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

[electrosoft]

4 hours ago, win32asmguy said:

Yep, got a case opened with eBay. Shame though as I was really looking forward to trying it out.

 

In other news, I finally seem to have an idea why P-cache undervolt was crashing. It appears that it is related to the cache speed which is 47x by default and may be too high at least with E-cores enabled. I tried 43x which did not crash immediately but instead after completing a CBR23 run.

 

I think cache undervolting is very important on this system, because it seems to share that limited power budget between the cores and cache. Less power used by the cache means more that the cores can use, correct?

 

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.

 

As for cache, yep and it can get real finicky >43x in many situations, shares power and can increase thermals. The fact it defaulted to 47x in a constrained environment is perplexing.

 

51 minutes ago, jaybee83 said:

btw, have a looksie 🤪

 

Oh Yes!

 

 

[jaybee83]

11 minutes ago, 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.

 

As for cache, yep and it can get real finicky >43x in many situations, shares power and can increase thermals. The fact it defaulted to 47x in a constrained environment is perplexing.

 

 

Oh Yes!

 

 

oh yes indeed 😄 i visually checked the dimensions on the NH55 ILM, the contact frame "should" fit nicely in there without getting in the way of any SMDs. the height shouldnt be an issue, since the IHS is the highest point and thus the contact to the IHS.

im currently cleaning up the stock IHS, still had the black silicone glue and the sTIM at the center. especially the latter is a biatch to clean up, but the Rockit Cool kit also helps in this regard 🤠

[win32asmguy]

5 hours ago, 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.

 

As for cache, yep and it can get real finicky >43x in many situations, shares power and can increase thermals. The fact it defaulted to 47x in a constrained environment is perplexing.

It was delidded but the IHS was in the same plastic packaging as the CPU without any tape to hold the packaging shut tight. So it looks like the IHS worked the packaging slightly open then started rubbing against the CPU PCB. The photo of it from the listing looks like those SMD's were intact and his rating was high.

 

So it looks like with the cache the magic number is 40x where I can undervolt and it doesn't crash with 10 mins CBR23. It is still kind of strange though as 47x is stable with a reboot.

[1610ftw]

What is the highest CPU power draw that can be achieved for this chassis so far? Wasn't it something like only 120W? Hopefully Prema will be able to help so that it can be increased to some degree.

[electrosoft]

5 hours ago, 1610ftw said:

What is the highest CPU power draw that can be achieved for this chassis so far? Wasn't it something like only 120W? Hopefully Prema will be able to help so that it can be increased to some degree.

 

I think @jaybee83 or even @win32asmguy pulled 140w'ish?

 

Now that I've worked out a lot of the kinks, even this MSI Pulse 12700k GL66 is pulling up to 121w and went from ~10k CB23 to 17.5k CB23. See if I can extract a bit more out of it.

 

 

[1610ftw]

1 hour ago, electrosoft said:

 

I think @jaybee83 or even @win32asmguy pulled 140w'ish?

 

Now that I've worked out a lot of the kinks, even this MSI Pulse 12700k GL66 is pulling up to 121w and went from ~10k CB23 to 17.5k CB23. See if I can extract a bit more out of it.

 

 

 

140W is quite impressive given the seriously underpowered heatsink!

 

17.5K on 121W for the GL66 is awesome, almost makes me furious how fast these tiny BGA books can be.

Did you get to those scores with MSI bios and other software tweaks only did you also repaste or do some other physical changes?

 

Speaking of tiny BGA books pulling ahead notebookcheck posted a CB23 score of more than 22K for a GE67HX which again is stunning if you think about the form factor. All while the X170 chassis is not getting any further use for desktop CPUs:

 

[electrosoft]

1 hour ago, 1610ftw said:

 

140W is quite impressive given the seriously underpowered heatsink!

 

17.5K on 121W for the GL66 is awesome, almost makes me furious how fast these tiny BGA books can be.

