Forums first comp: CPUz Benchmark

[Mr. Fox]

1 minute ago, ArchStanton said:

Another thing to keep in mind, very few (room temperature) liquids have the "heat energy transport" capacity of plain old water.  Methanol, in particular, is a significantly poorer coolant when comparing raw capacity for energy transport.  A given volume/s of a 50/50 mix of methanol and water will "haul away" less energy from your block than an equal volume/s of pure water.  I don't recall from school if this is all down to the specific heat of the two substances, or if other factors come into play as well (like "thermal conductivity")

 

 

This phenomenon is a pretty common "tradeoff" when deciding how strong to mix things like antifreeze.  Pure water cools the engine better (under "normal" operating conditions), but we need the added benefits of the additives in automotive antifreeze (freeze protection, boiling point elevation, corrosion inhibition, etc.)

 

Hope this is helpful as you fellas map out your "extreme" cooling strategies.

Great point. That is important to consider as well.

 

That said, ice doesn't flow inside of a loop very well, so something with poorer thermal conductance might work better than a clogged system, LOL.

 

I am not going to be making any kind of changes right away, but it is certainly something fun to think about for a project in the future.

[ArchStanton]

4 minutes ago, Mr. Fox said:

ice doesn't flow inside of a loop very well, so something with poorer thermal conductance might work better than a clogged system

100%.   It seems like "trade-off" is the name of the game in most facets of the overclocking realm.  Something most of you guys have known for a long time and I'm only beginning to see for myself.  

[Storm-Chaser]

Just now, J7SC_Orion said:

 

...check out this 'new-fangled invention' 

 

Are you going to get one? Problem is I am not running caseless, so I would have to put this computer on it's side to use the pot and unfortunately my water cooling system does not like laying flat and things go wrong. 

 

Just now, Mr. Fox said:

I wonder if running methanol inside of the loop would chew up o-rings, tubing and pumps due to lack of lubrication and acting as a solvent on the materials? I'd hate to have to start replacing things because they got ruined by the methanol.

 

EDIT: Running methanol is pretty extreme, so there is no guarantee that it will not damage components in your loop, so take what I say with a grain of salt. That being said, here is my take:

 

I've been running meth for about 3-4 months now and have had no problems whatsoever. No leaks or anything like that. The pumps seem fine as well, same RPM as will standard coolant. Then again I'm only running 40% methanol, so hard to say exactly how your loop will respond to 100% methanol.

 

Regarding the o-rings, I use Hylomar HPF, a sealant that is typically used for transmission cases and engines, but the important factor here is that it doesn't cure. It always remains pliable and can be removed easily but is still highly effective and wont wash away and expose the orings to meth. I used it on my loop and can say for certain it's highly effective. 

 

 

Edited April 8, 2022 by Storm-Chaser

[ArchStanton]

14 minutes ago, Storm-Chaser said:

the important factor here is that it doesn't cure.

Not the exact formulation/brand that we use at work, but those "gasket maker" compounds that never quite "dry" are some handy  to have around .  Often needed when the mating surfaces can move slightly more than "typical" in relation to one another.

[Storm-Chaser]

7 hours ago, Mr. Fox said:

run the lines through the window, LOL. With the heat here I wonder how fast it would evaporate? Probably would not last as long as your did in the winter. It is never cold here, even in the winter.

Be sure to get a wireless mouse and keyboard and extensions for other hardware. You could place a cover over the 5 gal bucket to help keeping it cool. 

 

Edited April 8, 2022 by Storm-Chaser

[J7SC_Orion]

2 hours ago, Storm-Chaser said:

Are you going to get one? Problem is I am not running caseless, so I would have to put this computer on it's side to use the pot and unfortunately my water cooling system does not like laying flat and things go wrong. 

 

(...)

 

...that pic was from my work space - but I haven't used that pot in a while; best to employ them when a new setup is on a test-bench.

