Ryzen 3000 listed online early on russian site.

[Deleted member 50521]

I think what's being lost in this illogical dicussion of practically is that if AMD improve upon Ryzen+, which is highly likely, they'll have a CPU that Intel can be worried about. They went from 3.8 to 4.4(?) Boost from Ryzen 1 to 2. Ryzen 3 can conceivably go just as far. The cores are irrelevant, the frequency, moreso.

I am also waiting to see how Sunny Cove turns out.

 

[EarthDog]

Nobody lost that point...I think we all agree.

My gripe overall is lack of a need for wide cpus and they are being pushed that way instead of faster clocks and notably higher ipc.

Ryzen went from 4 to 4.3 ghz 1800x to 2700x.

 

[phanbuey]

"Lisa, we need more SPEED!"

Lisa: "Done."

 

[Super XP]

Intel made that mistake with Pentium 4, making a CPU slower than the one before it (Pentium 3), hoping the frequencies will go up so much that it will get faster ( It did, but just barely ).
AMD made the mistake with Bulldozer, creating a CPU which was slower than Phenom before it, but hoping that multicore will really take off. Unfortunately, it didn't.

I doubt any of these companies will make that mistake again.

The Bulldozer was definitely something new and somewhat innovative, but AMD made the mistake and relied too much on automation. If they designed the chips as they did the Athlon 64, Bulldozer might have performed a lot better. Anyhow, Jim Keller saves the day with the Zen micro-architecture. I expect a significant IPC improvement with the upcoming 7nm ZEN, because of where its being manufactured. Can't Wait for an official AMD Announcement.

Bring on the CORES, because this is where AMD can remain quite competitive against Intel. They have the leg up at the moment, and they should capitalize on this for as long as possible.

YI was pretty disappointed that nothing after ivy Bridge actually has disconcernable performance gains. There were only what 4.5 Intel gens after that? Could take a step back and look at k6-3 as well.

Why would they be any gains? Without competition, this is what happens, Stagnation, and now look O_O Intel was caught with its pants down, and that's a good thing. It's time for AMD to show OFF the better product, ZEN, ZEN+, ZEN 2 & the future ZEN 3.,

 

[cdawall]

Why would they be any gains? Without competition, this is what happens, Stagnation, and now look O_O Intel was caught with its pants down, and that's a good thing. It's time for AMD to show OFF the better product, ZEN, ZEN+, ZEN 2 & the future ZEN 3.

I wouldn't go as far to say Intel got caught watching.

They had the new tech, this just sped up delivery. I am 100% a ok with that. You can tell they stop gapped products for whatever they have waiting on the horizon though. So maybe it was a slight surprise that ZEN was good enough to garner actually updating the tired quad core i7 MDT game.

 

[Super XP]

I wouldn't go as far to say Intel got caught watching.

They had the new tech, this just sped up delivery. I am 100% a ok with that. You can tell they stop gapped products for whatever they have waiting on the horizon though. So maybe it was a slight surprise that ZEN was good enough to garner actually updating the tired quad core i7 MDT game.

Intel got caught with its pants down, in terms of they underestimated ZEN's performance.

 

[Solaris17]

Nobody lost that point...I think we all agree.

My gripe overall is lack of a need for wide cpus and they are being pushed that way instead of faster clocks and notably higher ipc.

Ryzen went from 4 to 4.3 ghz 1800x to 2700x.

Do you think forcing higher core counts on the market albiet at a lower clock speed might make for a boom in heavily multi threaded applications? I recall IPC gains is what drove the ghz war to a close. I think their is some merit though the "fruits" of such CPUs is seldom used at current, it could change the tide of software dev frameworks.

imo anyway.

 

[Vya Domus]

Core count is easily scalable, IPC and clocks aren't. It's that simple.

 

[Metroid]

Core count is easily scalable, IPC and clocks aren't. It's that simple.

As long as core count gives an added performance to applications that is what counts for them, as single thread performance has declined since 2000 because it seems more transistors dont equal to more performance like used to, they prefer to add more cores to the die than transistors cause for them core count is more important overall. The question is, why they dont put the the total number of transistors of the 8 cores in just one core, a larger cache and there we go, we have a monstrous 1 core cpu, question is if it will be faster than those 8 cores combined, i guess not cause they have not even tried or they have not bothered about doing that to find out or maybe they know the result is and never talked about it or maybe they even talked about somewhere, however many dont know why they dont do a monstrous single core processor anymore like used to be.