Did you get to those scores with MSI bios and other software tweaks only did you also repaste or do some other physical changes?

 

Speaking of tiny BGA books pulling ahead notebookcheck posted a CB23 score of more than 22K for a GE67HX which again is stunning if you think about the form factor. All while the X170 chassis is not getting any further use for desktop CPUs:

 

 

If the X15 had a beefier heatsink and cooling so much more performance could be extracted. Hopefully something pops up on aliexpress sooner than later.

 

I ended up having to clean up a lot of garbage mistakes under the hood including several mis-aligned and double stacked thermal pads, incomplete thermal coverage and tape covering a memory chip for starters. After a few thermal paste replacements I ended up doing a shim mod and replaced the stock screws with another set. I was then able to switch to kingpin thermal paste.

 

Opened up the extended menus in MSI BIOS but the H chipset is pretty limited. Next stop is Throttlestop.

 

This is all without an undervolt yet.

 

Those full fat 8+8 12th gen chips do add some extra oomph. 🙂

 

 

 

 

[jaybee83]

4 hours ago, electrosoft said:

 

I think @jaybee83 or even @win32asmguy pulled 140w'ish?

 

Now that I've worked out a lot of the kinks, even this MSI Pulse 12700k GL66 is pulling up to 121w and went from ~10k CB23 to 17.5k CB23. See if I can extract a bit more out of it.

 

 

145 😛 

 

issue being the sustained loads, not the max though. the delid, liquid metal and copper IHS combo provided thermal headroom for another +340 Mhz, im hoping i can get another 100-150 Mhz with the TG contact frame. should be here by the end of this week.

after that, im hoping for undervolting capabilities with Prema's help. were currently waiting for some software bugfixes before we can proceed....

[Haleskinn]

On 7/8/2022 at 11:24 PM, jaybee83 said:

theres a sister model with an AM4 socket that can take up to a 5950X, but no other LGA1700 socket laptop as far as im aware. definitely not socketable CPU paired with a mobile 3080 Ti around. however, this might change by fall/winter this year, lets see....hoping for the best 😅

in any case, the 3070 Ti in this system is actually trading blows with a 3080 Laptop and only lacks like 5-10% vs a 3080 Ti due to higher available TDP and desktop CPU, have a looksie:

 

https://www.notebookcheck.net/Eurocom-Raptor-X15-review-LGA1700-Core-i7-12700K-in-a-laptop.620093.0.html

 

https://www.techspot.com/review/2433-geforce-rtx-3070-ti-laptop/

why stop at 330w? go bonkers 

https://eurocom.com/ec/configure(2,404,0)780WACDCAdapter

 

780w 😄 more headroom

[Haleskinn]

also quick question what is the max tgp for the 3070ti? and what are your guys 3dmark scores so far? also anyone done lm or any other cooling mods? also what about shunt modding gpu? 

[jaybee83]

11 minutes ago, Haleskinn said:

why stop at 330w? go bonkers 

https://eurocom.com/ec/configure(2,404,0)780WACDCAdapter

 

780w 😄 more headroom

haha main reason is practicality in terms of compatible psu connector. we know that the stock 230W, a 280W and now the newly discovered 330W psu models have a compatible barrel connector for our machines. that 780W monster would have to be modded in order to fit 😛

besides, well also need to first deactivate the EC power limits, otherwise the machine just shuts off once it reaches a certain total wattage. lets see how far we can push this thing before it catches fire 😛 

 

5 minutes ago, Haleskinn said:

also quick question what is the max tgp for the 3070ti? and what are your guys 3dmark scores so far? also anyone done lm or any other cooling mods? also what about shunt modding gpu? 

 

 

well lets see what we can put together here from this thread alone:

 

youve got your homework cut out for ya, son! 😉 

[electrosoft]

8 hours ago, jaybee83 said:

145 😛 

 

issue being the sustained loads, not the max though. the delid, liquid metal and copper IHS combo provided thermal headroom for another +340 Mhz, im hoping i can get another 100-150 Mhz with the TG contact frame. should be here by the end of this week.

after that, im hoping for undervolting capabilities with Prema's help. were currently waiting for some software bugfixes before we can proceed....