 

 

Edited April 8, 2022 by J7SC_Orion

[Mr. Fox]

On 06/04/2022 at 10:04, Storm-Chaser said:

Thanks for the comprehensive write up. Have any pictures on hand?

 

On 06/04/2022 at 10:18, Mr. Fox said:

I will grab some pictures for you. I had to order more fittings after reconfiguring my setup with the distribution manifold and ran out of fittings in the process. As soon as I get the chiller reconnected I will post them. I should have the fittings I ordered by this weekend.

The fittings arrived a day earlier than originally expected. Rather than photos, I made a very short video so I could describe the setup better. The video is still processing, so quality at the moment is not great. Time to hit the sack. You guys have a nice Friday and a wonderful weekend.

 

[Storm-Chaser]

38 minutes ago, Mr. Fox said:

 

The fittings arrived a day earlier than originally expected. Rather than photos, I made a very short video so I could describe the setup better. The video is still processing, so quality at the moment is not great. Time to hit the sack. You guys have a nice Friday and a wonderful weekend.

 

Whaaat??? I've never seen anything like that before. Plus, your attention to detail is basically 100%. 

 

Also I like that you have quick disconnect fittings on most of the important lines. It's got to be breeze to work on the cooling system because of that.

[ArchStanton]

Nice video @Mr. Fox .  A hypothetical for yourself, and the other sub-ambient OC'ers around these parts: if the radiators were internal rather than external and we had unlimited chiller capacity, would there be any likely performance benefits to circulating the chilled coolant through our radiators and using the case fans mounted to those radiators to take most of the entire system sub ambient rather than the CPU/GPU blocks alone?  I suppose answers will depend on a host of factors (i.e., how aggressive and how hot a DRAM "profile" are we running etc.), but I was just curious to hear some general impressions/perceptions related to this scenario.  Also, in this "whole rig sub ambient" scenario if we keep the fan RPM way up, and very slowly lower case temperatures, can we avoid condensation?

[Storm-Chaser]

6 minutes ago, ArchStanton said:

Nice video @Mr. Fox .  A hypothetical for yourself, and the other sub-ambient OC'ers around these parts: if the radiators were internal rather than external and we had unlimited chiller capacity, would there be any likely performance benefits to circulating the chilled coolant through our radiators and using the case fans mounted to those radiators to take most of the entire system sub ambient rather than the CPU/GPU blocks alone?  I suppose answers will depend on a host of factors (i.e., how aggressive and how hot a DRAM "profile" are we running etc.), but I was just curious to hear some general impressions/perceptions related to this scenario.  Also, in this "whole rig sub ambient" scenario if we keep the fan RPM way up, and very slowly lower case temperatures, can we avoid condensation?

If you turn the radiators and fans around in the case and pump chilled water through them it does likely have a reliability benefit if your enclosure is sealed up. Obviously this wont work very well if you are caseless, but it might be a cool idea to test out and see what happens. Basically, instead of using the radiators to heat the case you are using them to cool the internal components. 

[Mr. Fox]

2 hours ago, Storm-Chaser said:

Whaaat??? I've never seen anything like that before. Plus, your attention to detail is basically 100%. 

 

Also I like that you have quick disconnect fittings on most of the important lines. It's got to be breeze to work on the cooling system because of that.

Thank you. I am a big fan of QDC fittings and use them extensively on both of my systems. They make everything very convenient.

11 minutes ago, ArchStanton said:

Nice video @Mr. Fox .  A hypothetical for yourself, and the other sub-ambient OC'ers around these parts: if the radiators were internal rather than external and we had unlimited chiller capacity, would there be any likely performance benefits to circulating the chilled coolant through our radiators and using the case fans mounted to those radiators to take most of the entire system sub ambient rather than the CPU/GPU blocks alone?  I suppose answers will depend on a host of factors (i.e., how aggressive and how hot a DRAM "profile" are we running etc.), but I was just curious to hear some general impressions/perceptions related to this scenario.  Also, in this "whole rig sub ambient" scenario if we keep the fan RPM way up, and very slowly lower case temperatures, can we avoid condensation?

