Ryzen 3000 listed online early on russian site.

[SL2]

Talking mainstream/hedt here... not server/xeon/opteron.

1090T january 2010

i7 980X march 2010


Can't believe it's 9 years ago.

 

[Metroid]

The Phenom II was comparable and in some cases better than C2Q. It was the Phenom that was behind. Phenom II x4 was released 2009 x6 2010, C2Q was released 07. Phenom (one) was released 07. The C2Q didn't intro into Phenom, Phenom was after C2Q.

As I stopped using amd cpus since 2004, i remember I checked that phenom on many places and I really thought it was a tri core using 3 more threads and a marketing gimmick from amd, few days ago somebody said it was a real 6 cores and went looking on it and it was really a six core, although in many places they said it was a tri core at that time and even some places today they say that. If I really knew it was a true 6 cores i would have bought cause the price was very good, price performance was better than intel overall. Intel's 6 cores cpu was very expensive at that time, more than $1000.

 

[SL2]

As I stopped using amd cpus since 2004, i remember I checked that phenom on many places and I really thought it was a tri core using 3 more threads and a marketing gimmick from amd,

That sounds like FX/Bulldozer to me. Phenom II had real cores, basically an Athlon X2 with L3 cache and more cores. The same goes for Phenom I.

 

[Vario]

As I stopped using amd cpus since 2004, i remember I checked that phenom on many places and I really thought it was a tri core using 3 more threads and a marketing gimmick from amd, few days ago somebody said it was a real 6 cores and went looking on it and it was really a six core, although in many places they said it was a tri core at that time and even some places today they say that. If I really knew it was a true 6 cores i would have bought cause the price was very good, price performance was better than intel overall. Intel's 6 cores cpu was very expensive at that time, more than $1000.

They had a 3 core Heka, was often a defective (or sometimes functional and unlockable) quad deneb K10 phenom II that was locked into a triple core. It represented a budget option. The Phenom II was a good capable processor. FX Bulldozer was the turning point where they fell seriously behind for many years until today's Ryzen closed the gap. First gen FX was even slower than Phenom II at the same clock.
https://www.anandtech.com/show/2832/8

 

[toyo]

Really hard not to get excited by this, if true, which it sure seems like it. Makes my 8700K seem like an i3. Don't mind it, hopefully by the time new cheaper GPUs get made on 7nm and Intel answers these Ryzens, it will finally be reasonable to upgrade from a financial perspective.

 

[lsevald]

8 physical cores or more makes the most sense to me since we have had 8 separate running threads available in our gaming computers for many years now. Thinking developers might target that number when they need as many cores as possible for say High or Ultra quality in a game. If there is a 2-300MHz penalty to go 12 cores or more, I might prefer a 8c/16t cpu.

One of the things I really appreciated the last couple of years, is how stable my computers have been. If this is due to win7/10 or mature Intel based hw I don't know. Last time I used AMD, I was on Win XP. Hopefully this won't change if I upgrade to an AMD based PC next year

 

[lexluthermiester]

Thinking developers might target that number when they need as many cores as possible for say High or Ultra quality in a game.

There's just one small flaw with that logic; Most games need far more GPU processing power than CPU for graphical quality. Thus the need for powerful GPU's.

 

[robot zombie]

Personally, I think 6 cores is killer for your average "desktop" user. I'm routinely astounded at the sheer range/number of tasks my 2600 will tackle simultaneously without a single hiccup. I get so much more done, not just in work, but leisure, too. I have a Ryzen 3 machine AND a 2nd gen Ryzen 5. I can do the same things on either, but the difference in how I operate is there. The flexibility of the 2600 is great, even if you're not an enthusiast concerned about processing capability. App to app, Windows 10 makes great use of these 12 threads, and little by little, apps and games are getting better optimized to use those capabilities internally the way Win10 already does across apps and the system as a whole.

It's not so much getting things done faster with better multi-threading imo (that may be more attainable later, again with better in-app optimization) right now, it's the transition from task to task - not having to wait for this to finish before you start that. And starting this not bogging down the "that" you already have going. It just makes everything more seamless and stable. Everyone at least beyond gramgramz and her solitaire can benefit. Most use their PC's for a wide range of tasks. Some cannot be combined, and others you don't really need to, but that doesn't mean that being able to combine more tasks doesn't ever come in handy. In fact, I think it's always gonna improve one's experience on the machine, even if it's not in a way that can be reliably benched. It's one of those things where you try to go back and the difference becomes hard to ignore.

It's like... ...it may be hard to see the value in that now, but when hardware capabilities increase in novel ways, people often adapt new habits around them in unexpected ways. And then devs see that and better gear their software to streamline that way of operating. And little by little the way we do things gets better in ways nobody ever sees coming. To that's part of what makes tech so cool, to me, anyway. Discovery. Newer, better paths to the same goals.