 

lol, 145w nice!

 

The PL on this MSI is actually 200w (!) but it hits thermal wall around 121w so far. Going to try to get a uv going but also adjust the PL down to 80w and start scaling by 5-10w and see where it hits the perfect equilibrium point so it doesn't throttle and drop so much then try to climb again, throttle rinse/repeat.

 

The TG contact frame might be a major game changer if the socket is subject to the odd bending forces like desktops. If the retention system exerts the same amount of force as a desktop board on the laptop board which might not be as robust you might even experience even more from the TG than the desktop even.

 

 

 

 

[jaybee83]

Time for some tinkering 🤠

 

edit: aaaaand THATS why you should always do your homework BEFOREHAND, kids!

 

only after disassembly and while taking off the copper IHS did i notice that the screws fastening the ILM to the mobo are NOT the same as on desktop boards, thus the included key aint compatible *facepalm*

 

they look like some kinda hexagonal "screw-nuts", where ud need a tool to get a grip from the outside 🤨 guess ill have to put this endeavour on ice....ugh

[electrosoft]

2 hours ago, jaybee83 said:

Time for some tinkering 🤠

 

edit: aaaaand THATS why you should always do your homework BEFOREHAND, kids!

 

only after disassembly and while taking off the copper IHS did i notice that the screws fastening the ILM to the mobo are NOT the same as on desktop boards, thus the included key aint compatible *facepalm*

 

they look like some kinda hexagonal "screw-nuts", where ud need a tool to get a grip from the outside 🤨 guess ill have to put this endeavour on ice....ugh

Awwwww....

 

Copper IHS looks nice in there though. 🙂

[Clamibot]

2 hours ago, jaybee83 said:

Time for some tinkering 🤠

 

edit: aaaaand THATS why you should always do your homework BEFOREHAND, kids!

 

only after disassembly and while taking off the copper IHS did i notice that the screws fastening the ILM to the mobo are NOT the same as on desktop boards, thus the included key aint compatible *facepalm*

 

they look like some kinda hexagonal "screw-nuts", where ud need a tool to get a grip from the outside 🤨 guess ill have to put this endeavour on ice....ugh

 

You can use a pair of pliers to help you unscrew these. It's better to have the correct tool, but pliers will do as long as you're very careful. Make sure to use the smaller ones. There's less risk of breaking stuff that way.

[jaybee83]

10 minutes ago, Clamibot said:

 

You can use a pair of pliers to help you unscrew these. It's better to have the correct tool, but pliers will do as long as you're very careful. Make sure to use the smaller ones. There's less risk of breaking stuff that way.

no dice, there are two issues:

1) there is barely any room to maneuver to actually get a grip on those screws, even with the tiniest of pliers.

2) ill need those screws again when installing the contact frame, which sports ingressed holes, so again basically zero room to get a grip on those screws, let alone do exact screwing, since u require a narrow window of torque to avoid any issues with pin contact...

[win32asmguy]

3 minutes ago, jaybee83 said:

no dice, there are two issues:

1) there is barely any room to maneuver to actually get a grip on those screws, even with the tiniest of pliers.

2) ill need those screws again when installing the contact frame, which sports ingressed holes, so again basically zero room to get a grip on those screws, let alone do exact screwing, since u require a narrow window of torque to avoid any issues with pin contact...

Maybe a hex nut driver, like these from a iFixit kit?

 

[jaybee83]

6 minutes ago, win32asmguy said:

Maybe a hex nut driver, like these from a iFixit kit?

 

yeah that could potentially work, but it would need to be quite narrow in circumference on the outside to fit 🤔

[win32asmguy]

6 minutes ago, jaybee83 said:

yeah that could potentially work, but it would need to be quite narrow in circumference on the outside to fit 🤔

Yeah those precision tools tend to be better than the generic stuff at hardware stores.