 

[Vya Domus]

The question is, why they dont put the the total number of transistors of the 8 cores in just one core,a larger cache and there we go, we have a monstrous 1 core cpu,

I've already answered why, the scalability of something like that would be atrocious. A CPU core with dozens of ALUs and FPUs and one thread will be significantly slower than an 8 core CPU, there is simply not enough parallelism that can be extracted out of a single thread of instructions to fully max out such a core.

 

[Wavetrex]

I've already answered why, the scalability of something like that would be atrocious. A CPU core with dozens of ALUs and FPUs and one thread will be significantly slower than an 8 core CPU, there is simply not enough parallelism that can be extracted out of a single thread of instructions to fully max out such a core.

It is also the reason why SMT exists, and has existed since the age of Pentium 4 with HT.

There's only so much a line of instructions can be divided and parallelized before encountering a "conditional" instruction which splits the logic to unknown paths (in some cases somewhat known, via even more transistors called "branch predictor" *)
The problem with branch predictors is that in many cases they get it wrong, resulting in the CPU having to start over from last point which wasn't guesstimated.

Basically, in CPU design all "low hanging fruit" have been plucked a long, long time ago, and today all they can do is improve data moving latency a bit here, a bit there, extracting that last 2% of performance.
There is just no wait to make x86 based CPU's faster, or any CPUs faster (including CISC) for that matter, when talking about IPC for a single thread.
And of course frequency (clock speed) can still raise... as materials get better and manufacturing tech improves, but only a little bit... electricity (and light) have a maximum speed, it cannot be made to travel faster, not in this universe.

The ONLY way forward for more performance is at software level, when the software itself is designed to run in parallel on multiple cores.

It's probably possible to scale core counts in the millions in the next 100 years, even on silicon, by making three-dimensional CPU's, assuming the software is capable of dividing itself that much and tech finds a way to cool such 3D CPUs (via micro-water pipes maybe that go through the CPU? IBM has did it, experimental...)

 

[EarthDog]

Do you think forcing higher core counts on the market albiet at a lower clock speed might make for a boom in heavily multi threaded applications? I recall IPC gains is what drove the ghz war to a close. I think their is some merit though the "fruits" of such CPUs is seldom used at current, it could change the tide of software dev frameworks.

imo anyway.

There isnt a choice with the direction things are going now...mainstream platforms are 8c/16t!

That said, weve had hex cores and octos in the market for nearly a decade now and most software still hasnt caught up to really use it. Most games are ok with 4c/8t....which have been out for over a decade. So, IMO, the table was already set for software devs to already be working on these things over the past decade but here we are...still in a world where a quad with HT is enough for 90% of users to have an overwhelmingly positive computing experience.

Edit: this isnt like ray tracing and nvidia right??? Where we have our first card and with luck devs will follow... many core cpus have been in the market for quite a long time and software is well behind the curve even today.

Since amd couldn't trump Intel at the time, they went wide with bulldozer that long ago and started a trend that devs havent been able to keep up with.

For the gaming crowd, who don't do much else, yes.

This is for anyone who isnt a power user, really...not just gaming. Gaming is just a common use for many pcs. This is for the vast majority of people who have a PC in their house who email, web game, interwebs, some photo and video editing (like family vids and pics , not talking professional production.

 

[eidairaman1]

Rumors are Tumors. Just wait till we have it in hand

 

[cdawall]

Core count is easily scalable, IPC and clocks aren't. It's that simple.

Core clocks and IPC scale a lot simpler in applications than core count does. So from a design standpoint 100% easier to just strap on more cores. From a application dev standpoint I believe they might think amd is the devil.

 

[Vya Domus]

From a application dev standpoint I believe they might think amd is the devil.

You gotta do with what you have, it's not like it's intentional that they don't increase clocks and IPC, they simply can't do it to a great extent. At least AMD improved the situation somewhat in the consumer space, better than the mostly nothing that Intel has strapped on their CPUs.

