15S is a pretty tight fit with the strix z690. The Corsair case I have does not help either as there is barely 1.5 inch clearance at the top to unhook those fan clips.

[electrosoft]

36 minutes ago, win32asmguy said:

The 12900ks is much higher at the start but lower at the top end.. Curve is as follows:

 

8x 0.914v

18x 0.914v

40x 0.969v

48x 1.159v

52x 1.354v

53x 1.354v

55x 1.354v

 

47x 1.327v Cache

 

It did not run very well in the NH55 at stock. 135W at 33x P-cores, around 20400 in CBR23.

 

I also tried the locked 12900 non K in the strix z690. It could not read VF curve and the SP rating was crazy high - SP135 or so. Also undervolting had no effect either with the locked chip.

 

The NH-D15S is a pretty tight fit with the strix z690. The Corsair case I have does not help either as there is barely 1.5 inch clearance at the top to unhook those fan clips.

 

I spent an hour or so pouring through all my collected data, both personal and from other forums, of V/F curves and I can't find another 12900k that has that low of a V/F point @ 4200mhz so you clearly have a winner in that aspect. Even my 12900k isn't sub 1v. It is 1.014v @ 4200.

 

That goes to show you for our purposes with this particular model SP rating and overclocking has its limits on that low end. Only thing that would trump it would be a cooler running chip assuming you run into thermal throttling at 4.0-4.2. But if no thermal throttling just PL reached I would call it a wrap with that one. Congrats bro!

 

 

[win32asmguy]

1 hour ago, electrosoft said:

I spent an hour or so pouring through all my collected data, both personal and from other forums, of V/F curves and I can't find another 12900k that has that low of a V/F point @ 4200mhz so you clearly have a winner in that aspect. Even my 12900k isn't sub 1v. It is 1.014v @ 4200.

 

That goes to show you for our purposes with this particular model SP rating and overclocking has its limits on that low end. Only thing that would trump it would be a cooler running chip assuming you run into thermal throttling at 4.0-4.2. But if no thermal throttling just PL reached I would call it a wrap with that one. Congrats bro!

Wow thanks for checking all of that data.

 

It is a great chip for the NH55! I should delid it as there is thermal throttling.

[jaybee83]

On 7/30/2022 at 7:55 PM, win32asmguy said:

Wow thanks for checking all of that data.

 

It is a great chip for the NH55! I should delid it as there is thermal throttling.

i absolutely insist on delidding that chip! 😃✌🏼

 

btw, new upload from GN comparing the TG contact frame at 35 bucks vs. Thermalright frame at 4.35 bucks from AliExpress *lulz*

surprisingly, they are statistically performing the same even though tolerances on the TR frame are way higher. however, installation needs much more finesse and skill to get it just right.

 

[electrosoft]

4 hours ago, win32asmguy said:

Wow thanks for checking all of that data.

 

It is a great chip for the NH55! I should delid it as there is thermal throttling.

 

Yeah, that 12900k is a prime specimen for the NH55 that would love a good delid especially with looking to optimize thermals since it has such a good V/F at that critical juncture.

 

Looking at the early Dell numbers for their Precision w/ 12950HX, those numbers are going to have to improve drastically quickly to get in the ballpark of what the NH55 is putting out especially with a good chip. I can't wait to see the results with a good delid.

 

After that, some pairing tests see if you need to use Nanogrease Extreme or if the spread and pressure is good enough to step up and use something else or even liquid metal if the pair is good enough and a foam bridge if you want to be daring.

 

 

[electrosoft]

1 hour ago, jaybee83 said:

i absolutely insist on delidding that chip! 😃✌🏼

 

btw, new upload from.GN comparing the TG contact frame at 35 bucks vs. Thermalright frame at 4.35 bucks from AliExpress *lulz*

surprisingly, they are statistically performing the same even though tolerances on the TR frame are way higher. however, installation needs much more finesse and skill to get it just right.

 

 

Maybe order a couple (or more) of those at 4.35 and see if you can drill some new mounting holes to match and shoehorn it into the NH55....

 

Just about everytime a good delid results in better temps sometimes extreme. (10c+).

 

 

[jaybee83]

back to flashing and testing, w00p w00p!

 

fingers crossed no hard bricks this time around 😛 

[jaybee83]

were FINALLY getting somewhere! original quote: "it seems like the evil spirits have finally left your machine, exorcism accomplished" 😂🤠

 

ok we finally got a solid baseline now and can successfully flash custom code. now its off to getting actual shit done. wish us luck guys!

[electrosoft]

28 minutes ago, jaybee83 said:

were FINALLY getting somewhere! original quote: "it seems like the evil spirits have finally left your machine, exorcism accomplished" 😂🤠

 

ok we finally got a solid baseline now and can successfully flash custom code. now its off to getting actual shit done. wish us luck guys!

 

That is awesome to read! Progress! 👏

 

 

[win32asmguy]

 

Welp, finally a video review of the NH55JNNQ.

[Taraquin]

I think a 12400 would be a awesome deal in this notebook. Easy to cool since it only uses about 30-40W in gaming and 70-80W in CB, better with UV.