 

How bad is the IHS warped by the stock mounting frame?

[jaybee83]

17 minutes ago, win32asmguy said:

Yeah those precision tools tend to be better than the generic stuff at hardware stores.

 

How bad is the IHS warped by the stock mounting frame?

i havent done any proper checking of it yet, i assume the more massive copper IHS is more resistant to warping than the stock IHS. the thermal paste application looks ok from what i can gather. but of course, best way to check would be to compare perf and temps with and without the contact frame 😄 

[win32asmguy]

On 7/13/2022 at 12:25 AM, 1610ftw said:

 

140W is quite impressive given the seriously underpowered heatsink!

 

17.5K on 121W for the GL66 is awesome, almost makes me furious how fast these tiny BGA books can be.

Did you get to those scores with MSI bios and other software tweaks only did you also repaste or do some other physical changes?

 

Speaking of tiny BGA books pulling ahead notebookcheck posted a CB23 score of more than 22K for a GE67HX which again is stunning if you think about the form factor. All while the X170 chassis is not getting any further use for desktop CPUs:

I have seen bursts of 136W with the 12900k in the NH55. Not bad given Clevo designed it to target 65W LGA CPU's. The undervolting really helps boost performance and lower noise I am very glad its possible now.

 

The BGA chips are ok but I think performance testing in the reviews are kind of useless and potentially misleading because we don't know if the results are because of good/poor silicon or good/poor thermal contact both of which can highly vary even with the exact same model.

[jaybee83]

1 hour ago, win32asmguy said:

I have seen bursts of 136W with the 12900k in the NH55. Not bad given Clevo designed it to target 65W LGA CPU's. The undervolting really helps boost performance and lower noise I am very glad its possible now.

 

The BGA chips are ok but I think performance testing in the reviews are kind of useless and potentially misleading because we don't know if the results are because of good/poor silicon or good/poor thermal contact both of which can highly vary even with the exact same model.

man i cant wait to get undervolting and ram tuning unlocked on this machine!

[1610ftw]

36 minutes ago, win32asmguy said:

I have seen bursts of 136W with the 12900k in the NH55. Not bad given Clevo designed it to target 65W LGA CPU's. The undervolting really helps boost performance and lower noise I am very glad its possible now.

 

The BGA chips are ok but I think performance testing in the reviews are kind of useless and potentially misleading because we don't know if the results are because of good/poor silicon or good/poor thermal contact both of which can highly vary even with the exact same model.

With the power envelope of the NH55 I am quite impressed that you are managing to go up to 24K in CB23 and with that roughly double of what Clevo intended to be the target wattage.

 

I also see the issue with the BGA chips and the silicon lottery - it basically comes to buying a unit, running benchmarks on it and then returning it if you are looking for the best of the best. That being said I am not aware of how much variance one has to expect with BGA chips but if it is as much as with the 12900K then the differences in results seem to easily go up to 10% if one hits the silicone lottery.

 

As for the scores I have never seen a CPU or GPU score on notebookcheck or with the usual YouTube reviewers that I couldn't beat quite easily with some tweaking and proper application of thermal paste so when a score is high that is something that I am noticing as usually those scores are always quite low. As an example I believe that their CR23 of the Raptor X15 was only 17K with the 12700K which is silly as I would expect at the least 20K - shows how good they are with this as they only take the out of the box performance which basically guarantees subpar results. So I found that GE67 quite noteworthy as even under these conditions it managed to get such a high score.

 

[1610ftw]

6 minutes ago, jaybee83 said:

man i cant wait to get undervolting and ram tuning unlocked on this machine!

You're such a tease, get a 12900K already and let us know how far you can take it with the help of @Prema 😄

 

[jaybee83]

4 hours ago, 1610ftw said:

You're such a tease, get a 12900K already and let us know how far you can take it with the help of @Prema 😄

 

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 😛