First try on Dice, Results and Questions

[Lochekey]

So I had a chance to try my hand at Dice cooling today and figured I would share my results and hopefully get some questions answered. I apologize in advance for the crappy pics, I snapped them in haste with my phone.

First, Pics of the setup.








Second, some nice results (Unfortunately I cannot submit any of these at the bot yet as I do not want to lose my novice status for a little yet)
Please note that I did not do a lot of tweaking on these as my main goal was to learn the basics of handling the dice. Also, everything was run on Win7 so superpi took a serious hit.








Third, my questions,

1: During this testing I had an issue with the acetone gelling up almost to the consistency of runny mashed potatoes. From reading other threads I do not believe this is normal. I'm assuming I got a bad batch of acetone that had water in it?

2: When running XTU it kept reporting temps in the 90-100c range and noted that the cpu was throttling. I'm assuming this was from a bad mount. I had no way of monitoring the actual temps as my temp probe is currently at my place of work. Please see pics below for the mount after I removed it.

3: When running any multi threaded bench I did not seem to gain much if any headroom as compared to my water system. I was able to see a marked improvement in my max clock and single threaded bench runs. I am assuming that this is the same as in question #2 and ultimately due to a bad mount.

4: To help with not having a repeat of a bad mount how does everyone check their mounts, I do not have a feel for this mounting system yet and didn't want to crank on it for fear of crushing the cpu.








Well I'm sure I will come up with more questions as we go along but this should get us started.

Thanks,
Lochekey

 

[dejo]

what paste is that?
the mount doesnt look to bad to me. I would guess that 1 or more cores are just weaker than 1 or 2.
I would also guess that you have water content in the acetone.
as a precaution I always laid a cover over the top of the pot when doing really intensive loads as the acetone can really get to boiling and bubble some out the top.
when you get it going again and get the acetone to act proper- I would still expect some temps up into the 30's or so on really heavy benches

 

[Dr. McCoy]

Huh - I've never heard of acetone gelling up so I guess condensation did get into the pot and cause it.
I normally get mine from a place like Autozone or something - Of course I don't know where you got yours from but from the looks of it I'd have to be "Eh" about getting more from there.

From what I can see with the contact pattern of your TIM to the pot it doesn't look bad and it does appear to be consistent across the surface of the pot and CPU lid. Setting it up so you get a nice, even contact pattern is what you want and looks like you got it.

Just don't forget when you set the pot in place and it's ready to go, run it without freezing it at first with your system at stock settings to heat up the TIM so it spreads out and embeds itself for the best contact pattern possible. You don't have to go into the OS itself, just let it sit and monitor temps in the BIOS and once they begin to get warm it should be ready for freezing. I normally let it go up to about 45c or so before dropping anything in the pot but I do shut it down first, add some DICE then pour in the acetone.
I'll then let it sit for a minute or two, fire it up and go right back into the BIOS to see what it's doing. If it looks good I'll shut it down again, add more DICE and then let it pull temps down where I want them to be before I really tweak it and let it go into the OS. Be sure to doublecheck temps while tweaking to see how it's behaving after you set it up for freezing it and personally I'll even go in the BIOS again after tweaks are applied to see how it's reacting to the tweaks before booting.
If it's all good OK I'll esc key the BIOS without saving settings and let it go.

It's all about observation and taking it a step at a time because once you have it tweaked and in the OS, either it's on or it's not.

EDIT: Forgot to mention once the TIM has been heated up and spread out, you'll need to recheck mounting torques of the pot BEFORE adding any DICE to the pot. I've seen torques actually change once the TIM has spread out since it will thin itself from a glob or lump to a film across the CPU lid like it should.

 

[Lochekey]

Cooler Master Ice Fusion.

I quickly learned to cover the pot with something to keep the spray down.

I'll plan on getting some new acetone before my next try. I thought about trying methanol as well since it has a lower theoretical temperature but I don't know how this would actually work.

I'm hoping to have my temp probes home for my next try as well, it should give me better control.

- - - Updated - - -

I got the acetone at a local hardware store. Will try a different supplier next time.

I'll make sure I set the paste by heating up next time. I took it directly cold so that may have been part of the problem.

 

[Dr. McCoy]

If you add DICE to it first you won't have the acetone volcano erupt.
Add DICE to the pot first, then acetone. Pour it in slowly until you get the amount in you need and I'd let it get the pot cold before adding any more DICE. Key here is to get the pot AND acetone down to temp before stuffing more DICE in. Just a few chunks will do that, powder will have a tendency to be more volitile related to the overflow. All you need a just a few chunks once you begin chilling it, once you hear the acetone settle down then it's ready for powder or whatever you drop in.

 

[Lochekey]

I just took the blocks I got and smashed them up and then just was dropping in whatever was in the bag size wise. I assume the smaller the size the better the cooling due to more surface area but I didn't want to get the size to small and have it be unmanageable.

 

[Dr. McCoy]

Think of it like this:
Powder will give the best cooling, pulls temps down the fastest and coldest BUT goes away the fastest too, not really stable in terms of consistent cooling while under load.
Small chunks will give good cooling and last for a bit, is a little more stable in terms of HOLDING temps under load vs the powder.
Larger chunks are the best for maintaining temps but won't pull temps down as far as powder or smaller chunks will. Works best used with shots of powder along the way for longer runs such as 32M Super PI or 1024M WPrime for example. The larger chunks work just fine on their own for longer runs too so you don't have to use them alongside powder if you don't want to but adding powder with these is best for keeping temps down the furtherest (? ) and longest during a run.

