Considering that good heatsink compound is cheap and abundant, no.

Best and worst thermal compounds have about 2-3C difference at best so go wild and use them on all non critical components I would say.

They're not terrible and will work perfectly for everything that won't get insanely hot (modern high end GPUs and CPUs).

Delidding is done to remove additional thermal resistances out of the equation and thus improve heat transfer. It isn't nearly as important on older CPUs where silicon was soldered directly to the heatspreader as the thermal resistance is much lower than some sort of silicon to thermal paste to metal interface.
Also unlike heatsink pushing on the CPU, the heatspreader is just glued in place and if it was to get loose in some way you get increased thermal resistance that pressure alone cannot fix. The pastes used often harden and i such a thing happens you lose huge amount of transfer capability.

mr tibs wrote:

proof. if this was the case why are people deliding CPUs? best to worse is closer to a 10c difference. https://overclocking.gu […]
Show full quote

appiah4 wrote:

Best and worst thermal compounds have about 2-3C difference at best so go wild and use them on all non critical components I would say.

proof. if this was the case why are people deliding CPUs? best to worse is closer to a 10c difference.
https://overclocking.guide/wp-content/uploads … est_update1.png
https://play3r.net/wp-content/uploads/2017/03/5GHz-Load.png

1. The crap TIM that Intel uses.
2. The IHS is not mounted tight enough to the die, which makes the thermal transfer horrid.

10c... pfffft. If you delid an old Althon 64/x2 or Opteron, you will drop load temps by about 20c.

I still have my Opty 175 that is dellided.

I too have a container full of these "soft paks" from Ebay, that I've never used. Feels weird to just toss them, but I have two tubes of Arctic Silver 5 purchased ~6 years ago that never run out and surely work better.

cyclone3d wrote:

If you delid an old Althon 64/x2 or Opteron, you will drop load temps by about 20c.

I got about a 10 C drop at max load with my Opteron 185 re-lidded.

I've seen people try ketchup and toothpaste as a thermal compound on overclocked CPUs and get 2C or 3C higher temperatures than Arctic Silver.

I would use the soft packs without any worry on any hardware.

When I bought a heatsink once it came with a tube of "Cooler Master" branded paste that was awful. It was chalky, I think using no paste might have been better than using that stuff. Maybe the tube was 100 years old, but I think the Hyper-212 was a pretty fast selling heatsink in 2010, so I'm not sure.
But other than that, every tube of heatsink paste I've ever had seems about the same to me. Including a 99cent tube that I bought at Radio Shack a long time ago. Come to think of it, I liked that (metallic?) tube better than any of the modern plastic tubes or syringes that seem to be the normal packaging style nowadays. I'm currently using a syringe of Arctic Silver Ceramique which is a pain-in-the-finger to squeeze down.

When I first started using paste on my K6-3, it dropped the temperature by 5C versus nothing at all. And that was a thin application - I learned about it's existence from the internet and the internet told me not to goop it on. It helped but it wasn't exactly a revolution. So I've been pretty skeptical about the added value of any of the premium stuff. When it comes to quality, I'm more concerned with how quickly it dries out than any tiny difference in how it performs when fresh.
If I had a bunch of junk packets to get rid of, I'd use them on random builds that I wasn't real concerned about. I have a pretty lax attitude about thermal paste unless it's for a hot running chip in a system that I expect to stay assembled for years.

The most concerning application for me was actually on a car. I was putting paste on an ignition control module which experiences much bigger temperature variations (both hotter and colder) than any CPU. It also would stay assembled for a long time, and I assumed the surfaces might not be as smooth. So durability, endurance and "gap filling" were the big criteria.
I remember seeing some specs somewhere which implied Arctic Silver's more "premium" product (Arctic Silver 5) wasn't as suitable as their cheaper stuff. I think it was because it was inferior on rough surfaces. I don't remember if I saw any specs about temperature ranges. I think I ended up using Ceramique.

3dprophet wrote:

I've seen people try ketchup and toothpaste as a thermal compound on overclocked CPUs and get 2C or 3C higher temperatures than Arctic Silver.

