Dell Precision 7670 & Dell Precision 7770 owner's thread

[MyPC8MyBrain]

1 hour ago, MyPC8MyBrain said:

next... cooking with fire :)

cant upload more images to my previous post,
here are the changes made in ThrottleStop TPL settings to achieve last test result,

[Kataphract]

Thanks for all the photos and write-up ! 

Well, that is dissappointing, but the photos really say it all. That's a very small amount of copper in total, looks smaller than XPS17 and Razer Blade, for laptop that is twice as thick. And no vapor chamber despite being the highest config cpu+gpu wise. And it throttles without GPU even being engaged, meaning mixed workloads are even worse.
Yeah this is nothing repasting or faster fans could have fixed... 

There was no point in putting HX range into these models sadly outside of marketing.

[MyPC8MyBrain]

did even more testing,
despite scoring pretty constantly now 22+k in CB R23 the unit ultimately still thermal throttling getting there,
i noticed in a mixed stress test that the unit does manage to handle its relative power reduced -60% before it settles in,
even in that mixed test when temp reached equal Librium GPU was around 81 with 91 hotspot, CPU was at 76,
when i say mixed stress test that's all options in AIDA64 selected and the system allowed to run for an hour,
even went the route of liquid metal on the GPU which didn't improve anything 
reverted back to paste before shipping the unit back

 

1 hour ago, Kataphract said:

no point in putting HX range into these models sadly outside of marketing

 

a more robust vapor chamber design could easily run HX without thermal throttling,
i am sure the blade 17 with its vapor chamber cold handle the HX with more grace,
these two sad heat pipes cant carry the load short of capping the CPU itself,
this powerful CPU can only run in this chassis at 40% capacity atm,

also noticed the fans this unit equipped with don't really move much air ,
even when at full rpm there is relatively poor air exhaust in the back,
 

 

[TwistedAndy]

14 hours ago, Aaron44126 said:

I don't think ThrottleStop has been updated to properly support hybrid architecture undervolting yet.

 

Actually, it supports hybrid CPUs pretty well. There're separate controls for P and E cores.

 

14 hours ago, Aaron44126 said:

Yeah, since Dell/Intel have limited undervolting capability because of potential security issues, this is what you have to resort to.

 

Yes, this vulnerability is called plundervolt.

 

8 hours ago, MyPC8MyBrain said:

there's only two heat pipes one for the GPU and one for the CPU being shared between both,

when this thing fires up it’s just too small to handle the heat,

 

Yes, I also noticed it and decided to go with XPS 17 9720 instead of 7670 three months ago.

[TwistedAndy]

1 hour ago, MyPC8MyBrain said:

also noticed the fans this unit equipped with don't really move much air ,
even when at full rpm there is relatively poor air exhaust in the back,

 

Dell is using DOO fans both in XPS 17 and 7670/7770. Fans are pushing air in both directions: through radiator and through the motherboard and cooling system. On your image air inlets are actually outlets.

 

This approach works great on XPS 17, but may be not so efficient in 7670/7770 due to smaller area.

[MyPC8MyBrain]

19 minutes ago, TwistedAndy said:

This approach works great on XPS 17

the XPS 17 has a vapor chamber thermal design,
it is different than the 7x70 series heat pipe thermal design,
i had some hand on time not long ago with the XPS 9710, excellent piece of equipment,
i got the 9710 to run full blast with zero thermal throttling no matter what i dished at it,
 

[Ionising_Radiation]

9 hours ago, MyPC8MyBrain said:

after editing @win32asmguyIA AC/DC LoadLine 

 

next... cooking with fire 🙂

(i may have to eat my words and keep this baby after all),

That's a mental improvement. It is outrageous that Dell is leaving > 50% of the peak performance of the CPU on the table with poor engineering decisions. 

[TwistedAndy]

1 hour ago, MyPC8MyBrain said:

the XPS 17 has a vapor chamber thermal design,

 

Yes, correct. I was talking about Dell DOO (or Dual Opposite Outlet). This approach is applied regardless the presence of vapor chamber. It allows using fans with larger diameter in smaller casing, pressurize the inner laptop volume, and use the heatsink as an additional radiator.

 

Actually, that's why Dell put an additional small radiator on the heatsink in Precision 7670 and 7770 🙂

 

You may find more details on how it works in the interview with a Dell engineer.

[MyPC8MyBrain]

1 hour ago, TwistedAndy said:

On your image air inlets are actually outlets

hmm... you think? 

