Go Back   HWBOT forum > News > HWBOT Frontpage News

HWBOT Frontpage News Official HWBOT articles and frontpage news

Closed Thread
 
Thread Tools Display Modes

[Video] GPU Thermal Paste Application, the Buildzoid Methodology

It never ceases to amaze me just how much overclockers agree on virtually nothing. Not ever it seems. From motherboard insulation, to simple CPU pot mountings and even paper towels, consensus is pretty damn hard to find. One such topic of debate is the simple act of thermal paste application. There are different ways to apply thermal paste to a CPU or GPU before mounting the cooler; a small ball, long strips, crosses etc are all techniques favored by some overclockers, and equally derided by others. One man who is never short of an opinion or ten is Buildzoid. Today we bring you the Buildzoid way to apply thermal paste to a GPU in preparation for some LN2-based, subzero overclocking.

The graphics card in question is the latest AMD Radeon RX Vega 64 card, but in fact the advice Buildzoid has to share would I?m sure be applicable to most high-end, full bore GPUs. The GPU LN2 pot he uses is the ECC Raptor 4 from der8aeur which as Buildzoid notes, offers sufficient clearance for the chokes on the PCB. Buildzoid has successfully got his Vega GPU down to temps of -130 C without cracking and reckons thermal paste application is a major factor in making this happen.

His thermal paste of choice is the popular Thermal Grizzly Kryonaut. He starts by applying direct from the tube to the GPU Vega core, dabbing a small amount to each of the corners. He then mounts the LN2 pot briefly so that he has an outline of where the paste connected with the pot. He then applies paste to the LN2 pot in a relatively thick lines from corner to corner to the other, then spreading it out to cover the surface of the pot. He then goes back to the GPU, adding plenty of paste to the edges of the GPU. He is very generous with the paste.

He goes on to reveal plenty of techniques that he?s picked up from other overclockers and offers plenty of advice along the way. Thanks for sharing as ever Buildzoid. You can catch the video here on the Actually Hardcore Overclocking YouTube channel where there?s a ton of live extreme OC sessions, PCB breakdowns and more.

You can discuss this post in the HWBOT Community Forum here.


Dec 5, 2017 - article - youtu.be
  #21  
Old 12-10-2017
FireKillerGR FireKillerGR is online now
transformer
 
Join Date: Apr 2011
Location: Greece
Posts: 918
Thanks: 218
Thanked 59 Times in 46 Posts
FireKillerGR
Send a message via Skype™ to FireKillerGR
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

Quote:
Originally Posted by mickulty View Post
Like round 1 of the ROG OCS?
had to google to find what this was about lol

If you consider this as something that made the 1070 a competitive card that tons of people would bench then I think I should skip the rest of the conversation

PS. by competition I meant ranking-wise between members, not actual competition as contest

Last edited by FireKillerGR; 12-10-2017 at 01:03.
The Following User Says Thank You to FireKillerGR For This Useful Post:
l0ud_sil3nc3 (12-10-2017)
  #22  
Old 12-10-2017
l0ud_sil3nc3's Avatar
l0ud_sil3nc3 l0ud_sil3nc3 is offline
robot overlord
 
Join Date: May 2010
Location: UNITED STATES
Posts: 1,418
Thanks: 270
Thanked 271 Times in 145 Posts
l0ud_sil3nc3 is a jewel in the roughl0ud_sil3nc3 is a jewel in the roughl0ud_sil3nc3 is a jewel in the roughl0ud_sil3nc3 is a jewel in the rough
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

Quote:
Originally Posted by FireKillerGR View Post
Didn't want to interfere but come on man... This doesn't have any substance at all.
He barely passed the water wahuhu guy who does no tweaking nor modding by something like what? 30 graphics points?

Imagine if there was an actual competition out there with guys benching this pointless (literally, no impact on the actual rankings) card
This.

Bench on LN2 and still get beat by H20 user on overall score. Sounds like someone who most definitely should be giving users advice on gpu benching. . .