I would use the soft packs without any worry on any hardware.

Well ketchup will almost certainly NOT last long even if it cools for a while, organic compounds do not take well to prolonged exposure to heat, they will continue to burn between the surfaces until they start to become non-conductive.

I don't know how long toothpaste would last, though.

The high heat lifetime of the soft paks is also my concern. I really have no idea how well they fare. I also bought some junk at Radio Shack in the past (though it is no longer called that in Canada). Anyway, I think the brand was "NexTech" or something like that. I have noticed that there is considerable separation of constituents in the tube with time. One is the white paste and the other is an oily substance. I haven't seem this happen with any other paste. If it separates in the tube, makes me wonder if it separates when in use.

I've been very happy with Arctic Silver 5 for many, many years. It's cheap and a little goes a very long way. I see no reason to switch to anything else. One application is good for 2-3 years or more in my experience.
Anecdotal datapoint: I have a gaming laptop that I bought in 2012, has i7 CPU and HD6970 GPU. I repasted it with AS5 one year after purchase because temperatures started to climb, and it was good until last year when I repasted it again. Runs like a champ.

appiah4 wrote:

3dprophet wrote:

I've seen people try ketchup and toothpaste as a thermal compound on overclocked CPUs and get 2C or 3C higher temperatures than Arctic Silver.

I would use the soft packs without any worry on any hardware.

Well ketchup will almost certainly NOT last long even if it cools for a while, organic compounds do not take well to prolonged exposure to heat, they will continue to burn between the surfaces until they start to become non-conductive.

I don't know how long toothpaste would last, though.

I wouldn't use it personally but I see it as a demo of how the main job of the thermal paste is to fill in the air gaps, and there is no need to spend big bucks on fancy pastes.

Tomato paste and water probably have a much lower thermal conductivity than silver to begin with. They will also dry out and change density over time. So it's hard to pin down the loss of performance to thermal conductivity.

feipoa wrote:

The high heat lifetime of the soft paks is also my concern. I really have no idea how well they fare. I also bought some junk at Radio Shack in the past (though it is no longer called that in Canada). Anyway, I think the brand was "NexTech" or something like that. I have noticed that there is considerable separation of constituents in the tube with time. One is the white paste and the other is an oily substance. I haven't seem this happen with any other paste. If it separates in the tube, makes me wonder if it separates when in use.

The Radio Shack junk I had in the '90s would separate in the tube and an application would eventually turn into a hard crust. That's my guess on what will happen long term with the soft pak stuff.

I've seen Arctic Silver 5 applications separate / thicken too over years. I'm not sure it reduces conductivity enough to be a problem though.

AMD recommends using a phase change TIM. Wax-like. It liquefies when heated. Most enthusiasts seem to mistake this stuff for dried up crusty paste. It can be found on some GPUs and some OEM CPU installations. It seems to me that this product would avoid the longevity problems with oil-based paste. It can also fill a larger gap than paste can (see the Radeon 9700 heatsink & shim design). But it's not commonly available in the DIY world.

Is the premise that in a more liquid state that the product will fill air gaps easier, that it is more thermally conductive when liquid, or that staying in the liquid state will keep it from drying out faster? Going on more viscous seems like it would result in a thicker layer compared to going on more liquid. However, if it is the case that the paste goes on thicker, heats up, then flattens with increasing heat (more liquid) that this could make for an even thinner and more ideal application.

feipoa wrote:

Is the premise that in a more liquid state that the product will fill air gaps easier, that it is more thermally conductive when liquid, or that staying in the liquid state will keep it from drying out faster? Going on more viscous seems like it would result in a thicker layer compared to going on more liquid. However, if it is the case that the paste goes on thicker, heats up, then flattens with increasing heat (more liquid) that this could make for an even thinner and more ideal application.

This is informative.
http://datasheets.chipdb.org/AMD/Other/26951_3.00.pdf

Nice. Looks like some appropriate bedtime w/iPAD reading material.

Interesting scientific paper I found. The thermal conductivity of Artic Silver is actually 1/10th of what they claim it is.

http://www.nrel.gov/docs/fy08osti/42972.pdf