Is a 5Ghz Ryzen Cpu coming?

Well yeah, obviously a 5GHz boost clock Ryzen is coming EVENTUALLY. It's not going to be the 3950X or anything else this generation. Even if we get a refresh 3970X in ~9 months I doubt it's more than 100Mhz improvement.
 
I expect the most we'll get this gen (and maybe next as well) is about 4.8, single core. I doubt feelings will be hurt significantly.

If it's cheap enough no one will care.
 
Lol, I'm sure the process will improve and frequencies will go up, that's normal. Not sure how soon it will be and on how many cores or for how long.
 
I don’t really care what the clockspeed ends up being. 5,000,000,000 is just a arbitrary number.

Actual performance is what matters.
This. Some imaginary milestone. If AMD or Intel could get a 4.8ghz CPU with 20% better if I would take that over a 5ghz CPU with crap ipc. Just look at Pentium 4s with high frequency and crappy performance.
 
This. Some imaginary milestone. If AMD or Intel could get a 4.8ghz CPU with 20% better if I would take that over a 5ghz CPU with crap ipc. Just look at Pentium 4s with high frequency and crappy performance.

It seems the 5ghz mark has made a lot of people loose their minds... and memories. MHZ is only a good way to measure performance when comparing apples to apples. Different designs are different.

I mean we could all be running IBM power 6 chips at 6ghz.... cause they made them. But we aren't. lol
 
I don’t really care what the clockspeed ends up being. 5,000,000,000 is just a arbitrary number.

Actual performance is what matters.

Clockspeed is part of a chip's performance. Its not the most important part of performance, but you cannot discount it either. Zen 2's IPC matches or exceeds Intel's offerings across the board, but it still loses in gaming and other low-thread applications due to its clockspeed.
 
Piledriver fx CPUs had no problem hitting 5ghz and nobody liked them..

Power efficiency on ryzen is incredible at 4ghz... Just give me another 20% IPC somehow :p (then we'll want to OC that too cuz nothing is ever fast enough)
 
What's so special about 5 GHz anyway? Frequency is likely to go down on all new nodes. Say hello to many core. It could've happened a decade but now it's just on time for the end of the world party at Rick's
 
Clockspeed is part of a chip's performance. Its not the most important part of performance, but you cannot discount it either. Zen 2's IPC matches or exceeds Intel's offerings across the board, but it still loses in gaming and other low-thread applications due to its clockspeed.

They lose to Intel at low resolutions and situations where the GPU isn't standing in the way... so 10+ year old game engines and for people with 1080p monitors. (who are still buying AMD for the price to performance win without a doubt).

Intels win is a hollow one. Anyone buying an Intel part today because it wins in games is a moron. Plain and simple.

Still when Zen2+ rolls around the improvements can go either way... more clock or more IPC improvements. I know if I had to choose I would take the IPC. Extreme clock speed only helps in niche situations. +IPCs is more beneficial in single and multi threaded workloads.
 
They lose to Intel at low resolutions and situations where the GPU isn't standing in the way... so 10+ year old game engines and for people with 1080p monitors. (who are still buying AMD for the price to performance win without a doubt).

Intels win is a hollow one. Anyone buying an Intel part today because it wins in games is a moron. Plain and simple.

Still when Zen2+ rolls around the improvements can go either way... more clock or more IPC improvements. I know if I had to choose I would take the IPC. Extreme clock speed only helps in niche situations. +IPCs is more beneficial in single and multi threaded workloads.

Clocks still matter in graphics and desktop publishing (Adobe Mostly) and Intel is pretty entrenched there. I wouldnt exactly call that niche but maybe someone else would?
Also stability matters there and the Z390 and X299 platform are currently that.

Maybe it will all change as the differences keep shrinking.
 
I don’t really care what the clockspeed ends up being. 5,000,000,000 is just a arbitrary number.

Actual performance is what matters.

5Ghz is significant because chip manufacturers have been trying to reach it for the last 10 years.

Speed is useful for intuiting the performance of one chip to another in the same processor family, and now even across vendors as the IPC's are about the same (finally).
 
5Ghz is significant because chip manufacturers have been trying to reach it for the last 10 years.

Speed is useful for intuiting the performance of one chip to another in the same processor family, and now even across vendors as the IPC's are about the same (finally).

Didn't AMD reach this mark early in this decade with the turbo on the FX 9590, yet the chip was a colossal failure anyway?

5ghz would be nice given the current zen 2 architecture of course.. Just can't have our cake and eat it ever it seems.

speaking of which, I'd like to see a zen 2 core @ 3ghz pitted against a 5ghz piledriver core to see how badly the FX chips get annihilated :p

These core clocks are obviously a big part of the reason Intel hasn't budged from 14nm+++++. What is the average consumer going to think if the next gen core i[x] only hits 4.4ghz instead of 5?
 
