Similar to @SpyderTracks thread Here I had picked up the kryosheet for my laptop a while back. It ended up getting lost so I did a repaste with conventional MX-4 and waited on the replacement. When it came, I had never bothered to change it as it was working OK. My plan for the laptop is to use it for an arcade cabinet down the line so just wanted normal temps.
After looking at the thread I thought I would give it a go to see what the experience was like. I've gone through liquid metal and standard paste, more so because the liquid metal is corrosive. I've never had particularly great results with pastes outside of the LM so it definitely works as it should.
Prior to doing anything I left the system for 15 minutes or so and then checked the min idle temps.
I did a run in Prime95 small FFTs before the sheet went on. This is relatively fresh paste with it only being a week old. I had similar experience with other pastes, with the LM being the exception.
This really wasn't surprising to me as small FFTs is a proper torture test. The closest I got to this in a gaming scenario was GTA5, which for some reason saw low 90s and crushed my CPU temps.
I let everything cool and then I ran a quick Firestrike test to see where it landed.
CPU temp 91c, avg CPU frequency just over 4.1 while set to 4.5 (4ghz Stock)
www.3dmark.com
So, I switched out the paste after a good clean with alcohol. Note, my CPU is de-lidded so I added the sheet on top of the die as well as on top of the heatsink. I did the same with the GPU die, all cut to size as accurately as I could. I'll be honest, the stuff was great to work with. It requires finesse but it wasn't as flimsy as I was expecting it to be.
I was surprised to find my low idle temps a bit higher than before. The minimum is probably more reflective here but it did tend to hover 3-5 degrees more at idle than the paste. This had me a tiny bit concerned.
So then it was onto the torture test, once again small FFTs expecting some sort of melt down.
Never in a month of Sundays did I expect that. If the liquid metal beat that (I can't remember) it was by no more than 1 or 2 degrees. I can't remember having such a low result after 10 minutes. The LM hardly lasted a year before eating and degrading the cooling efficiency so this is a game changer for any laptop as far as I'm concerned.
Lastly, back to firestrike for a run. One of the highest runs I've had in recent years.
CPU temp 78, frequency almost 4.3ghz
www.3dmark.com
Hilariously the GPU score actually tanked. Through all the years of tuning this laptop I've always found there's a fine line to balance between giving the CPU headroom and keeping the GPU fed with current (it's split between the 2 mouths to feed). There was so much extra headroom and potential from the CPU now that there wasn't enough power to keep the GPU going at full chat so it had to pull back. I can play swings and roundabouts by lowering the CPU and getting a GPU increase, but there wasn't much point really.
So yeah.... colour me impressed. If the longevity is anything like reported this is a winner for every laptop I have from now on.
PS.... I bought a big sheet so that I could cut it and use it for all 3 applications. This worked a treat and there isn't a lot of saving going for the one cut to size. If you're confident to cut the size you need you could get a few CPUs/GPUs out of one sheet.
After looking at the thread I thought I would give it a go to see what the experience was like. I've gone through liquid metal and standard paste, more so because the liquid metal is corrosive. I've never had particularly great results with pastes outside of the LM so it definitely works as it should.
Prior to doing anything I left the system for 15 minutes or so and then checked the min idle temps.
I did a run in Prime95 small FFTs before the sheet went on. This is relatively fresh paste with it only being a week old. I had similar experience with other pastes, with the LM being the exception.
This really wasn't surprising to me as small FFTs is a proper torture test. The closest I got to this in a gaming scenario was GTA5, which for some reason saw low 90s and crushed my CPU temps.
I let everything cool and then I ran a quick Firestrike test to see where it landed.
CPU temp 91c, avg CPU frequency just over 4.1 while set to 4.5 (4ghz Stock)
I scored 17 112 in Fire Strike
Intel Core i7-6700K Processor, NVIDIA GeForce GTX 1080 (Notebook) x 1, 16384 MB, 64-bit Windows 11}
www.3dmark.com
So, I switched out the paste after a good clean with alcohol. Note, my CPU is de-lidded so I added the sheet on top of the die as well as on top of the heatsink. I did the same with the GPU die, all cut to size as accurately as I could. I'll be honest, the stuff was great to work with. It requires finesse but it wasn't as flimsy as I was expecting it to be.
I was surprised to find my low idle temps a bit higher than before. The minimum is probably more reflective here but it did tend to hover 3-5 degrees more at idle than the paste. This had me a tiny bit concerned.
So then it was onto the torture test, once again small FFTs expecting some sort of melt down.
Never in a month of Sundays did I expect that. If the liquid metal beat that (I can't remember) it was by no more than 1 or 2 degrees. I can't remember having such a low result after 10 minutes. The LM hardly lasted a year before eating and degrading the cooling efficiency so this is a game changer for any laptop as far as I'm concerned.
Lastly, back to firestrike for a run. One of the highest runs I've had in recent years.
CPU temp 78, frequency almost 4.3ghz
I scored 17 342 in Fire Strike
Intel Core i7-6700K Processor, NVIDIA GeForce GTX 1080 (Notebook) x 1, 16384 MB, 64-bit Windows 11}
www.3dmark.com
Hilariously the GPU score actually tanked. Through all the years of tuning this laptop I've always found there's a fine line to balance between giving the CPU headroom and keeping the GPU fed with current (it's split between the 2 mouths to feed). There was so much extra headroom and potential from the CPU now that there wasn't enough power to keep the GPU going at full chat so it had to pull back. I can play swings and roundabouts by lowering the CPU and getting a GPU increase, but there wasn't much point really.
So yeah.... colour me impressed. If the longevity is anything like reported this is a winner for every laptop I have from now on.
PS.... I bought a big sheet so that I could cut it and use it for all 3 applications. This worked a treat and there isn't a lot of saving going for the one cut to size. If you're confident to cut the size you need you could get a few CPUs/GPUs out of one sheet.