 

[Johan45]

It does look to me like you could have used more pressure on the mount. The TIM looks quite thick still on the bottom of the pot. That XTU temp tells the tale. You won't see a huge gain over water but should be able to squeeze 500MHz out of it. I went with methyl hydrate/methanol it was cheaper and is supposed to get a bit colder than acetone. Also make sure on the mount that the towel or gasket doesn't get in the way. I have had that before. The towel was holding one side of the pot a bit high. The bottom of the pot is quite a bit larger than the CPU so it can happen a lot easier than you might think.

 

[Woomack]

You don't need much acetone. It's even better if there is just a little bit ( depends from the pot ) because only in acetone is full reaction. If you use too much then temps are not really better while it uses much more dice. Check your temps and you will see. Also if there is not so much acetone then it won't make that "volcano" effect and you don't have to cover it. There will be only that steam effect when you pour acetone for the first time.
It's more as a tip to save dice during long sessions. Some chips are using a lot of dice - generally everything higher from AMD, Intel 6 core+ chips and highest Nvidia graphics.

One more tip - to be sure that TIM is better spread and on Skylake to avoid TIM "cracking". At first run CPU with empty pot. Check if all cores have similar temps and wait till it goes up to ~80*C. If you are sure that mounting is good then you don't have to check cores, just monitor temps in BIOS. After that turn off everything and slowly lower temps till you see ~ -70*C.

I'm using "Technical Acetone" as standard acetone is nowadays more often mixed with other stuff and it's usually called acetone solvent ( I'm not sure if it's correct ). Probably that's why you have that gelling effect. If there is water then simply effectiveness is dropping. You can see a bit worse temps after really long sessions just because water from air is freezing on the dice ( especially when you bench on a full pot ).

 

[QuickFast]

+1 woomack but I don't think he has anything to monitor temps I never bring cpu up to temp- I go form pot at room temp to -15 in Bios I check all bios setting then boot into windows just to make sure there no cold boot bug and that I didn't forget to install anything then reboot into bios then add acetone.I like powder ,chunks drive me crazy but that's just me I like my pot to look like a sluppy form 7-11 but a little thicker just a enough for the acetone to bubble up, plus I kept a screwdriver in the pot to break the dry ice up if it stop boiling up and if breaking the dry ice up with the screwdriver don't let the acetone to boil up time to add more

All that said my beat you will fine out what works best for you in no time

Edit -kept pot full at all times no need for unnecessary moisture

 

[Blaylock]

I've purchased bad DICE before. It had water and some type of cleaning agent mixed (frozen) in with it. It smelled like Windex. You may have bought bad DICE and not acetone. Just another possibility. You can tell if DICE is bad if it melts. It should never melt.

Also If you look closely at the pic that Johan was talking about with the thicker TIM at the bottom of the CPU it looks like you compressed the foam with the pot. I would trim some more of that foam away to potentially get a better mount.

 

[QuickFast]

could be blaylock cheap TIM fails or not -look like to heavy on the paste to me and as long it doesn't fail which he wouldn't know because it's he's first time he wouldn't know, like to see a better paste, but I don't think it's the paste.From his pic's he's running a 1/2 pot or less

 

[Lochekey]

Wow I had a lot of catching up to do.


I'm gonna have to work on the mounting to get a better feel for it . I'll probably try mounting it a few times without running it to just get a better feel for working with the pot. I will try using less TIM next time as well


I'll make sure to monitor temps next time and will try bringing the pot up to temp and will retighten before pulling it down. I will have to remember to bring my temp probe home from work next time.

I think I'm gonna try to pick up some methanol and try that as I don't trust the acetone I have.

As for stirring the acetone do you think something like a chopstick would work? I hate to have metal tools around my mobo.

I well try running with a fuller put next time I was not quite sure how feasible that was.

I doubt it was bad dice as it did not act or smell funny but I plan to try a new dealer next time. I just have to find time to get there as it is an hour away. Beats paying $5 a pound though at the local dealer.

I know the TIM I have its cheap but it was what I had around is there a better choice that won't break the bank. I was using gelid gc extreme before but was going through it like water.

I think that about covered everything but if I missed something let me know.

 

[dejo]

do you still have the retention mechanism installed when using it. can be done both ways. but if its off you have to be very careful when removing the pot as you can damage pins on the board

 

[Lochekey]

do you still have the retention mechanism installed when using it. can be done both ways. but if its off you have to be very careful when removing the pot as you can damage pins on the board

The bracket is still on. My cpu is delidded and I never glued it back on so I rely on the hold down to keep it in place.

 

[QuickFast]

just fined out how to run a fill pot with out out it boiling over and your all good like I said you will fined out what works best for you

I will try to go live Saturday to show how I run my pot but I will be running AMD 754 put you should get the the point

 

[Lochekey]

just fined out how to run a fill pot with out out it boiling over and your all good like I said you will fined out what works best for you

I will try to go live Saturday to show how I run my pot but I will be running AMD 754 put you should get the the point

Thanks quickfast. If you are able to get the feed live let me know and I will tune in. Definitely interested in seeing how an experienced bencher does it.

 

[MaddMutt]

ONE of the best advises that I can pass on for DICE benching. Once you start your Benching runs do not stop until you are done. Even a short 30sec break could allow condensation to form under the CPU. That's what I got from the guys when I first started Use a rubber band around the towel to keep it on the pot. Did you have fans on top of the pot to blow away any extra coldness.

Thanks

MM

 

[Blaylock]

+1 to all of this. ^^^

 

[MaddMutt]

+1 to all of this. ^^^

Most of that advice came from YOU