...These core clocks are obviously a big part of the reason Intel hasn't budged from 14nm+++++.

I've been thinking the same thing... Time will tell of course. If nVidia makes a 7nm part and the clocks and performance suffer, it will mean that the 12nm/14nm++/16nm processes are actually the best we are going to get out of silicon, in regards to speed.

Downside is this just means shit will keep getting bigger and more expensive. $1k GPU? Hello new norm.. :(
The upside to larger (more cuda cores/shader units etc) GPU's, is that at least they are very good at parallel processing, and benefiting from those types of increases. The Desktop CPU is another story.
 
I've been thinking the same thing... Time will tell of course. If nVidia makes a 7nm part and the clocks and performance suffer, it will mean that the 12nm/14nm++/16nm processes are actually the best we are going to get out of silicon, in regards to speed.

Really means nothing of the sort- you cannot state that TSMC's "7nm" is representative of all processes marketed as 7nm based on AMD's designs and one Nvidia part. Further, while AMD is struggling with CPU clockspeeds, their GPUs are hitting the same speeds that they were on larger process nodes, i.e., just over 2GHz.

At this point, all we can really say is that AMD has as of yet proven themselves incapable of pushing clockspeeds with their Zen architecture.
 
This latest launch was completely FUBARd by AMD. I lost a huge amount of trust in them, especially since the CPUs are not hitting anywhere near their advertised boost clocks.
 
This latest launch was completely FUBARd by AMD. I lost a huge amount of trust in them, especially since the CPUs are not hitting anywhere near their advertised boost clocks.

that bad? I'm pretty happy with my 3700x, even though you're right and it's not single core boosting. It just plain games with all cores at 4250-4300. While I was expecting 4400 single core boosts, I wasn't expecting to see it able to hold 4.3ghz all core during a gaming session. It's hit or miss, but PBO is clearly working on my chip for the all core performance at least (if i turn it off those numbers drop)

I do want to see my box advertised boost though... And then I'd like to see somebody reverse engineer the AGESA code and lift the [possibly] artificial limits (e.g. 3600 max boost 4.2ghz when 3900x can do 4.6ghz), boy wouldn't that be swell :p, ignoring binning.
 
Anyone buying an Intel part today because it wins in games is a moron. Plain and simple.
Can we avoid these blanket statements? The 9900K is still the fastest gaming CPU. Is it the best all around CPU or best value? No, but for some people that only game, the FPS measurement is all they care about and plenty of people have the money to burn to simply upgrade when it becomes dated. There's no need to insult people because they may have different priorities.

And this is coming from someone who is buying a 3950X.
 
Can we avoid these blanket statements? The 9900K is still the fastest gaming CPU. Is it the best all around CPU or best value? No, but for some people that only game, the FPS measurement is all they care about and plenty of people have the money to burn to simply upgrade when it becomes dated. There's no need to insult people because they may have different priorities.

And this is coming from someone who is buying a 3950X.

I too would still grab a 9900k for a strictly gaming rig. You can bring the cost down with a slightly cheaper board and the price works out to about 3700x territory.

In my case I wanted the future ability to drop in a 16 core (4950x?) because my gaming rigs always get shifted over to my productivity rig after a couple years.
 
This latest launch was completely FUBARd by AMD. I lost a huge amount of trust in them, especially since the CPUs are not hitting anywhere near their advertised boost clocks.
In what way was this launch "FUBAR"? Aside from the clock speed issues with some of the pre-release BIOSes that were used for reviews, what's the problem? RAM compatibility is much better, X470 + B450 support for Ryzen 3000 were available on launch, and the clock speed concerns seem to be getting ironed out quickly (and we are talking about 100-150mhz at most, where you can simply overclock if that's a huge problem for you).

Your statement is a lot of hyperbole. Ryzen 3000 is a way smoother launch than the first gen Ryzen CPUs.
 
Your statement is a lot of hyperbole. Ryzen 3000 is a way smoother launch than the first gen Ryzen CPUs.

It think that's certainly true. But yet, like the 1st gen, one can't help but feeling there's a Zen2 "tock" that is needed. We'll see. Shoot, an "Intel response" might force a "tock" out of AMD.

Got to remember, Intel has a slew of high core count CPUs, just not at the desktop level.
 
It think that's certainly true. But yet, like the 1st gen, one can't help but feeling there's a Zen2 "tock" that is needed. We'll see. Shoot, an "Intel response" might force a "tock" out of AMD.

Got to remember, Intel has a slew of high core count CPUs, just not at the desktop level.