I have some observations to share, but I have to jump into a meeting at work that is going to run for about 7 or 8 hours today. I will be back this afternoon to share what I have observed in my tinkering.

[ArchStanton]

2 minutes ago, Storm-Chaser said:

If you turn the radiators and fans around in the case

Aye, optimal or not, all 3 of my radiator/fan installations are pure "intake".  I have a single 140mm rear and four 120mm side fans all operating as exhaust.  I presume if the air in my "lab" is cooler than the air in my case there will be a slight benefit .

[Mr. Fox]

8 hours ago, ArchStanton said:

Nice video @Mr. Fox .  A hypothetical for yourself, and the other sub-ambient OC'ers around these parts: if the radiators were internal rather than external and we had unlimited chiller capacity, would there be any likely performance benefits to circulating the chilled coolant through our radiators and using the case fans mounted to those radiators to take most of the entire system sub ambient rather than the CPU/GPU blocks alone?  I suppose answers will depend on a host of factors (i.e., how aggressive and how hot a DRAM "profile" are we running etc.), but I was just curious to hear some general impressions/perceptions related to this scenario.  Also, in this "whole rig sub ambient" scenario if we keep the fan RPM way up, and very slowly lower case temperatures, can we avoid condensation?

8 hours ago, Storm-Chaser said:

If you turn the radiators and fans around in the case and pump chilled water through them it does likely have a reliability benefit if your enclosure is sealed up. Obviously this wont work very well if you are caseless, but it might be a cool idea to test out and see what happens. Basically, instead of using the radiators to heat the case you are using them to cool the internal components. 

There is a lot to consider here. Including a radiator in a chilled water loop will always prevent the water from getting as cold as the chiller can get it. There could certainly be some benefits to having an air conditioned case, which is exactly the effect of having internal radiators on a water chilled system. Even with an open bench, if you have chilled water running through the radiator it will blow very cold air onto whatever components are in the line of fire from the fans. It can help cool hot memory modules and hot GPU back plates for sure, in an enclosed case or an open bench. This absolutely has merit for extended hours of intense gameplay if you are a gamer. Because it floods the inside of the the case, (or adjacent air space on an open bench,) with freezing cold air, I think it is more beneficial than the recently release TEC cooling units we have seen hitting the market because the air blown off the chilled radiators is extremely cold. The only downside to leaving the radiator in the loop is that the water cannot get "as cold" because the radiators are counter-acting the work of the chiller by attempting to warm the water to equalize it with ambient temperatures. The more effective the radiators are, the less effective the chiller will be. If the goal is to achieve the highest CPU and GPU overclock possible for chasing benchmark numbers, then you need to remove the radiator from the loop.

 

I think having the chilled air inside of the case may help with condensation because the dew point changes. Having strong fans blowing air vigorously over the CPU and GPU water blocks also helps mitigate the condensation. 

 

You can identify the dew point to target a low temperature limit using the calculator at www.dpcalc.org. You need to know the ambient temperature and relative humidity. I have a couple of small desktop hygrometer/thermometers placed near the computer that tell me the temperature and humidity in my office. The ones I have are no longer available, but they are very similar to these: https://www.amazon.com/TASOGEN-Thermometer-Hygrometer-Temperature-Fahrenheit/dp/B08P3QJJZL

 

8 hours ago, ArchStanton said:

Aye, optimal or not, all 3 of my radiator/fan installations are pure "intake".  I have a single 140mm rear and four 120mm side fans all operating as exhaust.  I presume if the air in my "lab" is cooler than the air in my case there will be a slight benefit .

I generally do the same. I think having positive pressure inside of the case is ideal. It helps with dust, too.

 

Edited April 8, 2022 by Mr. Fox

[Mr. Fox]

On 29/03/2022 at 13:00, Storm-Chaser said:

Since @Bastiaan_NL and others posted some great single core numbers I wanted to post a single core leaderboard as well... (I still have to update some of the notes)

Here is a new single core for me...