 

26 minutes ago, TwistedAndy said:

I was talking about Dell DOO (or Dual Opposite Outlet).

got it! followed your link, indeed that is possible i didn't verify their operation,
i wonder if blocking these internal outlets will improve airflow to the back radiator,
seems to be wasting allot of static pressure going toward the one populated SSD slot,
if i place my hand on the back where the radiator is there's barely air flow id expected from two fans
(gotta test this now :D ),

 

48 minutes ago, Ionising_Radiation said:

It is outrageous that Dell is leaving > 50% of the peak performance of the CPU on the table with poor engineering decisions.

i couldn't agree more,

this is a poor neutered design to a very capable cpu from a very capable company, 

there may be a way to remove the "cpu lock" through the bios than undervolt it,
not sure ill venture the route, the lock is likely artificially imposed through the bios designed to prevent overclocking not undervolting 😉 

[brunooo84]

10 hours ago, MyPC8MyBrain said:

here are some test result pics from my runs
single run next to 30min continues run  results prior to repasting,

 

 
 

single runs (post repasting)

 

after editing @win32asmguyIA AC/DC LoadLine 

 

next... cooking with fire 🙂

(i may have to eat my words and keep this baby after all),

 

 

(disclaimer...  results may be further hindered by my system running in single channel memory mode atm)

Hi, newbie here... 

Is there a tutorial that I can follow to do what you did and increase the performance of my 7770?

[Aaron44126]

2 hours ago, TwistedAndy said:

Actually, it supports hybrid CPUs pretty well. There're separate controls for P and E cores.

 

My information must be old.  I will have to check this out.

 

1 hour ago, Ionising_Radiation said:

That's a mental improvement. It is outrageous that Dell is leaving > 50% of the peak performance of the CPU on the table with poor engineering decisions. 

 

Agreed.  Everyone should make this tweak, it is clearly one of the easiest ways to get a boost to CPU performance.  I was pretty shocked when I saw the number pop out.  I'm going to make a post with detailed step-by-step and link it somewhere near the top of this thread for easier visibility.  And it's not like this is doing anything "risky" — as I understand it, it is just putting the "default" value for Alder Lake in there which Dell has set incorrectly in error.

 

Next, I will be trying ThrottleStop to raise the PL1 limit and see what happens...

 

42 minutes ago, TwistedAndy said:

Yes, correct. I was talking about Dell DOO (or Dual Opposite Outlet). This approach is applied regardless the presence of vapor chamber. It allows using fans with larger diameter in smaller casing, pressurize the inner laptop volume, and use the heatsink as an additional radiator.

29 minutes ago, MyPC8MyBrain said:

hmm... you think?

 

Yes, air comes in from the bottom and goes out both the "back" and "front" sides of the fan, so it actually circulates through the chassis.  (I'm not sure how much good this actually does, I think it is more to keep the chassis surface temperature down than to cool off the CPU/GPU.  But at least it gives some airflow over the main SSD slot?)  Also, despite the hinge sort of in the way, the left fan actually shoves air out of the back of the system as well.

 

 

11 minutes ago, brunooo84 said:

Is there a tutorial that I can follow to do what you did and increase the performance of my 7770?

 

In addition to @MyPC8MyBrain's recent posts, look at @win32asmguy's posts on page 5 of this thread.  People are still discovering things.  At some point we will have more of a comprehensive list of things to try with a sort of feel for benefit vs. difficulty for each.

[MyPC8MyBrain]

9 minutes ago, brunooo84 said:

Is there a tutorial that I can follow to do what you did and increase the performance of my 7770

though not new to computers i am also new here, what i did is described in my posts,
most of the information is in this thread, follow Aarons link in the post above me to apply the tweak,

 

  

8 minutes ago, Aaron44126 said:

I will be trying ThrottleStop to raise the PL1 limit and see what happens...

i had good results with PL1 set to 91 + locked MIMO

[brunooo84]

Thank you Aaron44126 and MyPC8MyBrain! 

[DoC]

Can anyone confirm the thickness of the thermal pads?

[MyPC8MyBrain]

  

2 hours ago, MyPC8MyBrain said:

there may be a way to remove the "cpu lock" through the bios than undervolt it,
not sure ill venture the route, the lock is likely artificially imposed through the bios designed to prevent overclocking not undervolting 😉 

 

i had to go there... to make things simple
here are the two commands needed to unlock and free Precision 7X70 for full Undervolring/Overclocking 

setup_var CpuSetup 0x10E 0x0   # Overclocking Unlock
setup_var CpuSetup 0x43 0x0    # CFG Unlock

 

Happy tuning everyone 😉 

 

those interested in the full BIOS Dump Decrypt here it is DELL 7X70 BIOS Decrypted.rar 

[DoC]

8 hours ago, MyPC8MyBrain said:

  

 

i had to go there... to make things simple
here are the two commands needed to unlock and free Precision 7X70 for full Undervolring/Overclocking 

setup_var CpuSetup 0x10E 0x0   # Overclocking Unlock
setup_var CpuSetup 0x43 0x0    # CFG Unlock

 

Happy tuning everyone 😉 

 

those interested in the full BIOS Dump Decrypt here it is DELL 7X70 BIOS Decrypted.rar 

Great work and thank you! I am waiting for replacement fan on my 7770 and will try it after that. I am still wondering what's the thickness of the thermal pads, has anyone measured them?