Can't wait to see that "tock".. bit more IPC and clocks? I'm in. The good news for us is that our existing boards should support those chips, which is a definite plus. I mostly got tired of my Intel boards being dead end from day 1. I know there are ways around it, but no support from intel gets annoying.

When you really zoom out, both companies have their annoying quirks.. You just have to pick which annoys you the least.
 
  • Like
Reactions: Auer
like this
They lose to Intel at low resolutions and situations where the GPU isn't standing in the way... so 10+ year old game engines and for people with 1080p monitors. (who are still buying AMD for the price to performance win without a doubt).

Intels win is a hollow one. Anyone buying an Intel part today because it wins in games is a moron. Plain and simple.

Still when Zen2+ rolls around the improvements can go either way... more clock or more IPC improvements. I know if I had to choose I would take the IPC. Extreme clock speed only helps in niche situations. +IPCs is more beneficial in single and multi threaded workloads.

I’d argue that the majority of PC gamers are still gaming at 1080p at 60 or higher hertz. Though, my point, was simply to point out that you cannot just discount clockspeed. It still an important part of the performance equation, even if it is no longer the most important part of it.
 
We aren't discounting clock speeds, we are discounting an arbitrary #. If AMD hits 4.9999ghz it's a fail, even if they get a 10% ipc increase? Of course not, the performance would be awesome. Know what would be better than 5ghz... 6ghz... Or 10ghz... Or... Any other imaginable number. If I put out a cpu that did 1 instruction per clock @ 4ghz (no pipeline), it would kill any 5ghz Intel or AMD. Noted that isn't happening any time soon, but my point is an arbitrary # doesn't mean much. Heck IBM power was hitting 5ghz in 2014, they should be the most popular chip by now using that logic.
 
If AMD hits 4.9999ghz it's a fail

The problem is not that they didn't hit 5.0GHz- it's that they advertised lower, and then didn't hit those lower speeds. Also, they're trading blows in terms of IPC with Intel's now quite old Skylake cores.

AMD has quite a bit of work to do to maintain pace, and the clockspeed limitations of the Zen architecture aren't inspiring.
 
Didn't AMD reach this mark early in this decade with the turbo on the FX 9590, yet the chip was a colossal failure anyway?

5ghz would be nice given the current zen 2 architecture of course.. Just can't have our cake and eat it ever it seems.

speaking of which, I'd like to see a zen 2 core @ 3ghz pitted against a 5ghz piledriver core to see how badly the FX chips get annihilated :p

These core clocks are obviously a big part of the reason Intel hasn't budged from 14nm+++++. What is the average consumer going to think if the next gen core i[x] only hits 4.4ghz instead of 5?
I have that chip and it only does 5GHz if you have above average watercooling and one of like the two or three strongest motherboards available for the platform. It doesn't accomplish nearly as much computational work at that speed as the current Ryzen architecture would with subambient cooling to get it there. I'm not sure if anyone has compared Zen2 vs Piledriver IPC but we already knew that Zen1 crushed it...
 
AFAIK officially they don't have plans for a Zen 2+, right? but IF they wanted to, then in theory yes they could use the refinement of the process to increase those 200-300mhz to the top end or lower the power consumption.
Again the thing is that there is no plan at the moment for a zen 2+, there is the plans for the zen 3 so i would temper my expectations.

And i am currently an AMD fan.


edit to add:
https://www.eteknix.com/amd-details-longterm-zen-cpu-roadmap/
Here's the current roadmaps, as you can see on the 7nm+ they only have plans for Zen3 at the moment. Zen2 is what we have and what we have is Zen2, temper your expectations of anything extra being thrown at it.
 
Clockspeed means nothing without an architecture that makes use of it. This is why Athlon64's at 2.0 - 2.6 Ghz were worlds faster than Pentium 4's at 3-4 Ghz. It's a disproportionate argument. Like saying an engine that spins at 10,000 RPM is more powerful than an engine that spins at 6000 RPM.
 
AFAIK officially they don't have plans for a Zen 2+

If they don't have plans for Zen 2+, then they should put out the 'for sale' sign.

They have plans for Zen 2+.

They haven't announced those plans yet, but there should be little doubt that they likely have pre-production samples of the Ryzen 3000 successor already running in their labs, and the second successor already designed and ready for tapeout.
 
Maybe, maybe not. Didn't they say that they were planning to move to a new socket too?
from what i remember the plans always were for only the first 3 ryzens to be on the current socket and then to move to a new socket + that refinement in the Zen 3.

If i am wrong, then it will be a nice surprise tbh, but i would rather not count on it happening and be pleasantly surprised than to lie to myself about it (and i am pretty sure that they would have known if they were going to do a zen2+ when they released the last roadmaps). We could look up their investors calls to see if they talked about the possibility, but i will admit that it would be too much work for me and today i am a bit lazy ;)
 
Back
Top