 

[EarthDog]

Intel hasn't needed to 'strap anything on' for the last decade. Perhaps now there is competition, we can see a return to innovation instead of complacency from both camps.

The GPU side of things REALLY needs a high-end competitor... I think we are all praying Navi is within a few/several percent of w/e NV flagship is out at the time.

 

[Vya Domus]

I think we are all praying Navi is within a few/several percent of w/e NV flagship is out at the time.

I would be rather curios to see what's going to happen if they don't. Everyone argued for years AMD's products are crap or that they aren't doing anything and they should get rid of their GPU division, we now start to see what that's really going to be like. Now people are praying that doesn't happen ? Huh, how about that.

 

[Vario]

Wish they never acquired ATI in the first place.

 

[EarthDog]

I would be rather curios to see what's going to happen if they don't. Everyone argued for years AMD's products are crap or that they aren't doing anything and they should get rid of their GPU division, we now start to see what that's really going to be like. Now people are praying that doesn't happen ? Huh, how about that.

Im not sure anyone said that (until the post after yours, LOL!)... surely wasn't me!

That said, much of the same is going to happen. Continued inflated prices in part due to a lack of competition in that market.

Huh, how about that.

lol...always polarizing, you.

 

[Solaris17]

You know I am hopeful intels GPU segment is worth anything performance wise even if just a few points from AMDs cards.

With AMD aquiring ATI and Intel re-dipping into the market (assuming it isn’t vaproware) it makes me wonder if Nvidia would be willing to pay the royalties on x86 to AMD and Intel to dip into the cpu segment.

Though with windows now supporting arm it may not be needed.

 

[EarthDog]

I dont see NV getting into x86... That is an uphill battle if I ever saw one, honestly.

 

[TheoneandonlyMrK]

I dont see NV getting into x86... That is an uphill battle if I ever saw one, honestly.

There's two Chinese firms and a Russian one afaik that have/are doing just that so not impossible if you're backed by a Big backer.
And Navi is mainstream not high end so no highend gpu battle yet" rumoured ".
Seriously, i got an 8 core years ago to the same diatribe of there is no need, all 8 got pounded for years and that chip still meets minimum spec on all games, an Intel four cores(from that era) cannot now say the same, it lives on gaming at 1080 fine whereas no one that called me out on fx8350 buying stuck with their quad anywhere near as long , it cost £159 your daft if you think that a bad buy(6years gaming and crunching).

Bring on the cores and let the pc elite have an actual edge on joe regular i say.

 

[moproblems99]

Nobody lost that point...I think we all agree.

My gripe overall is lack of a need for wide cpus and they are being pushed that way instead of faster clocks and notably higher ipc.

Ryzen went from 4 to 4.3 ghz 1800x to 2700x.

So I get your point and I don't see the need to move from 8C/16T for a little while but....can you post the frequency increases from Intel over the last few generations?

Edit: For quite a few generations, we got barely more than a few mhz here and there. Let's just be happy that it is moving.

 

[EarthDog]

I agree with your sentiment as well. My point there about clockspeeds wasn't to put intel in a more positive light, but to correct a small piece of misinformation about the amd processor boost clocks.

Seriously, i got an 8 core years ago to the same diatribe of there is no need, all 8 got pounded for years and that chip still meets minimum spec on all games, an Intel four cores(from that era) cannot now say the same, it lives on gaming at 1080 fine whereas no one that called me out on fx8350 buying stuck with their quad anywhere near as long , it cost £159 your daft if you think that a bad buy(6years gaming and crunching).

bulldozer and any newer derivative at 1080p gaming is a slug compared to the same gen intel.

If one can actually use all the cores/threads, you have a point. But as I mentioned earlier, games are fine with 4c/8t so most Intel's of that same generation wrap up and smoke amd in most activities. That ~40% ipc deficit (w/e intel had out before ryzen is my comparison) is hard to make up. Obviously this varies by title and settings. So yeah... it works... but to what end? Its a lowered glass ceiling without a doubt.

 

[GoldenX]

Zen2 could be barely better than Zen+, doesn't matter, what matters now is Intel answer to being attacked on all fronts.
It's unbelievable that AMD got to 7/10nm before Intel, after "surviving" with FX for so long. I think it's a combination of good innovation from Zen and lazy complacency from Intel.