(Rumor) Ryzen 7000 thermals are as hot as a neutron star

I thought they choose the height they did so that it would remain compatible with the existing coolers on the market so they didn't need to redesign new brackets for everything.

You'd think there would be cheaper ways to raise it up, though. An extra PCB to add the needed height difference with just vias should work.
 
TSMC is admittedly struggling with its stacking process and is still working with BESI on a solution, the yield rate is not something that can really handle a product-wide mainstream launch.
Interesting, seems interesting AMD went with the 5800x platform to do the "limited run" I would think the 5600x would work best for most gamers, but maybe better ROI with the 5800x side of things?
 
Interesting, seems interesting AMD went with the 5800x platform to do the "limited run" I would think the 5600x would work best for most gamers, but maybe better ROI with the 5800x side of things?
Cost is the same for AMD on the 5600x as it would be on the 5800x as they needed to match die size, but on the 5800x they get a shot at the crown, not so much with the 5600x.
 
I thought they choose the height they did so that it would remain compatible with the existing coolers on the market so they didn't need to redesign new brackets for everything.
While true, there are inherent bracket differences, especially if you need a modified back bracket. But the real issue is "the cost" in terms of temps due to the decision.
 
While true, there are inherent bracket differences, especially if you need a modified back bracket. But the real issue is "the cost" in terms of temps due to the decision.
I don't disagree, I don't remember the last time I said to myself, "I am building a new multi-thousand dollar PC, thank god I can save myself from buying a new cooler while simultaneously making my current machine nonfunctional and harder to resell!"
 
Super thick IHS to maintain a not guaranteed compatibility. In other words, they may have made a big mistake.


Holy crap a 20°C drop? Or about 36°F just by removing the clunky piece of aluminum from the top? Definitely one of those situations where AMD did fuck up big time, or largely didn't care about the higher end users who would push it to it's fullest and beyond.

That said, make sure you watch to the end for the "RIP 7700x" :D

edit: Bah! physics guy converted the wrong way.. big Freedomheit temperature change.
 
Last edited:
I can foresee the "Ryzen 7000 Delidding" forum thread here, and derbauer selling lots of de-lid kits. The temp drop with a de-lid (90C -> 71C) is massive enough that some enthusiasts will happily void warranty to not only drop temp but gain frequency / OC headroom. Who knows, AM5 could end being a real tinkering and tuning platform together with all the other new tech (DDR5, etc).

AMD may never outright state why the IHS is thicker than previous, and is arguably bottlenecking heat transfer (at least relative to AM4), but I reckon their engineers would say "the delta doesn't matter, it won't make any realworld difference to most users, and we're gaining a degree of AM4 cooler compatibility".

Nevertheless, the idea of all 16 physical cores humming at @ 5.2 - 5.4Ghz @ 71C is making me feel something.

View attachment 513929

I'm waiting to see what the x3d versions look like first, but if I was buying one of the launch SKUs, I'd definitely but a delidder and bracket.

It would be nice if ek made a mounting kit as well like the one I had for my 4770, but that was more convenience than anything. I'll design and make a standoff setup if needed.
 
If reusing your existing cooler saves you $100, you're half way from a 7900 to a 7950x, a sacrifice for compatibility is worth it for most folks. If the huge temperature difference makes a difference, if it's less than 5%, nobody will honestly care.
 
  • Like
Reactions: Axman
like this
If reusing your existing cooler saves you $100, you're half way from a 7900 to a 7950x, a sacrifice for compatibility is worth it for most folks. If the huge temperature difference makes a difference, if it's less than 5%, nobody will honestly care.

Also, if the "standard" temps for the regular parts is ~5-20 degrees C higher than the last gen parts, but the norm for their own gen, then 3D-stacked parts can run a lot hotter than the last gen, and get better effective cooling than the current non-3D parts.

People keep saying that 95 at full load is bad, but what if it's there to normalize people? What if it's also there to incentivize people to move on to the refresh parts sooner?
 


Looks like you an drops temps a TON while keeping the same performance.

Here's a guy who undervolted the CPU quite a bit and got a 40 degree (C) drop with a small performance hit, but running at 55°C seems kind of cool. Note, I don't know if the guy is a complete hoser or anything, it just was a recommended video considering all the similar ones I've watched recently.
 
My 3900x cpus will often be 90c when running all core prime workloads with D15 or even a good 240 aio with stock settings. To me thos isnt a surprise. Also have not had any of them degrade or die even running 24/7 like this for weeks at a time.
 
No problem with my hardware running 100c if it isn't a threat to chip. Run it hot run it fast.
I agree. People need to move out of a closet if they worry about heat out put. My PC is in wide open basement and I don't notice a rise in temp running my 5950x and 3080ti balls to the wall.l let alone gaming.
 
Love the undervolting and underclocking for optimum computational efficiency. That's what most of us do in DC (not Washington).
 
That's good to hear. Now we just need to reel in these mobo prices some.
The B550s should be nice, most of us at home don't need all the bells and whistles of the 670s, B550E will be more than enough for most consumer workloads, with enough room left over for the entry-level enthusiast. The 670's are for people who want a Threadripper but can't get a Threadripper.
 
Why are they comparing to a 5950x at 3853 MHz? Doesn't it boost higher than that? Mine boosts pretty much exactly to that all core in Cinebench right now, but mine is screwed up. It used to go higher.
Ha, I just found the video you pulled the screenshot from where it showed a 5950x boosting to 3853 MHz all core in Cinebench. It's from OptimumTech. I had also researched other people having their 5950x chips with low boost and he had the same problem. Looks like he just rolled with it. As an aside, I believe that problem is motherboard related. I have a whole thread I started on the topic as mine suddenly started that behavior one day and I haven't been able to fix it. Hopefully soon, I'll be able to test my chip on another X570 board to prove it.
 
Holy crap a 20°C drop? Or about 36°F just by removing the clunky piece of aluminum from the top? Definitely one of those situations where AMD did fuck up big time, or largely didn't care about the higher end users who would push it to it's fullest and beyond.

That said, make sure you watch to the end for the "RIP 7700x" :D

edit: Bah! physics guy converted the wrong way.. big Freedomheit temperature change.
I think AMD and Intel need to start selling CPU's again without IHS. AMD did back in the Athlon XP era, and it was fine.
 
No it wasn't. There were a ton of people cracking cores and slipped screw drivers ripping through MB but I do agree that they should have a bare die sku.
If you slipped a screw driver you're gonna damage something. This is why you just design a better heat sink mount and not leave it to the aftermarket to solve it.
 
Yeah, when they started using heat spreaders... shipping the chips naked again would be a return to the Athlon days of old.
At least on the consumer side, I seem to remember the athlon 64 being the first heatspreader chip from them. Ah, the memories
 
Back
Top