On 29/03/2022 at 12:47, Storm-Chaser said:

Okay guys managed to finally get things updated. I wish I could say I'd have more time to continually update real time but I just can't do it with everything else going on. Consequently, updates will be about once a week or so if I can manage that.... And going forward please post relatively new results for this we want it to be fresh hardware results if possible from here on out. Thanks. I colored the intel scores to further draw a delineation between intel and AMD results. If u don't think it fits we can revert back to red only for overall score category. Thank You. If you see anything out of place let me know and I will get it fixed.

 

 

 

 

Looks like post #1 has not been updated with this new info for multi-core.

[Storm-Chaser]

1 hour ago, Mr. Fox said:

Here is a new single core for me...

Looks like post #1 has not been updated with this new info for multi-core.

Fixing tomorrow.

[schuck6566]

I just want to say,all the specialized cooling talk here is corrupting me. I saw a 3 and 1/2 ton ac unit for sale on the local sale site and my FIRST thought was "If I'm gonna switch over from air cooling,I bet THAT would keep my system chill..." of course I was then hit with the * am I going to put that in a 10th floor 1 bedroom apt.  

Edited April 9, 2022 by schuck6566
added images of local site listing.

[J7SC_Orion]

5 hours ago, schuck6566 said:

I just want to say,all the specialized cooling talk here is corrupting me. I saw a 3 and 1/2 ton ac unit for sale on the local sale site and my FIRST thought was "If I'm gonna switch over from air cooling,I bet THAT would keep my system chill..." of course I was then hit with the * am I going to put that in a 10th floor 1 bedroom apt.  

 

...yeah, but just think how much you can $ave on water-cooling gear per year, for just a 'little' effort 

 

On a similar note, a friend of mine owns a wholesale fruit an vegetable business...his warehouse has two 'cold rooms' at 400 square feet each. They can be cooled down all the way to about -4 C, though they usually are at around +5 C. While I have his permission, I have yet to try out some oc'ing in those...the systems I build are simply too big and heavy to safely transport the 5km or so.

[Mr. Fox]

Well, I finally pulled off the crappy (and ugly) T-Force heat sinks and replaced them with the RAM jackets and finned heat sink that I had on my GSKILL DDR4. I was getting errors in TM5 due to the memory overheating right near the end of the test run and now no errors and 12°C cooler. I am also glad that the RGB puke rainbow is gone with the LEDs hidden under the heat sinks now. I never have cared much for RGB memory... chintzy gimmick that requires bloatware filth to control. Thanks, but no thanks.

 

I honestly don't know how memory manufacturers think they are doing us a favor by slapping thin aluminum sheets with an adhesive foam "blanket" to snuggle the ICs in their own heat. The stock ones are never much good, but they'd probably be good enough to leave alone if they actually used good thermal pads instead of that adhesive foam rubber crap to trap heat. It also makes them a real pain in the butt to remove from the modules. They make them so difficult to remove, maybe they are scared someone might discover a memory module hiding under the tacky dress.

 

 

Edited April 10, 2022 by Mr. Fox

[ArchStanton]

8 hours ago, Mr. Fox said:

It also makes them a real pain in the butt to remove from the modules. They make them so difficult to remove, maybe they are scared someone might discover a memory module hiding under the tacky dress.

As someone who is considering replacing his RAM heatsinks in the future (not much point till I've learned quite a bit more about tuning RAM), I'd be very interested in any tricks of the trade.  I'm sure there are other videos and/or text-based tutorials out there already, but I'd be all ears for anything you wish to share.  Though perhaps Storm-Chaser's life would be simpler if we didn't plop it down in his thread .