And the thing is, while 4 cores is more than serviceable, 6 or even 8 is still a perceptibly appreciable improvement. It's subtle but important. You don't realize where multitasking capabilities are holding you back until you're free to surpass what you're used to, because you've trained yourself to operate comfortably within your hardware's limitations. And that's not to say that quad cores are BAD at multitasking by any stretch. If that is what's available to you and is enough for you to have a good flow doing whatever it is you do, that's what you should go with. But what if you could have better flexibility for less money at little to no compromise on the single-core performance you need?

I mean, a lot more can be done with cross-core stuff than has been yet, while clocks and IPC have been smashing the same wall for so long that we're about to hit the point where there will be little to no variation between best and worst options on that front. It's becoming less meaningful of a comparison. With improvement being so stagnant there, multi-core aspects will naturally become the distinguishing features, even if not with Ryzen 3. That time is coming.

It's one thing when it costs significantly more - but that falls apart when you look at something like the 2600, which gives you 6c/12t at perfectly reasonable clocks/ipc for ~160 dollars. There are situations in that spot where I can see even a typical user benefiting more from having something like that than say a blistering-fast quad-core for another chunk of paper. Those 12 threads may be a screen door on a submarine, but the 6 physical cores are absolutely getting used over the course of the machine's service life. And it probably costs a little less.

With speed and ipc there are diminishing returns, too. Just like with crazy-high core counts, not everyone will be able to notice or make use of a couple bumps there. Not all of us are high-refresh-rate gamers or doing heavily CPU-bound rendering or whatever. Not everything runs appreciably better/more stable just because a CPU happens to be that much faster, even if the difference is measurable. Many things are already fast or simply don't speed up much more, much like nvme SSD's are not always the upgrade you'd think they'd be coming from a 4x slower SATA SSD. I think its wise to ask yourself whether you really need something that is singularly fast over something maybe a bit more flexible and all-around more steady and efficient. Gas-guzzler or diesel?

The way I see it, we're moving in a direction where conversations about whether we need more cores will be nullified, likely should we ever have available to us CPU's that both have more cores and at least match the clocks/ipc of the lower-core-count speedsters of yesterday aaand possibly even be a better value. Whether you think you need the cores or not, you benefit. Dual-cores become dinosaurs, while fast quad cores become entry level! Even if it's a luxury for you, that luxury comes free, or because it comes free, what you need is now available for less in order to still compete and sell.

That, to me, is the ideal form of progress in tech. Over time, you get better performance for what you pay while yesterday's best goes down in value simply by merit of comparison to what is available now. That's how it should be. Obsolescence through innovation. That is why I think people are getting so excited. Most are more than happy to pay what new tech is worth, but nobody wants to spend the same money they spent on their last iteration of a part to get a sidegrade, or pay for new iteration that performs as well as the tier above the last gen version did for the same money that hardware ran you when it was new. Modern computing got to where it is today because of advancements towards making better tech cost less.

We'll see. For me, the excitement isn't over the fact that I can have more cores for my money. Now that I have them, I'm never going back, but honestly what I think would be truly awesome to see would be AMD actually competing with Intel on efficiency, clocks, AND raw ipc, core counts be damned. Throw a couple extra cores per price bracket on top and you will have some seeriously competitive offerings.

I dunno, to me it represents interesting possibilities. There are legit arguments to be had over whether everyone needs a $300 8-core. But what about a $240 6-core? Or a $150-$170 4c/8t that's fast enough to game and still a kickin multitasker? If this is truly the new precedent, the question becomes not "Do we need these cores?" but instead "What will we do with all of these cores?" Lotta potential usage there - but then, actually it doesn't matter much right now. The only real justification needed for the market to take it in is price and all-around performance, multi and otherwise. That's the appeal being garnered here - the hope is that the merit of it goes well beyond more cores.

Sometimes I think part of the critical attitude towards "moar coarszz!" comes from the fact that AMD tried to push that agenda one before, and the things were hot garbage for the money... ...I mean they were really very hot. Completely lopsided performance and usage profiles. Not putting words in anyone's mouth, but there's gotta be a sour taste there for those who remember. This time around it's more than a nice idea though. They've already started to deliver with Zen+. That's your proof of concept. And with a significant node shrink on top of refinements to an already known-good design, I don't see why Zen 2 wouldn't be a very nicely-balanced bag of chips. It's all gonna come down to pricing. And with the jump to 7nm, that really could go either way.

Like I said, it's about what's available to you for the money. I'm very interested in seeing if AMD doesn't turn around and provide many people the most suitable options available for the money, cores notwithstanding. They do that, they win this battle. And then maybe Intel comes back with something good!

 

[EarthDog]

1090T january 2010

i7 980X march 2010


Can't believe it's 9 years ago.

So...yeah... 9 years and 10% adoption. While 80% use quad or less.

 

[TheoneandonlyMrK]

So...yeah... 9 years and 10% adoption. While 80% use quad or less.

I think the threads popping up on here(site) with quads, intels quads having issues in games is proof enough quads are approaching eol for gaming, yes those users could Oc or update the cpu but ht adds effective cores , fours been moving towards the edge a while, this year its getting to it imho.

