AMD Zen 4 CPU with 5.2 GHz Boost and RDNA 2 iGPU surfaces
Click here to post a comment for AMD Zen 4 CPU with 5.2 GHz Boost and RDNA 2 iGPU surfaces on our message forum
moab600
Hope AMD can reach even higher speed, though IPC and architecture improvement are the best...
Intel probably gonna reach 5.8GHZ and 6.0GHZ with 300W+...
Venix
speed of light = 299.792.458 m/s 6ghz is 6 000 000 000 hz ....so by the time a nand gate switches from 0 to 1 in 1 cycle (1hz ) the electricity that for the shake of simplicity we assume is equal to the speed of light had time to travel only 5cm .... ! Even if we overcome the friction and tunneling and be able to push past 10ghz or something we will hit a barrier with the speed of light ! random thought i had XD
mohiuddin
^ wow.
Fediuld
That is why we need real quantum computers 😀 where light it not the issue.
k3vst3r
Thought electricity was 1/100th speed of light through electrical wires.
rl66
Then unbeatable plaid speed...
kanenas
Very likely to do so as this is ES
AuerX
So will computers eventually turn in to time machines once we break the speed of light?
fantaskarsef
Depends on the medium it travels through, and it could be faster under certain conditions (especially lower temperature). What they're trying to do with super conductivity.
This is true, but pardon my noob question, aren't we still way ahead of the barrier you described?
Let's assume lightspeed, for the simplicity, was the maximum for our electron / energy to travel, or say, put a 0 to a 1 inside that switch / gate. I'm 100% sure that this is not how to name it, but let's just humor those who really are pros on this 😀
Light speed's 3 x 10^8 m?
A gateway's width is about 15nm for DDR5, right? Which is roughly 1,5 x 10^-8 m ?
So theoretically, as the electron would have to "travel" 15nm, it would only take it 5 x 10^-17s?
Wouldn't that technically, very, very simplified, mean that we're talking about a barrier in the range of about 10 Peta Hertz? So about 1 million times more than a GHz?
Sorry, asking this not because I want to discredit you, but because I think that barrier's way further ahead than the limitations of our resources to construct said hardware, which brings us further, further away from reaching that perfect speed of light and closer to what k3vst3r said, and probably below this. Even if we consider 1/100th the speed of light for electrons to travel, we're still about 10.000 (ten thousand) times ahead of a GHz until we reach that limit?
Sorry guys, I'm just bored at work.
Mpampis
We're nowhere near the speed of light barrier guys, measurement units do play an important role you know.
H83
If AMD can offer an eight core capable of reaching 5.0Gh on all cores without producing too much heat, that´s perfect for me.
I like high speeds but no need to exagerate, specially on CPUs with multiple cores.
Rattlehead99
We already have a transistor working at 604Ghz, yes 604. And antennas working at 300Ghz. The travel distance of electrons is minimal, you don't send them through a tube, they are already next to where they need to be, they just need the transistor to open, they move less than a nanometer.
You can view the transistors in a CPU like one big transistor made from many smaller ones in a parallel connection)which is why they can't break 5Ghz easily and why they don't scale linearly with voltage at and above relatively low speeds(3-3.5Ghz)
The best CPUs have very wide cores with medium-high frequencies like 2-3Ghz, similar to the M1 right now, or back in the day Athlon 64(K8) and Athlon XP(K7) which have/had mediocre frequencies, but for their respective process nodes and times very big cores. The minus is that they are very expensive and not good for servers.
Servers need many small, weak, power efficient cores, which is what AMD and Intel have been making for the past 15 years or so.
BLEH!
Hopefully it has more than 8CUs/512 GPU cores...
Venix
@fantaskarsef @k3vst3r @Rattlehead99
fanta ! nothing wrong with your post it was napkin math to begin with and for the rest i am saying if we somehow solve the tuneling and friction / heat issues and we are able to push the ghz waaaaay past what we are now then eventually in this magical scenario lightspeed will be the barrier ...and that assumed light spee = electrecity speed ! i did not claim we are close too ... it is still mind bogling that you only travel 6 cm with the speed of light until the gates switch once !
@k3vst3r i thought it was over 0.9 of the speed of light with small variations depending on the mediums resistance ...apparently i was way off on that too !
@Rattlehead99 i am not sure antenas apply all you have to do is be able to read the syn wave but again i am not a physicist !
and to all of you i very welcome corrections ! talking about such things is fascinating !
Valken
WHEN?! Because I cannot find a 5800X3D and if it is within 2 months, I will hold out for this!
asturur
Will be end of year / november.
If it was 2 months we would have pictures of retail boxes already
Rattlehead99
The Radeon 680m in the R7 6800U/H is 12CUs already at 6nm(7nm++) and that's for laptops. There were leaks that said that Zen4 APUs will have 16-24CUs.
fantaskarsef
Yeah it's weird... and the next mind blowing thing in relation to this: We discuss such things, and honestly, just two generations ago, with average education, we probably would not have the education to even discuss this. The common knowledge about physics to even vaguely imagine the stuff here, and even getting a mindset to think about these things, have developed so recently it's cool that we can discuss it here!
Also, @Rattlehead99 : Yes you are absolutely right, electrons don't even have to travel that far! Which only makes it theoretically more hehe.
But yeah, I think the physics is a lot more complicated, But I just like to let the mind slip into those realms with the fellow gurus here, the physics behind that which is our hobby is pretty cool after all.
CPC_RedDawn
Well this is ES sample and Lisa did show Zen4 running Halo Infinite and said that the CPU was running at 5GHz ALL CORE which is extremely promising and that was quite a while ago now so this 5.2GHz could also be ALL CORE load clocks. I predict it will hit 5.5-5.6GHz single core boost and an all core boost of 5.0-5.2GHz with a base of 4.5-4.7GHz
I think Intel will boast a lot about their next CPU's which I believe will have massively increased clocks for the P cores possibly around the 5.8GHz mark if not higher. With their E cores also getting a small bump in clocks but even better efficiency. They need to focus on that for their next lineup, Alder Lake got destroyed in reviews for being way too hot and power hungry.
ES sample. Rumours are pointing towards the desktop chips having RDNA3 architecture not RDNA2 😉 This sample could just be a placeholder for testing purposes.
Airbud
https://miro.medium.com/max/3840/1*3goWdZza8eR26JFW60PP6A.jpeg
:D