Benching 1070 is a waste of time the only thing it's good for is mining lol.
__________________


Quote:
Originally Posted by sin0822 View Post
whats LOD?
Quote:
Originally Posted by Gunslinger View Post
If you can't beat 'em, report them! lol
  #23  
Old 12-10-2017
Wimpzilla Wimpzilla is offline
maintenance bot
 
Join Date: Oct 2017
Location: France
Posts: 43
Thanks: 0
Thanked 11 Times in 7 Posts
Wimpzilla
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

So first let get the data:

This sub say custom H2O: http://hwbot.org/submission/3602453_...070_8143_marks
Where is the picture of the rig proving it, there is no picture? Who tell me this was under water aside the fact it was specified in the sub? Not i don't believe @H2o vs. Ln2 but as scientific guy i believe proof, here there is no proof i'm sorry!
This is the picture submitted with the sub: http://hwbot.org/image/1873106/thumb.jpg
For what i know it could be a chiller or something like that! There is no T° specified anywhere that could prove it again, neither from internal sensors or any external probe!?

More the GPU clocks reported are lower than the other sub in the top tier of the list, but with an higher overall score!
Why higher score, i believe because of a cpu @4.7Ghz against the 4.4Ghz from the OP! Did not bother to check the RAM clock, you get the point.

So again if one know that benching 3DMark GPU = having the best CPU/RAM clocks (nonsense), how the gentlemen can leverage the OP GPU overclocking skills when the scores are highly tied to the CPU score!

If not enough one can apply all the optimizations on the system as OS and drivers. I mean benching a GPU that have almost no image output, only artifact is considerated fair here, yeah whatever.

Once again if one get the data, its hard to say whats going on here, but what can be said for sure, that the OP is skilled as the top tier gentlemen, if one would analyze the data i would say that H2o vs. Ln2 submission is the outsider from the whole data set, not the OP one!

So again what we are speaking here, i will tell you: nonsense!

Why, let me post here the GPU scores of the first 6 submissions:
-7603
-7634
-7764
-7549
-7412
-7270

So here you go, OP have better GPU score that the top submission, but still gentlemen criticize!
In my world this is called defamation and it utterly punished by law, but here mods let it go as it is nothing.
Why, i don't know! Is this normal, nope! Is this the first time it happen, nope either!

Again pathetic!

Edit: Worth to mention that the OP posted R E A L pictures of the bench rig, what top tier gentlemen do not!
__________________
E=Mc²

Last edited by Wimpzilla; 12-10-2017 at 08:25.
  #24  
Old 12-10-2017
newlife newlife is offline
transformer
 
Join Date: Aug 2012
Location: Adelaide,South Australia
Posts: 613
Thanks: 78
Thanked 100 Times in 58 Posts
newlife
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

While I'm not the biggest fan of him you guys should give him a break because unlike you guys who are fine to bunny about him, he actually puts videos and such out to help others and I know for a fact that his information on AMD gpus has come to use to others

I'm also pretty disappointed that the OC community would gang up on another overclocker like this because he actually attempts to help the community unlike most of you people
__________________
The Following 6 Users Say Thank You to newlife For This Useful Post:
Casanova (12-10-2017), FireKillerGR (12-10-2017), Frito (12-11-2017), GeorgeStorm (12-10-2017), Infuriare (12-11-2017), mickulty (12-10-2017)
  #25  
Old 12-10-2017
FireKillerGR FireKillerGR is online now
transformer
 
Join Date: Apr 2011
Location: Greece
Posts: 918
Thanks: 218
Thanked 59 Times in 46 Posts
FireKillerGR
Send a message via Skype™ to FireKillerGR
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

just a clarification, I jumped in to comment that the pov of Mick re the 1070 doesn't make sense to me

Also, system photos are pretty much useless on top scores as they are all on ln2. They are just good on showing what pots you use and promo your sponsors if any.
We aren't messing with chilled water and claim it to be h2o in order to participate in comps that have temp limits (stuff that can happen in lower leagues). Photos change nothing 99% of the times.

Regarding all the rest like hanging, judging publicly or praising for sharing (others') info; I just wanted to avoid commenting since I think both sides were kinda extreme imho
  #26  
Old 12-10-2017
Casanova's Avatar
Casanova Casanova is offline
grunt bot
 
Join Date: Dec 2015
Location: Brasília-DF - Brazil
Posts: 129
Thanks: 62
Thanked 48 Times in 32 Posts
Casanova
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

Quote:
Originally Posted by newlife View Post
While I'm not the biggest fan of him you guys should give him a break because unlike you guys who are fine to bunny about him, he actually puts videos and such out to help others and I know for a fact that his information on AMD gpus has come to use to others