 

[EarthDog]

As I've said around the site, a hex is the sweet spot now. 4c/8t handles the VAST majority of games. Only a small handful show improvements with more cores and I can think of one that scales with more than 8.

That said, 8c/16t is incredibly futureproof if bought. More is a waste of money and why I'm annoyed we see more from amd on mainstream

 

[cdawall]

As I've said around the site, a hex is the sweet spot now. 4c/8t handles the VAST majority of games. Only a small handful show improvements with more cores and I can think of one that scales with more than 8.

That said, 8c/16t is incredibly futureproof if bought. More is a waste of money and why I'm annoyed we see more from amd on mainstream

Yea well everyone needs that new 3990wx or whatever with 64/128. I know I do. How else will I browse TPU?

 

[robot zombie]

As I've said around the site, a hex is the sweet spot now. 4c/8t handles the VAST majority of games. Only a small handful show improvements with more cores and I can think of one that scales with more than 8.

That said, 8c/16t is incredibly futureproof if bought. More is a waste of money and why I'm annoyed we see more from amd on mainstream

No, you're absolutely right.

I see both sides of it though. More than 8 might be a waste but those being more prevalent bodes well for the availability of what is actually needed in most cases. The presence of 12, 24, 1000, whatever as a high-end, premium option places the now ideal 6 completely in the midrange... the value of having them decreases this way, and still the performance is good and usable.

I think most people, even those less in the know will be looking at 8, max. Everyone knows the rest are niche products. I think AMD knows that too, just from looking at where their bread has been buttered this past year.

So to me if they wanna put out a few chips with crazy core counts, let them. People are only willing to go so far for their needs. The value-availability tradeoff comes into play. Just because you can have more doesnt mean youll take it. It feels better to be able to get what you optimally need for less than it does to spend extravagantly just because the market is sudenly such that it is now possible to do so. Most people will take the former when asked.

 

[Mussels]

View attachment 114237

he was talking monitor resolution

 

[Super XP]

There's just one small flaw with that logic; Most games need far more GPU processing power than CPU for graphical quality. Thus the need for powerful GPU's.

The CPU tells the GPU what to do, i.e., passes information to the GPU in most games I believe.
But would be interesting if somehow some GPU tasks such as Physics could be offloaded to the CPU. The CPU with all its cores have muscle that can be utilized efficiently.
Let the GPU deal with all the ultra high images and the like.

How many hundreds of thousands of people do you know? That's what it is going to take to change the steam survey.

I would assume the majority of Steam users can't really afford much perhaps? lol

 

[Final_Fighter]

he was talking monitor resolution

lol, made my day.

 

[GoldenX]

320x240 ought to be enough for anybody.

 

[SL2]

I'm pretty sure AMD will present some new CPU's based on ZEN 2, but this confuses me.. IMO it's bad choice of words.

When you read high-performance computing, do you think

1 - High-performance computing, supercomputers.

2 - This.

 

[xkm1948]

I'm pretty sure AMD will present some new CPU's based on ZEN 2, but this confuses me..
View attachment 114242

When you read high-performance computing, do you think

1 - High-performance computing, supercomputers.
View attachment 114241

2 - This.

Definitely HPC means Supercomputer clusters.

I bet 10 kitties it is just gonna be launch of Zen2 server chips and brief mention of RyZen 3000

 

[SL2]

I bet 10 kitties it is just gonna be launch of Zen2 server chips and brief mention of RyZen 3000

Yeah, EPYC Rome with its chiplet thingies hasn't been launched yet.. only previews so far.

It's 9 months since the launch of Ryzen 2000, I don't really expect AMD to end it so soon..

 

[phanbuey]

Yeah, EPYC Rome with its chiplet thingies hasn't been launched yet.. only previews so far.

It's 9 months since the launch of Ryzen 2000, I don't really expect AMD to end it so soon..

They might because they know intels got nothing, and if it's fast enough people and oems will jump on it.

The last thing they want to do is give intel time to recover. Especially since their boy Jimmy Keller is somewhere in an intel basement cooking something up.

 

[SL2]

The last thing they want to do is give intel time to recover.

The same can be said about server CPU's, which I guess is much more lucrative than desktop, especially these days.
I think they've already said that server comes first, and I have no idea if AMD/TSMC is capable to ramp up both in such short time.

They've already launched Zen+ based mobile CPU's this week, and Zen 2 server CPU's are expected. I'm not sure they'll present in even more CPU's today, seems like a lot.

My bet is on a preview, but no launch, similar to EPYC Rome in November.

I hope I'm wrong though.

 

[Metroid]

few hours before the end of intel and its 4 cores till the end of the world concept, it lasted 10 years, it is time to end it.

 

[SL2]

few hours before the end of intel and its 4 cores till the end of the world concept, it lasted 10 years, it is time to end it.

Yeah we've heard that oversimplification a few times now. Even the 8700K is over a year old, and the 1800X is close to two years old.

 

[phanbuey]

Time flies.