[Aaron44126]

1 hour ago, DoC said:

I am still wondering what's the thickness of the thermal pads, has anyone measured them?

See @win32asmguy posts on page 5, there is a photo with the pad heights labeled.

[MyPC8MyBrain]

now that's impressive 

Multi Core and Single Core

 

Multi Core 30 min stress test

your system miles may vary...
these settings work on my Precision 7670,
(with liquid metal application on the CPU side)

  

[Kataphract]

20k on 30min test? That's amazing. One test missing is 30 minutes throttle post LM repaste but without adjusting voltages. At least I don't see that on previous page.. I presume you're already through :- ).
Looking at the chart, it's effectively just -125mV offset? Are the Power Limits on second chart necessary?

 

[MyPC8MyBrain]

4 hours ago, Kataphract said:

One test missing is 30 minutes throttle post LM repaste but without adjusting voltages.

 

14k score single run and 13.5k for 30 min run pre repaste,
see below (on the right 30 min run, left single)

24.5k is a 74% improvement from Dell's original baseline
(after going through many hoops to get what we paid for), 

 

On 10/8/2022 at 6:07 PM, MyPC8MyBrain said:

here are some test result pics from my runs
single run next to 30min continues run  results prior to repasting,

 

 
 

 

4 hours ago, Kataphract said:

it's effectively just -125mV offset?

 

in the Fiver control window there's a chart on the right corner summarizing changes,
 

-125mV on the main core with IccMax 255.75

-75mV on P cache with IccMax 255.75
-75mV on E cache
and all Cores set to 4.7Mhz
 

4 hours ago, Kataphract said:

Are the Power Limits on second chart necessary?

no one setting makes this huge leap, its the cumulative effort of all settings together,

16 core CPU beating a 24 core Xeon and only 20% shy off the top 32 core Threadripper is respectable,
considering the CPU is essentially still capped off,
and resides inside a tiny chassis with lets face it not so great cooling solution,
(though some may contest the cooling effectiveness due to these scores,
consider the system is still throttling unless i also set PL1 at 80 and cap PL2 at 100
which puts CB R23 score at 19k)  

[win32asmguy]

5 hours ago, MyPC8MyBrain said:

16 core CPU beating a 24 core Xeon and only 20% shy off the top 32 core Threadripper is respectable,
considering the CPU is essentially still capped off,
and resides inside a tiny chassis with lets face it not so great cooling solution,
(though some may contest the cooling effectiveness due to these scores,
consider the system is still throttling unless i also set PL1 at 80 and cap PL2 at 100
which puts CB R23 score at 19k)

 

I am just glad we were able to improve things up from the initial worse multi core results. I still wish we had a bios option to force the fans to run at a minimum speed on AC power!

[MyPC8MyBrain]

same here :)

 

22 minutes ago, win32asmguy said:

I still wish we had a bios option to force the fans to run at a minimum speed on AC power

i posted the full bios dump, 
i'm sure there some fan controls there that can be further modified,
 

 

you can also change active power plan passive/active fan policy,
passive will allow some heat buildup before activating fans

vs. active where fans will always run to keep things cool,

[Kataphract]

6 hours ago, MyPC8MyBrain said:

 

14k score single run and 13.5k for 30 min run pre repaste,
see below (on the right 30 min run, left single)
 

I meant post-repaste, but without adjusting voltages yet :- ). Did you jump straight after repasting into adjustments or did you test purely repasting alone as well?
I am trying to separate the effectivness ratio of power settings vs the repaste, to make sure what was more influential.
LM repaste is risky but has no impact on stability, while the power settings are a potential issue for workstation.

[MyPC8MyBrain]

14 minutes ago, Kataphract said:

did you test purely repasting alone as well?

i sure did, if you follow the link to the full original post there will be more pics with each step described,

[Kataphract]

31 minutes ago, MyPC8MyBrain said:

i sure did, if you follow the link to the full original post there will be more pics with each step described,

I see only single-run for post-repaste/no voltage mod yet. I can read ;- ) I didn't mean offence from ignorance or my laziness, I formulated my question correctly the first time:
"30 minutes throttle test post LM repaste but without adjusting"  Am I blind..?