MY build and all the benchmark numbers

[deathstar13]

Picked this up to come with my delid device.

Amazon.com

Gonna have to learn to be stingy and get some nail polish, I have Kapton tape handy also.

If anyone is interested in using this tool after me that has been helping me along PM me so we can talk about further use.

 

[EarthDog]

Never heard of that stuff or the company....

Here's a review for best pastes (link is to liquid metal- https://www.tomshardware.com/best-picks/best-thermal-paste#section-best-liquid-metal-thermal-paste


Just be careful as that stuff is conductive. The review i read says it also doesn't last as long. I would've stuck to one of the more common, reviewed products instead of something seemingly random. Amy reviews is found were from nobodys. Hope it works out!

Edit: here's an image from the review showing the 'difference in performance among popular/known brands...




...and that's assuming you apply it right in the first place. 1C for liquid metal.

 

[Janus67]

2 new toys to play with.

Amazon.com

Amazon.com

Oh wow, yeah, just be super careful. Not many things you can do that will instantly break your cpu, but that is one of them, and then being super careful when re-mounting the CPU with the AIO as you can crush the die and it's over.

Let's just say I wouldn't be risking it to gain 100mhz for something that would not make a visible difference in games.

 

[deathstar13]

i grew up on Athlons with bare dies and shorting pins on top to unlock them. So i know how delicate a bare die is. They use to chip around the edges and some would die due to that.
I have a Thermalright cpu holder on mine so i need to be carfull i still make contact.
Im not doing it to gain Mhz but rather to gain lower temps as im fine at my ram/cpu overclock.
To me i like to experiment and learn, Lets just hope i dont learn a big $400 lesson. There is so many things to take into consideration before i attempt this but the adrenaline rush will be worth it.
I ordered some blue nail polish to match my system even tho it will never be seen so i can keep the liquid metal off the caps and resistors on top of the cpu.
My goal is a 20C drop in temps but i think thats just wishfull thinking.

 

[deathstar13]

Never heard of that stuff or the company....

Here's a review for best pastes (link is to liquid metal- https://www.tomshardware.com/best-picks/best-thermal-paste#section-best-liquid-metal-thermal-paste


Just be careful as that stuff is conductive. The review i read says it also doesn't last as long. I would've stuck to one of the more common, reviewed products instead of something seemingly random. Amy reviews is found were from nobodys. Hope it works out!

Edit: here's an image from the review showing the 'difference in performance among popular/known brands...


View attachment 370951

...and that's assuming you apply it right in the first place. 1C for liquid metal.

Always the voice of reason, I appreciate that!
This is just me doing my thing, Living and learning and the more prepared i am the better my odds. Those tests im sure were done on a heat spreader and not bare die where they actually make a difference. Ive stuck cryo freezers to my cpu back in the day so i know the dangers of not being prepared mentally for the job at hand. Plus you have heard it a million times from me i dont do what the average guy does i go over the top for the fun, learning and thrill.

 

[EarthDog]

Those tests im sure were done on a heat spreader and not bare die where they actually make a difference.

It doesn't matter what the surface is. Thermal conductivity of the product doesn't change. That's what's being tested, not bare die vs ihs.

Living and learning and the more prepared i am the better my odds.

This is true. And at the same time, what scares me when I read your posts.

May the force be with you.

 

[Janus67]

i grew up on Athlons with bare dies and shorting pins on top to unlock them. So i know how delicate a bare die is. They use to chip around the edges and some would die due to that.
I have a Thermalright cpu holder on mine so i need to be carfull i still make contact.
Im not doing it to gain Mhz but rather to gain lower temps as im fine at my ram/cpu overclock.
To me i like to experiment and learn, Lets just hope i dont learn a big $400 lesson. There is so many things to take into consideration before i attempt this but the adrenaline rush will be worth it.
I ordered some blue nail polish to match my system even tho it will never be seen so i can keep the liquid metal off the caps and resistors on top of the cpu.
My goal is a 20C drop in temps but i think thats just wishfull thinking.

I get it, my first build for myself was a 1.4ghz athlon Thunderbird. Let's just say things are more costly now than they were 25+ years ago with less headroom (I miss my 2500+ Barton)

 

[deathstar13]

It doesn't matter what the surface is. Thermal conductivity of the product doesn't change. That's what's being tested, not bare die vs ihs.


This is true. And at the same time, what scares me when I read your posts.

May the force be with you.

Any heat spreader is acting like a sandwich which doesn't transfer heat as well as direct contact. I think the liquid metal on the die is more effective than on a heat spreader. I feel the surfaces are different on the die and that makes the LM more effective.
I do agree with you on my noobness in many aspects, But i learn fast and have experience in ways you fine folks dont see. I still say im a noob towards todays PC building.

Post magically merged: Apr 13, 2025

I get it, my first build for myself was a 1.4ghz athlon Thunderbird. Let's just say things are more costly now than they were 25+ years ago with less headroom (I miss my 2500+ Barton)

I think my first cpy was 66Mhx or so and i thought it was awsome to get a 1KMhz Athlon. I believe i had a Thunderbird also. Im finding out that todays cpus have very little headroom and ram OC'ing is almost as bad. Sure you can OC it to 8000Mhz but the cas is so loose and its 2:1 my 6400Mhz Cas32 1:1 is faster. And it wont OC for crap.

 

[freeagent]

I finally got 6400 to pass TM5, yay! only took 1.55v to do it lol..

