AMD Zen 4 desktop processes up-to 16 cores, 170 W TDP
Click here to post a comment for AMD Zen 4 desktop processes up-to 16 cores, 170 W TDP on our message forum
Matt26LFC
Wasn't surprised that Zen3 had no core increase, little surprised Zen4 won't get a bump
TheDeeGee
170 Watt you say?
My 11700 uses about 40-45 during gaming, my old 4770K used 95 ^^
nizzen
And?
I want 400w+ powerdraw cpu with crazy good performance 🙂
Performance > powersave
Agonist
Ok. And my 5600x uses 65w @ 4.7.
Your point is dumb.
Silva
Gaming is generally a light load most modern CPUs don't push over 25% performance.
Do a render or some other intensive task and measure the real consumption when the CPU is at 100%.
Plus, your 11700 is 8 cores, 170W should be 16 cores.
I'm actually surprised if this is not bs or there is not an iGPU, why the power increase on common desktop CPUs?
Canada hit temperature records and some places in my country also hit the same temperatures.
I still want a place to live in 10/20 years, so efficiency is important.
Also those 400W don't just turn into zeros and ones, it turns into heat that has to be dissipated on your room.
I do feel a difference and I only have around 250W of TDP on my PC, imagine people who have double or more of that...
I guess in winter you don't need the heater!
If you really need performance you already have HEDT platforms pushing around 300W, don't look at common desktop components.
fantaskarsef
Shouldn't a 5nm CPU use less power than a 7nm CPU if both are comparable in architecture (e.g. both from AMD)?
Or do they use the extra power to boost frequency?
svan71
"Canada hit temperature records and some places in my country also hit the same temperatures." so you believe a record temp based on temps going back one hundred years is reason for alarm on a planet that is how old? Your "record temp" is an excuse for you to be record taxed, with zero to show for it.
Undying
All i need is zen4 8cores and some crazy fast ram 5000+ and this can be a nice upgrade from zen1/2.
Denial
It's a good thing ice core samples exist that show us historic temperatures and CO2/Greenhouse levels then.
Sorry but this isn't quantum physics. We can measure CO2 levels in the air pretty easily and simple experiments, ones you can literally do in your house, show CO2's effect on infrared radiation. Global temperature averages, whether you think they are relevant or not, are trending upwards towards the 2-3c and ~25M higher sea levels we've seen during the pliocene period for this level of CO2 (400ppm). We're shooting way past that - aiming for a nice 600ppm by 2100 all happening in your 100 years of recorded vs several million. Pretending this isn't going to have, or already has, a massive effect on various populations is absurd.
In the meantime a lower TDP costs me less money in both electricity and AC cost. It's just a net benefit to everything as @Silva said.
jura11
Another Climate change denier?
schmidtbag
I'm not surprised. Consumer-grade CPUs aren't going to need more cores than what AMD offers for another few years. If there is no IPC increase or high peak clock speeds, then we have more of a reason to be concerned.
Undying
V-cache without any architectual change will be adding 15% more performance. AMD confirmed zen4 will use it with new arch improvement and benefit from lower latency and fast ddr5 ram. This should be 40% increase over zen3 at minimum. It would smash.
kapu
Excellent point ..............................-- in another world another dimension maybe 🙂
Undying
But in this one tittle was "waste of sand" 😀
BlindBison
How relevant of a negative would this be for games? I've heard some people like "TechDeals" claim that more cores and threads can improve system responsiveness and "smoothness"/frametimes in games, but then other testing outlets claim there's essentially no gaming performance improvement in most titles beyond/above a certain core/thread count at this time. DOOM Eternal was one of the very few cases where I saw the 3900X could beat the 10900K so maybe the future with better multithreaded games core/thread count will continue to matter more than it does now, but going off the tests I've seen raw single threaded throughput still matters more for games after you reach a certain core/thread count at least.
More cores/threads means more heat/power draw, no? For example, doesn't the 3900X run hotter than the 3700X which in turn runs hotter than the 3600? Personally I think I prefer they keep the core/thread count the same and just do what they did with Zen 3 where they focus on making the cores and threads they've got faster/reducing latency between CCXs, etc. Correct me if I'm wrong of course, but seems like that would be the more beneficial route for gaming. The uplift between Zen 2 and Zen 3 for games was huge and for my uses 16 cores/32 threads is overkill unless more games start being programmed like DOOM Eternal (hopefully they will be). For example, I was playing Resident Evil 8 which is DX12 iirc yet that game didn't appear to scale with core/thread count well at all and most of my 3900X just sat idle looking at HWiNFO.
