AMD Thinks Next-Gen Ryzen Chips May Run Even Hotter

The FPS Review may receive a commission if you purchase something after clicking a link in this article.

Image: AMD

Is the Ryzen 8000 Series going to run any cooler than their current, Zen 4 counterparts? Sadly, probably not, as David McAfee (AMD Corporate VP and General Manager, Client Channel Business at AMD) was recently asked by QuasarZone during a media roundtable as to whether AMD would be solving its “temperature issues” anytime soon—namely, an irony where Ryzen chips feature a lower TDP but higher temps than Intel’s CPUs—and shared his belief that “high heat density will be maintained or further intensified.” harukaze5719 illustrated this concept last year, sharing a comparison that showed how Intel’s chips compared to AMD’s Zen 4 offerings in the power consumption and temperature departments.

QuasarZone: One of the criticisms regarding AMD desktop products is CPU temperature. The CPU power consumption is clearly lower than that of competitors, but the temperature is higher. Will these temperature issues be resolved in the future? Wouldn’t it be possible to induce heat dissipation by attaching a dummy die next to the CCD die?

McAfee: We are working closely with TSMC to put a lot of effort into process technology. At the same time, we must be able to guarantee the quality and stability of semiconductors. As more advanced processes are used in the future, we believe that the current phenomenon of high heat density will be maintained or further intensified. Therefore, it will be important to find a way to effectively eliminate the high heat density generated by such high-density chiplets in the future.

Also, if you look at the TDP 65 W product, it has excellent overall performance. Through these product examples, I believe that this is an important factor to consider when planning a future roadmap to ensure a good balance between TDP and heat generation.

Join the discussion in The FPS Review Forums...

Discussion (12 replies)

Join Discussion →
Grimlakin

That is interesting. I wonder if it's the addition of CCX's that are causing the higher power consumption.

Brian_B

"Grimlakin, post: 78230, member: 215" wrote:

That is interesting. I wonder if it's the addition of CCX's that are causing the higher power consumption.


No, the interesting thing is that it's lower power but higher temps.

Zarathustra
Zarathustra

I'm not convinced just running hot is a problem.

You just have to select the substrate appropriately, so the heat cycles don't kill anything, and design the cooling appropriately.

Elf_Boy

Less energy = more energy?

I wonder if this is a hot spot kind of thing - the Intel CPU package spreads the heat better so it has an even temp where the AMD package has it all in just a few spots over the chiplets making those spots hotter even though over all heat output is lower?

Is so then AMD could address with a better (more expensive) cpu package.

E

Maybe it’s time to go back to bare dies. Add a shim if you’re worried about cracking the die. Just getting rid of the IHS would make a big difference

Zarathustra
Zarathustra 👍 1

"Elf_Boy, post: 78242, member: 438" wrote:

Less energy = more energy?



I wonder if this is a hot spot kind of thing - the Intel CPU package spreads the heat better so it has an even temp where the AMD package has it all in just a few spots over the chiplets making those spots hotter even though over all heat output is lower?



Is so then AMD could address with a better (more expensive) cpu package.

I don't know for sure, but I suspect this has something to do with the thicker heat spreader they went with for AM5, combined with the tight fitment of chiplets making it more difficult for the heat to escape.

If I had to guess, they did this to allow for enough space to insert bigger 3D caches in future models, and they didn't want to have to have a separate cooler for those models.

The truth is, "running hot" doesn't really matter that much anymore. We know that most CPU's can withstand their TJmax of 90-100C (depending on the model) or indefinitely without problems. In the past we would want to bring the temp way down for better overclocking, but modern CPU clocks don't scale with temperature as much as they used to, so what's the point?

Brent_Justice
Brent_Justice 👍 4

If it's designed to run at xxxC, then it's fine.

People worried about it dumping more heat into your room is not how things work, Wattage is key, not temp.

Zarathustra
Zarathustra

"Brent_Justice, post: 78256, member: 3" wrote:

If it's designed to run at xxxC, then it's fine.



People worried about it dumping more heat into your room is not how things work, Wattage is key, not temp.

Oh, I didn't realize that was the worry

Yeah, that goes without saying.

Elf_Boy

My old Bulldozer system that was very much a concern. My 8150 (later 8370) and could heat my entire 2 bedroom apartment by itself. This was not so good in the summer.

Were about to test if my current system can heat my 1 bedroom condo though a winter.

Zarathustra
Zarathustra 👍 2

"Elf_Boy, post: 78259, member: 438" wrote:

My old Bulldozer system that was very much a concern. My 8150 (later 8370) and could heat my entire 2 bedroom apartment by itself. This was not so good in the summer.



Were about to test if my current system can heat my 1 bedroom condo though a winter.

Yes, but that's not what we are talking about here.

Take two scenarios:

Scenario 1:
CPU power use is 95W. CPU runs at 60C

Scenario 2:
CPU power is 95W. CPU runs at 90C.

In both of these cases the CPU will deliver the exact same amount of heat to the room, as the exact same amount of heat is being produced by the CPU.

The only difference is that in scenario 2, something is going on that is not allowing the heat to escape as fast, and because of this, it is running hotter.

The only difference in the impact on room temperature is that the hotter CPU will actually heat up the room slightly more slowly, as it's not transferring that heat to the room as efficiently.

LazyGamer

"Zarathustra, post: 78247, member: 203" wrote:

I suspect this has something to do with the thicker heat spreader they went with for AM5


Most of AMDs 'heat' issues with AM5 are centered on making the IHS even thicker than it was on AM4, because the new LGA socket sits lower and AMD wanted to keep the same package dimensions for OEM heatsinks (for those that weren't aware).

"Endgame, post: 78244, member: 1041" wrote:

Maybe it’s time to go back to bare dies. Add a shim if you’re worried about cracking the die. Just getting rid of the IHS would make a big difference


This makes a huge difference, but it's unlikely that AMD (or Intel) would want to take the risk.

E

they may not want to take the risk, but they may not have a choice as nodes continue to shrink.

Tsing Mui
News poster at The FPS Review.

Recent News