 

[EarthDog]

Any heat spreader is acting like a sandwich which doesn't transfer heat as well as direct contact. I think the liquid metal on the die is more effective than on a heat spreader. I feel the surfaces are different on the die and that makes the LM more effective.

Brother, youre not understanding my point. Peanut butter between two pieces of bread or by itself on the spoon is still peanut butter.

The thermal properties of the tim DOES NOT CHANGE. What changes is the iridium solder tim is being replaced by something better and/or (because you can relid too and get improvements) as well as losing the ihs. You're right, but that wasnt my point and irrelevant in testing tim anyway. I know how and why delidding works, lol.

All I'm trying to show you is the difference between most tim is negligble/within 1/2C... that and I've never heard that weird brand you chose so lord knows how it performs. Surely you saw a random yt video and here we are. Again, all good, just sharing the facts so you can do whatever it is you please, but armed with ACTUAL knowledge of the whats and whys.

 

[freeagent]

I would focus on stabilising your system. I am not certain you are actually stable.. anywhere.

Tis important.

 

[67Elco]

Peace out lol...

 

[deathstar13]

I finally got 6400 to pass TM5, yay! only took 1.55v to do it lol..

View attachment 370953

Im lucking as my ram will run Expo specs 1:1 undervolted at 1.35v, I thing the hottest ive seen them was 32C running Prime large TFT's, Today ill spend the day stress testing this OC. It passed a Cinebench run and some Prime yesterday but not enough to prove stability. Really im only testing the cpu as the ram was tested at these stock undervolted settings previously.

Brother, youre not understanding my point. Peanut butter between two pieces of bread or by itself on the spoon is still peanut butter.

The thermal properties of the tim DOES NOT CHANGE. What changes is the iridium solder tim is being replaced by something better and/or (because you can relid too and get improvements) as well as losing the ihs. You're right, but that wasnt my point and irrelevant in testing tim anyway. I know how and why delidding works, lol.

All I'm trying to show you is the difference between most tim is negligble/within 1/2C... that and I've never heard that weird brand you chose so lord knows how it performs. Surely you saw a random yt video and here we are. Again, all good, just sharing the facts so you can do whatever it is you please, but armed with ACTUAL knowledge of the whats and whys.

If you feel this is such an issue with the liquid metal i can just use Noctua H2 on the die.
I need all the knowledge i can get as to why im here, This is like being back in EE classes as a teen.
I do need to know how to clean the solder off the die so its clean and bare, Know how thats done?

 

[freeagent]

96GB is a massive load to push around. Testing 64GB is a chore, yikes.

 

[deathstar13]

96GB is a massive load to push around. Testing 64GB is a chore, yikes.

My old butt fell asleep and missed most of the day, But ill run more stability test and make sure everything is stable.
Also very nice timings at 6400Mhz as im at CL32, Basically i just run stock settings with a .05v undervolt.

Does anyone know if i need something like this after i delid my cpu.

Amazon.com

I currently have this but its no the same im sure.

Amazon.com

My Mem controller seems to be rather strong, I just dont have the full knowledge to push it further, And if i did id only go too 6600Mhz so i can stay 1:1
I can try 6600 2:1 and id bet it would pass but would give worse results.In that situation i feel id need 7600Mhz 2:1 to be faster in tests and i mush prefer stability over pushed limits.

 

[freeagent]

My old butt fell asleep and missed most of the day

Haha I had a 2 hour nap too.. got my hair cut so you don't see so much white, and so I don't look like I am going to rob you.

My mem controller is also strong I think, at least average anyways.. average is good.. right? Right? Whatever..

My mems though, not sure about them, they don't feel confidence inspiring. I am running higher voltage than I would like to be.. but whatever.. warranty if I need, and I have a spare set if I need I guess. But they will be fine I am sure.

As for delidding, it is not something I have ever considered. Even back when it started it was a hard nope from me.

I have crunched cores on S462.. not a good feeling, or sound.

 

[deathstar13]

I usually wear a Mohawk dyed black, But im gonna let it grow out and look like a normal old man i suppose.
I just had Prime fail on me and seems like i ram issue so im going back to stock 1.4v on my mem and start over.
So does that mean you dont want to borrow the delidder once i do mine? You guys seem so against me doing it that it makes me nervous now.
I gotta figure out the correct mounting for my AIO with a delidded cpu.

 

[freeagent]

Lol nah I am good on the delidder.

I am not against you doing it, you are a big boy

I just know what it feels like to kill a CPU, especially a good one.

 

[deathstar13]

Ive never killed a die but ive chipped them on the edges, Back in Athlon days my die looked rounded off and it still worked. But alot of those chips had stiff organic like frames so they never were flexible. Flexing the chip would be very bad on these so i need to go slow and tighten the AIO down in steps.

 

[Janus67]

.
So does that mean you dont want to borrow the delidder once i do mine? You guys seem so against me doing it that it makes me nervous now

To be honest, I wouldn't consider delidding a processor of mine unless I was wealthy enough to shrug and go buy a couple more at the store to try again. To me the juice isn't worth the squeeze to save some temps (which won't matter when gaming to be in a safe zone) outside of just because I know that I could.

Again, it's your money and time and effort, it would just make me too anxious to be having an anxiety attack while tightening the cooler knowing that a bit too much pressure just made an expensive keychain. (As I made about 20 years ago when I didn't leak test long enough on a water-cooling build and killed an expensive gpu)