Speculation: Ryzen 3000 series

[ub4ty]

I see and I agree, I think we are approaching the apex of what can be expected for improvements.

This has happened before. The focus then just shifted to storage/memory improvements and interconnects therein. Lots of improvements to be made in this area with lots of subsequent performance boosts.

I think we are approaching a point where less will be more, like very low powered machines that do everything thrown at them good enough for nearly everyone.

We already went through this phase with embedded/mobile computing and Intel dropped the ball big time allowing ARM to corner the market. Next up is a new batch of competitors and RISC. AMD nor Intel will be able to impact this trend.

Things are moving away from the CPU not towards. Low power embedded/mobile is dominated by ARM due in part to the largely open/licensable/customizable tech. Intel and others keep centering on proprietary and absurdly expensive and it simply continues to keep them out of volume adoption. Look how long Optane storage has been out and how little it was adopted. Meanwhile, standardized NvmE is king. Greed feeds.. your opponents.

I used to think there would be some sort of manufacturing change that would totally reinvent the PC business however I can’t imagine an application for home use that would require a massive step up in performance.

Well, you can always look to enterprise for what lies ahead for desktop consumers. Most of the action in enterprise is in storage technology and interconnects.

I’m sure we’ll see crazy stuff regarding big data machines, I’m not so sure we’ll see anything but cheaper chips with lower lifespans on the consumer end.

IMO, the big data cycle is almost near an end. Cloud computing/big data have been rolling along for some time on the same tired ethos. Something big will come along and disrupt it just like it disrupted the previous stagnant period in computing.

Lifespans aren't changing. However, compute is becoming much cheaper as it should over time. What the desktop user needs is a new use case for all of this compute and they likely will get one.

 

[Kenmitch]

Not really, even if AMD beats the crap out of Intel for two years they’ll get lazy and intel will comeback hard. That’s the way things work.
Good to see new processors with actual new technology in them instead of the steady intel tweaks we’ve been seeing for the last 5 years.

I don't really think the future holds what you envision. Guess time will tell in the end. We at least have to wait and see if the latest chatter on AMD is true or not. If they manage to pull off what the leaks are implying then the future looks much brighter than most would have guessed. Either way it'll be interesting too see what happens.

 

[ub4ty]

I don't really think the future holds what you envision. Guess time will tell in the end. We at least have to wait and see if the latest chatter on AMD is true or not. If they manage to pull off what the leaks are implying then the future looks much brighter than most would have guessed. Either way it'll be interesting too see what happens.

Yea, if AMD pulls off what this leaks imply, I will be pleasantly surprised. That being said, I think they're bogus for a number of reasons. In just a week, we'll have the answers. So, I'm ignoring all of these click-bait 'leaks'.

 

[DarthKyrie]

I don't really see why AMD would choose to beat a 9900K with a 3600X for half the price. Surely anything that beats the 9900K on a similar platform (AM4) should have at least a similar price. Anything else would be daft financially.

Market share and mindshare are the 2 most important things to AMD right now. Market share ensures that people will program with AMD in mind instead of only worrying about Intel. Mindshare is needed in order for market share to happen, and the quickest way to earn mindshare is doing what AMD is doing with Ryzen.

 

[PotatoWithEarsOnSide]

It'd be nice if the reviewers actually published their test methodologies in advance of their reviews this time, as when the 9900K came out there was an absolute circus of ridiculousness exposed.
AMD certainly don't need the negativity that initially came with the 9900K, which upon further examination was established to be down to the dubious practices of the mobo manufacturers, yet Intel took the rep damage. That incredibly dubious paid for advanced review didn't help either.
Any benchmarks revealed at CES need to be clear and fair, simply because we're all tired of dirty tricks by now. I'm probably asking for too much here, primarily because all PR and Marketing departments are full of devious barstewards.

 

[ub4ty]

It'd be nice if the reviewers actually published their test methodologies in advance of their reviews this time, as when the 9900K came out there was an absolute circus of ridiculousness exposed.
AMD certainly don't need the negativity that initially came with the 9900K, which upon further examination was established to be down to the dubious practices of the mobo manufacturers, yet Intel took the rep damage. That incredibly dubious paid for advanced review didn't help either.
Any benchmarks revealed at CES need to be clear and fair, simply because we're all tired of dirty tricks by now. I'm probably asking for too much here, primarily because all PR and Marketing departments are full of devious barstewards.

I doubt anyone has a dire need to break their neck to acquire and utilize these next gen processors when there are current gen that run all the way up to 32 cores. That being said and given this reality, it's best you wait until sound reviews come out before judging a product completely. If everyone practiced this vs. going nuts over leaks and pre-reviewers, there would be no incentivize to push ridiculous write-ups. Look at the whole Nvidia RTX debacle. Everyone warned people and all the red flags were there. Still people broke their necks gobbling them up... and got what they had coming to them.

Ryzen and its performance is already established and is the benchmark. The value is there. The question then becomes how much more can AMD bump the performance in their new gen over their old gen. TBQH, Intel is nowhere on the scope for me in these comparisons. I could care less about the 9900k and what it achieved because its a $539 processor that cost more than AMD's 16 core. It's a dead product and its not even moving volume. A group of die hard intel fans who like to spend money on gaming hardware for stats are buying it. Others have moved on to AMD already.

The last thing on my mind is what Intel is doing. I could care less. They're irrelevant currently. All I expect is details about the product line-up and specs. Bonus if prices are discussed. Bonus if a little bench demo is done. The real review/benchmarks will come once the product has been launched and is in a range of reviewers hands. I like anandtech tbqh for good CPU reviews. If you're expecting or looking for things before then, you should expect to get swindled.

