Took me 5 years to realize I was thermal pasting wrong

I was at a shop in Austin last month helping a buddy with a build and he watched me put a pea-sized dab on the CPU. He asked why I wasn't spreading it. I told him the pea method works fine. He showed me his thermal camera after boot and the temps were 8 degrees cooler on his identical setup just because he spread it thin with a spatula. Been doing it the lazy way since 2019 and never thought to check. Anyone else find a better method after swearing by the pea drop for years?

3 comments

3 Comments

anthony7633d ago

Oh man, the pea method is actually fine for most CPUs unless you're using a really big chip like Threadripper. Did you check if your buddy's cooler was mounted tighter than yours too?

david_reed222d ago

anthony763 says "unless you're using a really big chip like Threadripper" and that just blew my mind lmao. Wait so you're telling me you actually tested this on a Threadripper and still had issues? Because I built a buddy's Threadripper system last year and we did the pea method, and that thing was a nightmare to get cooling right. We swapped coolers like four times thinking the mount was bad. Turns out we just needed way more paste than we thought lol. I'm honestly shocked you got it to work at all with a pea on a chip that huge. Most people I know with Threadrippers are doing the cross or spread method because of the chiplet layout. Makes me wonder if your cooler just had really good pressure or something because that seems insane to me.

carr.abby2d ago

Right? The cross method is honestly the only thing that makes sense for Threadripper with those chiplets spread out like a checkerboard. I feel like the pea method would leave half the dies barely covered unless you used like a whole gram of paste.