I'm also pretty disappointed that the OC community would gang up on another overclocker like this because he actually attempts to help the community unlike most of you people
This.
I like being part of XOC, but sometimes i get really disappointed with some top overclockers.
Buildzoid is doing and showing overclock the best way: an addict and fun hobby, demystifying for the layman and probably future extreme members.
__________________
The Following 2 Users Say Thank You to Casanova For This Useful Post:
Infuriare (12-11-2017), mickulty (4 Weeks Ago)
  #27  
Old 12-10-2017
Wimpzilla Wimpzilla is offline
maintenance bot
 
Join Date: Oct 2017
Location: France
Posts: 43
Thanks: 0
Thanked 11 Times in 7 Posts
Wimpzilla
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

Quote:
Originally Posted by FireKillerGR View Post
Also, system photos are pretty much useless on top scores as they are all on ln2. They are just good on showing what pots you use and promo your sponsors if any.
You are missing the point completely, it's not about showing how much LN2 your pot is smoking or which brand you are using, it is an open competition and one must show the rig and modifications done to allow other to judge what you are doing. To be also able to reproduce if needed the conditions of the submission and compare results!

Because i still don't get a clear picture on how a water sub beat the four previous scores done under LN2/dice on this GPU submission by a far margin! Would be cool at this point if one give me a clear and neat explanations based on fact and data.

And you call me a noob/inexperienced, no problem with that, i do not have any experience on LN2 overclocking even if i clock from long time and had swam into LN2 because of my job. Nevertheless it not mean that i should not learn or leverage what one is doing. If it would not be for public events and some video on utube i would not even know what a pot look like. Go figure a complete noob that wish get some info about and give it a try.

Also, please the next live competition, hide every top tier clockers under a single little box where they bench and nobody could see what they are doing, just output the scores result.
Because it is what happen now on every common submissions.

There are principles that apply to competition that should be a standard, some of these standards here are clearly missing! And this brought already some issues on how this stuff is done.

And yes i trust none of you, after had leveraged that users cheat on scores or use parts and information not accessible to the public to reach they marvelous points!

In a competition elite does not mean private/hidden, it mean the most skilled users!
__________________
E=Mc²

Last edited by Wimpzilla; 12-10-2017 at 18:06.
  #28  
Old 12-10-2017
Mr.Scott's Avatar
Mr.Scott Mr.Scott is offline
Fully capable of playing games
 
Join Date: Jan 2009
Location: UNITED STATES
Posts: 2,197
Thanks: 843
Thanked 361 Times in 185 Posts
Mr.Scott is on a distinguished road
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

I admire your boldness.
__________________

Warp9-systems, the Antichrist of competitive benchmarking.
Loved by thousands, hated by millions.
  #29  
Old 12-10-2017
FireKillerGR FireKillerGR is online now
transformer
 
Join Date: Apr 2011
Location: Greece
Posts: 918
Thanks: 218
Thanked 59 Times in 46 Posts
FireKillerGR
Send a message via Skype™ to FireKillerGR
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

Quote:
Originally Posted by Wimpzilla View Post
You are missing the point completely, it's not about showing how much LN2 your pot is smoking or which brand you are using, it is an open competition and one must show the rig and modifications done to allow other to judge what you are doing. To be also able to reproduce if needed the conditions of the submission and compare results!

Because i still don't get a clear picture on how a water sub beat the four previous scores done under LN2/dice on this GPU submission by a far margin! Would be cool at this point if one give me a clear and neat explanations based on fact and data.

And you call me a noob/inexperienced, no problem with that, i do not have any experience on LN2 overclocking even if i clock from long time and had swam into LN2 because of my job. Nevertheless it not mean that i should not learn or leverage what one is doing. If it would not be for public events and some video on utube i would not even know what a pot look like. Go figure a complete noob that wish get some info about and give it a try.

Also, please the next live competition, hide every top tier clockers under a single little box where they bench and nobody could see what they are doing, just output the scores result.
Because it is what happen now on every common submissions.

There are principles that apply to competition that should be a standard, some of these standards here are clearly missing! And this brought already some issues on how this stuff is done.

And yes i trust none of you, after had leveraged that users cheat on scores or use parts and information not accessible to the public to reach they marvelous points!

In a competition elite does not mean private/hidden, it mean the most skilled users!
Am I the one missing the point? It is an open competition yes and no, you are not obligated to share with everyone else so they can reproduce your score. By the rules, you only need to show a screenshot (in oc-esports comps system photo too) and have to have a valid result.
Why would I be obligated, to share the ocp mod I found, or the thermal paste trick, or the pot mounting, or or or, so someone who might have been benching for way less than me, or even for the first time, beats me? So he can degrade all the effort and experience that someone before him had to put on making a single score?

