Of course they need to avoid heating up the product that is the reason the scale and density of the chip makes it harder to cool it efficient as it less much less contact to cool it.
Cooling a CPU is a product of the heat being generated, die size, and the coolers abilities. The new server CPUs are going to have close to 500W TDPs and still be able to be air cooled. The reason is the larger surface area on the dies spread out the heat making it easier to remove. Desktop CPU dies are so small that you are now getting an insane level of heat per mm2.
Your post doesn't elucidate why base clocks are dropping several hundred MHz between Zen 4 and Zen 5. The node did change but TSMC N5 and N4whatever are in the same 5nm FinFET family.
its N3 for EPYC, for desktop its N4(P, X, E still unclear), IPC is usually unrelated to node change per se, clocks and power draw are the direct perf areas attached to node improvement, as IPC is mostly dependent on uarch changes/improvements/optimizations (aside from small perf gains in cache speeds etc due to the node).
if i were you i'll wait for 2 things, one 9800X3D (if you game much) and 2, Intel "Official LAUNCH"(in stores), of ARL, as it will more likely than not force a price reduction on current Zen5 lineup, plus never buy products at launch date, its always safe to wait 3-6 months and give chances for kinks and bugs to be fixed/polished (plus the price reduction)
i feel they'll release Z6 for AM5, Z5 is looking more like Z3 was to Z2, an optimization rather than a full grounds up uarch. i think they backed down and pushed the big changes to Z6 with the new N3-N2 node and new packaging etc for the last AM5 hurrah, they need to up the memory channels count in desktop since Zen3, they got away already with 2 gens holding back and probably a 3rd with Z6 but past that the bottleneck of not doing it is massive.
they have shown you further proofs from multiple sites/reviews plus the overall tech community accepted perf numbers, you just keep moving/bending the goalpost, cherry picking test and outliers test results , while just outright misunderstanding what is being said or stated and making a fuss about it trying to falsely disprove known performance numbers, at 65w TDP= 88w PPT for AMD; and for intel go figure what its PL2 is at 105w TDP, 2nd, at 65w TDP for both 7950X and 13900K AMD wins overall, at 105w TDP intel starts to stretch its legs, but under that they simply lose. What jeremy told you was simply this, that below 105w TDP (not PPT or PL MAX) AMD wins, and above it, Intel wins, and that at 65w "TDP" AMD usually ties and sometimes wins vs 13900K at 105w "TDP" more often than not. at no point he claimed AMD wins all (TDP is not the usual max power draw for either of them).
you just shooted yourself in the foot right here, of the 3 the one with the least amount of media power is by far (but REALLY REALLY FAR) AMD, media power, industry power, mindshare, of the 3 even now compared to intel AMD still loses, the fact you directly just made it about review bias just discovered your true intentions and obvious bias against AMD.
I didn't attack anyone, I just quoted the content creators themselves which said that unless they praise AMD, they are done. Hardwareunboxed is a prime example, he made 30 videos about Intel degrading power limits etc. and 0 videos about the zen 4 parts imploding due to high cache voltage. Nough said i think
oh yeah enough said, i already understood what are your intentions, no point in wasting any more time reading you at all....
I'm just quoting what the content creators are saying man. Why are you so aggressive and making personal attacks?
Isn't the Zen 5 CCD made on N4P, while the Zen 5C CCD is made on one of the N3 nodes?
IPC is one of the ways designers reinvest dividends from the node change to get more performance (the other way is just to increase clock speeds). Rocket Lake provides a good example of what can happen when using a bigger & more complex uarch without a node change.
Grossly exaggerated. Millions of people buy these products before that window and most of them do just fine. Not everyone has the same risk tolerance (or lack thereof) as you seem to.
Fixed that, for you. The R5 7600X is currently $200 on Newegg. I know you just can't seem to let go of "launch prices", but as soon as Black Friday rolled around (only 1.5 months after launch), it dropped to as low as $234 and the best price never rose about $250, after that. The first time it hit $200 was almost a year ago!
The reason the price dropped is fierce competition. Check zen 3 prices, the 5600x was over 300 for more than a year, intel had to drag them kicking and screaming with alder lake to drop prices and give support to older motherboards. I had to wait for exactly 2 years to get support for my b350, at that point it would be just stupid to get a zen 3 cpu. The best option was the 450$ 5800x 3d, but with that money I could buy a 12700f + a brand new motherboard. Amd prices are just wack man.
Whatever the reason, it's simply not the case that AMD is currently selling it at $300, as you would have us believe!
I haven't seen pricing information on the 9000 series. Better yet, let's wait until the official launch date and see how much it's actually selling for.
I suppose I shouldn't point out that the 10600K/11600K were ~$280-300 6 core CPUs and that the only reason they dropped in price is that the 5600X was better? It would almost seem like both AMD and Intel are publicly traded companies who try to maximize profits.
Those dropped because they are 3-4 generation old CPUs....retailers are the ones that want to get rid of the last of them they still have around.
No, you shouldn't. You should point out that the 10600k was launched alongside the cheap 6core models, like the 10400f (173$ msrp), the 10400 (200$ msrp), the 10600kf (260$).
I know engaging with you is a waste of time because you love moving goalposts and refuse to admit when you're flagrantly wrong but let's do it anyways.
The 5600 isn't faster than the 11600k but whatever, that's besides the point. Having a better product doesn't mean you have to increase the prices, you can release the damn cpu on time and at the usual 199$ msrp and the competition, if it's worse, has to drop prices.
Facebook
Twitter
Instagram