Sort of off topic, but the PC version of RE8 could seriously use a round of patching (broken film grain/broken anti-aliasing/weird slow motion issues after alt tabbing on my system/some users I saw were reporting missing animations here and there depending on their rig seems like/CPU utilization on high thread count chips seems relatively poor/I get asset streaming stutters running through doors and I've got a pretty beefy rig/changing graphics settings on the fly tanks performance for some reason -- Digital Foundry confirmed many of these issues in their video while Modern Vintage Gamer demonstrated that the current DRM implementation is responsible for a lot of the game's stuttering which is totally unacceptable).
JamesSneed
If the highest end16 core part is able to hit 5Ghz on all cores everyone will be like 170w is not bad. Its all relative.
JamesSneed
Considering how much extra transistor budget 5nm will give that 40% number could actually be low. I expect the same huge IPC increase for Intel when they move to there 7nm which should be a bit more dense than TSMC's 5nm.
Farasicnsee
Ah, the name-calling retort. Why not use 'skeptic', or 'realist'? Do I get to call you a 'Chicken-Little Alarmist' (of 'the sky is falling fame)? Why don't you try to go to sites that are downgraded by google b/c they contradict the AGW narrative? The reporting of scientific results shouldn't be treated/listed as lies just b/c they contradict expectations. THAT IS NOT SCIENCE!!! It is a result that should be checked to see if it can be reproduced. All science should be able to be freely questioned and examined.
Short story - the temperature records have been altered or erased by people looking for reasons to adjust away said contradictions to their favored theories, the RCP scenarios are wildly out of touch with reality. If you look at them closely, they are based on models that contain a vast number of conditions/variables that cannot be measured (WAG inputs) and 90% of which grossly overestimate any warming. In addition, most of the paleoclimatological proxies smooth out temperatures by a factor of 50-100 years (or more), and the ice core records only resolve (at best, for the most recent layers) to an 8-16 year smoothing, so annual variation is not really well represented, the peaks (and valleys) were likely higher.
The latest heat records were not caused by 'global warming', but by a shift in the jet stream that amplified the effects of local geographic influences.
Go look at wattsupwiththat dot com - I won't post a link, b/c it has been labeled as a 'denier' website, even though the majority of its articles are by ppl genuinely interested in getting the facts straight. One of the recent articles is 'the state of climate science in 5 minutes'. Go there and look.
D1stRU3T0R
Wasn't PCIe4? This is some good news, at least the 5 meme can live on
AM5, 5nm, DDR5, PCIe 5 (quick tell me something with 5 exept 5GDDRX so we can have an 5 (V) logo as we had the 5 logo for asrock and with amd)
https://www.asrock.com/mb/photo/980DE3U3S3(M1).png
Denial
It is freely questioned and examined.
Hmm I wonder if changing global temperatures could cause wobbles in the jet stream? For the record the science isn't out on this one yet.
I'm sorry, I meant freely questioned and examined by scientists.
_
It's labeled as a denier website because Anthony Watts it's a climate change denier. He failed out of college and now he's failing out of life because he's a loser. All of his stupid blog posts have been thoroughly debunked time and time again.
This whole thing is offtopic but I'm just going to live you a few links that refute the claims you posted here - I'm not going to explain them because honestly it's not worth my time - had to battle enough of you people on reddit.
https://www1.ncdc.noaa.gov/pub/data/ghcn/v3/techreports/Technical Report NCDC No12-02-3.2.0-29Aug12.pdf
https://newsroom.unsw.edu.au/news/science/no-bureau-meteorology-not-fiddling-its-weather-data
https://skepticalscience.com/print.php?n=1414
https://judithcurry.com/2014/07/07/understanding-adjustments-to-temperature-data/
https://www.deccanherald.com/content/61233/uk-climategate-inquiry-largely-clears.html
Again it's offtopic so this is my last post on it.
Idk, Intel didn't really gain much IPC, it was mostly just AVX512 improvements which doesn't do much for general computing. I think 40% for Zen 4 is really high. I'd expect closer to 20%. I think IPC gains are going to get much harder going forward.
