AMD's latest Ryzen 7000 chips face severe criticism for their design flaws.
AMD's latest Ryzen 7000 chips face severe criticism for their design flaws.
We can reach a consensus on that, but given the company's goal of maximizing profits, it makes sense they chose this path since their existing user base would likely adapt better. If the opposite were true, AMD's strategy of encouraging cooler purchases might have sparked stronger resistance.
I just added it, didn’t think a separate reply was needed. But honestly, laptop CPUs are quite different from desktop ones, and the same goes for GPUs—this is why NVidia offers models like the RTX 3080 Max-Q and various other versions. You’re not missing much except the idea that high temperatures can harm them, but extreme heat combined with heavy power use causes even more damage. I don’t fully grasp the technical details since I’m not a silicon expert. It seems to be a topic some tech enthusiasts have discussed over time, like Steve from GN mentioned it once. Basically, it’s about how tiny transistors behave under intense power loads. Using everyday examples helps: imagine standard 12AWG wire, rated for about 200°C. Most modern wire jackets can handle a bit more before melting. But if you crank the temperature to 200°C and then apply a massive 100-Amp load at 12 volts (about 1,200 watts), the wire will melt and break almost instantly. This is due to metal’s properties—becoming more flexible and softer under stress. There’s a similar principle with silicon; transistors can manage heat or power, but not both at once. That’s why laptops can run hot and last a long time, while desktops start to feel the heat. As transistors shrink to just 5 nanometers, damage becomes more likely. That’s all I can share. To be honest, 100°C isn’t ideal in general, but laptops usually handle lower temps thanks to their low power draw. I think high-end machines should adopt liquid metal thermal paste as a standard, given how hot and dense these CPUs are becoming.
Correct me if I'm wrong here, I've only built a couple AM5 systems so far for people and haven't had the chance to really sit down and tinker as of yet. Isn't the whole point of Precision Boost just that? Isn't it's purpose in life just to boost the CPU as high as possible within the thermal (and power) headroom given? Does that not mean that even if you did have a thin IHS and the worlds most expensive water block with triple 360 rads... it would just boost as high as it could go if Precision Boost is enabled, up to it's 95C limit regardless? That kind of... makes everything invalid now, no? I mean yeah with the thicker IHS you're going to have a harder time achieving the highest possible clocks, but that "I want to run as fast as possible at sub-80C" party time we had previously is gone now, I think. What with Precision Boost and Intel's TVB coming into play with the 10900K and beyond, the more headroom you give these chips the more they use it. I mean I haven't had to manually overclock a Ryzen system since the 5000s came out thanks to the systems just saying "hold my beer" and running amok. I'm sure I could tune it better if I wanted to put the time in, but I honestly don't have any inclination to do that as of yet. (No one's paying me to do it right now). And of course, part of this lies on the motherboard manufacturers, we can't forget that. ASUS ROG boards and many others come configured to absolutely DUMP voltage into the CPU unless you stop it. I've never seen a CPU work so hard for so little before. Undervolting (I don't like using that term here, because it's not accurate really) has become more of a necessity than a hobby now. Anyways I'm not saying this is a conspiracy or disagreeing with the opinions here, I'm just pointing out that we've seen this coming for a long time. Thicker IHS or not, thermal design is shifting, and these chips are being designed to run at redline now. We won't really know how this will change longevity until a few years from now when things start moving to the used market and new generations come out. I do enjoy your take though @WallacEngineering and I share your worry on longevity. It's been a while since I left the electrical engineering field, but the rules don't change overnight and I'm sure we're going to see something start to shake loose if things keep going like they are. I could be wrong though.
Isn't this because you're increasing the temperature even further? For instance, 100A across a single meter of 12AWG would equate to around 5mOhms, adding roughly 50W of additional heat on top of the existing 200C conditions. Could that really be the reason it would melt? Based on your example, running a CPU at 100ºC and turning it on would generate extra heat that could cause it to fail, wouldn't it? Ignoring external influences (like unusually high ambient temps), isn't heat mainly linked to power consumption? More power means more heat, but with the added challenge that we can't dissipate it as quickly as before due to the high thermal density and tight contact points of smaller transistors. I still think 100C is extremely damaging, whether it's 10A or 100A, and I'd love to see evidence proving otherwise.