"Hey Wimpzilla, here is my screenshot, my system photo, my mods, my tweaks, my personal ID, and some backup scores so you can submit them and save you some time. No worries pal"

Me, showing a photo before, during or after my benching sessions by itself means jack $hit tbh. But being obligated to share info about how to reproduce my score is simply idiotic to say the least.

As for the question between h2o vs buildzoid's score.

First of all, h2o didn't participate in any competition so he wasn't obligated to show his rig.

Additionally, running a gpu modded @high clocks doesn't mean you get efficiency by default. Buildzoid could hit some kind of power limitation which would limit the clocks of the gpu (you see 2350 which could drop @2250 during load )

Now, the only "fault" I see towards this subject in general, is that it wasn't mandatory to have an xtu screenshot before. Something that many people took into advantage but has been addressed now.

PS. About the latest part. Galax 1080 Ti's Hof xoc bios was given to anyone who bought the card and wanted to participate into the OPEN qualifiers.
Aside of this, I see nothing confidential that other people don't have access to.

PS 2. You can check our videos to see how we broke GPUPI. If you had participated in the qualifiers, and had qualified to the GOC finals. You could actually witness a live reproduction of it. People there, who are 80+% of my league's competitors could see the OS we were using, the tweaks we did, and how we benched the card.
So you might not trust me or whoever else here, but I trust YOU on the current lack of skill/experience and effort.
Especially the latest based on your whole perspective here.

Last edited by FireKillerGR; 12-10-2017 at 19:17.
  #30  
Old 12-10-2017
Wimpzilla Wimpzilla is offline
maintenance bot
 
Join Date: Oct 2017
Location: France
Posts: 43
Thanks: 0
Thanked 11 Times in 7 Posts
Wimpzilla
Default Re: [Video] GPU Thermal Paste Application, the Buildzoid Methodology

Well sharing a technique do not mean master it. The time you spent master it would anyway give you the advantage on a newcomer even if he know how to do, i'm sorry. This could be applied in any technical, scientific area, unless you spend time to teach to the newcomer each peculiarity of the technique to allow him to reach the same level of mastery of yours in a lower amount of time.
Hope this point is clear enough for you!

After that, reading your words you clearly state that the HWBot normal competition is under shadows until you reach the public, where no shadow could exist because more easily spotted and utterly punished.
So on public area each one get the same hardware, outside the public area do what ever you want. This is not the definition of public competition, where regulatory statement do not correspond with opinions. Otherwise law enforcement would be a joke right.
Hope this point is also clear enough!

Aside that, being ignorant, disrespectful and threatening other of inexperienced to back up you pride is funny. As far i'm not experience under LN2 techniques you are threatening me of noob without knowing my actual skill and knowledge in other areas that are not mounting a pot and benching on LN2.

More than that i asked to back up explanations by data in a neat way, you just come out with a basic power limitation excuse. If there was a power limitation and clocks would not boost as high as they should, please explain me why the GPU score is higher than the other.

Obviously you did not take the time to read what i posted and asked before and jumped on the train: "let belittle this one too to settle this easily!", sorry this stuff don't work with me. Data and scientific proof or its is just you opinion that do not worth a penny!

Unfortunately for you nope, you are the one that seems poorly skilled based essentially on your own words, go figure!
__________________
E=Mc²

Last edited by Wimpzilla; 12-10-2017 at 19:39.
Closed Thread


Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Actually Hardcore Overclocking | The buildzoid way to apply thermal paste on GPUs for LN2 sdougal Overclockers on YouTube 0 12-04-2017 20:50
[Video] Buildzoid PCB Breakdown: Galax GeForce GTX 1080 Ti HOF sdougal HWBOT Frontpage News 1 08-19-2017 00:01
Gamers Nexus Test and Benchmark Thermal Paste Methodology on AMD Threadripper sdougal HWBOT Frontpage News 0 08-16-2017 08:36
ASUS Z170 non-k unlock elmor Elmor's Lab 99 04-25-2017 21:38
[AT] FLIR One Thermal Imager Gets an Update, New Infrared Cameras with Lepton Inbound hwbotrss Tech News RSS 0 01-04-2017 21:10


All times are GMT +1. The time now is 10:32.


Powered by vBulletin® Version 3.8.4
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
Copyright HWBOT 2004 - 2015