[Bastiaan_NL]

Maybe it's an idea to create a new topic with general system tuning tips, questions and guides, I bet all of us would benefit from that.
We only need someone who is willing to create that topic and keeps it up to date with links to useful posts

[schuck6566]

5 hours ago, Bastiaan_NL said:

Maybe it's an idea to create a new topic with general system tuning tips, questions and guides, I bet all of us would benefit from that.
We only need someone who is willing to create that topic and keeps it up to date with links to useful posts

Great idea! Wait,isn't there some rule that the person who comes up with the idea is responsible for it afterwards??? lol

[Bastiaan_NL]

8 hours ago, schuck6566 said:

Great idea! Wait,isn't there some rule that the person who comes up with the idea is responsible for it afterwards??? lol

I was looking for a subforum to start this topic but I couldn't find the correct one. I'd be willing to do it though

[pio]

9 minutes ago, Bastiaan_NL said:

I was looking for a subforum to start this topic but I couldn't find the correct one. I'd be willing to do it though

Software or Operating Systems subsections would probably be a good choice, no?

 

I'd definitely give it a read.  Shoot, these days with as powerful as modern rigs are, I don't even bother with tuneup stuff like that anymore, but honestly it probably would help just to rid the junk.

[damric]

Ok so I ironed out a very good stable 24/7 overclock on this Ryzen 5600G. I'm still working on max bench overclock, but this is perfectly stable with 104 base clock + PBO. Interesting that it beats the new i5.

 

 

 

AMD Ryzen 5 5600G @ 4627.37 MHz - CPU-Z VALIDATOR

VALID.X86.FR

---

[xv2vk6] Validated Dump by damric (2022-04-11 16:20:49) - MB: ASRock X470 Taichi - RAM: 32768 MB

 

[Mr. Fox]

9 hours ago, pioneerisloud said:

Software or Operating Systems subsections would probably be a good choice, no?

 

I'd definitely give it a read.  Shoot, these days with as powerful as modern rigs are, I don't even bother with tuneup stuff like that anymore, but honestly it probably would help just to rid the junk.

It would be the perfect place for that if it relates to OS or software. What about a hardware mod? (see below) If I am not mistaken, that gave rise to the suggestion. But, it would be good to have a place to share that info for hardware or software mods/tweaks.

 

On 10/04/2022 at 06:21, ArchStanton said:

As someone who is considering replacing his RAM heatsinks in the future (not much point till I've learned quite a bit more about tuning RAM), I'd be very interested in any tricks of the trade.  I'm sure there are other videos and/or text-based tutorials out there already, but I'd be all ears for anything you wish to share.  Though perhaps Storm-Chaser's life would be simpler if we didn't plop it down in his thread .

I used the thin and flat end of a nylon pick tool (this type) and very slowly pushed it between the aluminum skin they called a heat sink and the double-sided foam tape "blanket" they used instead of a thermal pad. Push it about 1/8th inch or so, pause, push more, pause. This will allow the tension of the nylon tool getting wedged in there to "peel" the adhesive off the memory ICs and PCB. I then used acetone to remove the leftover adhesives and did a final cleaning with alcohol. Go slow and keep the pick tool near the edge connector side of the DIMM to avoid the surface mounted ICs that are mounted near the top edge of the memory PCB. You don't want to tear off the SMDs pushing the pick tool into the danger zone near the top of the PCB.

 

If you are using DDR4 memory heat plates, use a slightly thicker thermal pad and/or place Kapton tape on the surface-mounted components to insulate them from accidentally making direct contact with the metal heat sinks. DDR5 has extra components on the memory PCB that are not found on DDR3/4 modules (see image below). The heat sinks for DDR4 fit fine on DDR5, just take the extra step with the Kapton tape on the SMDs to avoid a short. You could probably use a small piece of electrical tape, Scotch tape or even a few coats of clear nail polish if you don't have Kapton tape. I recommend Kapton tape. Every PC enthusiast should have some handy. It is strong and heat resistant. It will even stand up to the use of a hot air soldering station without melting or burning up. (It is actually designed for that.)

Some of these would be nice to have. They are out of stock now.

https://bartxstore.com/shop/custom-ram-copper-heatsinks-for-ddr5-ddr4/

 

Edited April 11, 2022 by Mr. Fox
added image of DDR5 to show SMDs to avoid