 

[Topweasel]

It'd be nice if the reviewers actually published their test methodologies in advance of their reviews this time, as when the 9900K came out there was an absolute circus of ridiculousness exposed.
AMD certainly don't need the negativity that initially came with the 9900K, which upon further examination was established to be down to the dubious practices of the mobo manufacturers, yet Intel took the rep damage. That incredibly dubious paid for advanced review didn't help either.
Any benchmarks revealed at CES need to be clear and fair, simply because we're all tired of dirty tricks by now. I'm probably asking for too much here, primarily because all PR and Marketing departments are full of devious barstewards.

There has been enough evidence that the PL2 setting for the 9900k is the default detected setting on top of Intel telling all reviewers to make sure the PL2 is set to 210. The white paper suggests that the defualt should be 1.5* TDP, but that is a note appended to table and not an actual stated part of the CPU configuration. I don't want to make this a war between fan groups, but I find trying to blame the mobo manufacturers a bit a deflection. That should be considered the defualt activity of the 9900k and not a nefarious way the mobo manufacters try to make the mobo's look better.

 

[mattiasnyc]

Market share and mindshare are the 2 most important things to AMD right now. Market share ensures that people will program with AMD in mind instead of only worrying about Intel. Mindshare is needed in order for market share to happen, and the quickest way to earn mindshare is doing what AMD is doing with Ryzen.

I don't disagree, but suppose new Ryzen CPUs come out and one of them performs better than the 9900K. The 9900K is now at $540 on Newegg and Amazon. You want to tell me that this new 9900K-killer is going to sell for $270?

How much does AMD's currently fastest CPU cost?

I think we can safely assume that if it is logical to price a future Ryzen that's better than the 9900K at half the price then any current CPU that's clearly less good should currently be priced at less than half.

Is that the case?

 

[DarthKyrie]

I don't disagree, but suppose new Ryzen CPUs come out and one of them performs better than the 9900K. The 9900K is now at $540 on Newegg and Amazon. You want to tell me that this new 9900K-killer is going to sell for $270?

How much does AMD's currently fastest CPU cost?

I think we can safely assume that if it is logical to price a future Ryzen that's better than the 9900K at half the price then any current CPU that's clearly less good should currently be priced at less than half.

Is that the case?

AMD's current fastest CPU on AM4 is the 2700x which retails around $250. With that said when they first released Ryzen the 1800x could be had for $500, AMD priced it at $500 because Intel had nothing in the regular desktop market that could touch it because Intel only gave consumers a max of 4 cores for almost a decade. Intel's equivalent 8c/16t was priced at almost $1000 and could only be had on HEDT and their $500 MB. So AMD pricing their product at half of what Intel charges isn't new.

If AMD had the kind of market share that they had back when they released the Athlon64 I could see them pricing the 9900k killer much higher, but Intel made sure that AMD lost all that share with their illegal tactics.

 

[Shivansps]

Ryzen 3 3300 at $99 still makes no sence to me, they JUST LAUNCHED the worthless 240GE at $75 w/3CU, were they are going to place Picasso? The refresh of Raven Ridge is Picasso, a rebranded 2400/2200G... If they launch a 6C for $99 were they are going to even place those APUs?

Best case escenario having the 2400G at $99 is still a awful idea because there is no need to split the volume intro 2 skus like that. And of the "big APUs" (Picasso) were going below $100 why on earth they launched the Athlons 220 and 240 with just 3CU. Something is wrong here.

 

[ub4ty]

I don't disagree, but suppose new Ryzen CPUs come out and one of them performs better than the 9900K. The 9900K is now at $540 on Newegg and Amazon. You want to tell me that this new 9900K-killer is going to sell for $270?

How much does AMD's currently fastest CPU cost?

I think we can safely assume that if it is logical to price a future Ryzen that's better than the 9900K at half the price then any current CPU that's clearly less good should currently be priced at less than half.

Is that the case?

I don't care about pricing until AMD specifies the pricing.
There is no sense in debating/discussing this as no one beyond AMD has an impact. As a consumer, I'll ultimately decide whether or not their pricing makes sense and that will determine if I buy their products or not. To be quite frank, as an owner of a 1700 that I paid about $180 for, I'm not willing to pay much more for an 8 core. $270 to low $300s sounds about right and as I already stated : Intel doesn't set market norms anymore. AMD set the pricing norms. So, go buy AMD's pricing. I have no clue why people keep bringing up the 9900k as if its some benchmark staple of 8 core computing... It's not. It's a last ditch effort to stay relevant that has an arrogant and nonsensical price. No one in their right mind should pay $540 for it. So, there's no reason for AMD to set prices based on it. Intel is irrelevant. AMD has itself to compete against regarding their previous lineup.

How much does AMD's currently fastest CPU cost?

https://www.microcenter.com/product...pper-1950x-34-ghz-16-core-tr4-boxed-processor
$540 - 16 core

https://www.amazon.com/AMD-Threadripper-2990WX-Processor-YD299XAZAFWOF/dp/B07G25SD1P
$1,700 - 32 core

Time to read through the thread and become acquainted with the current market norms.

 

[StinkyPinky]

Perhaps we are coming to an end on what a single core x86 can do (on silicon), but no reason why parallel computing cannot continue to improve and become more efficient and cheaper.

And, of course, I still think ARM has a lot more left in the tank even with single core performance.

 

[mattiasnyc]

I don't care about pricing until AMD specifies the pricing.
There is no sense in debating/discussing this as no one beyond AMD has an impact.

Ok, fine. I didn't say you cared and nobody's forcing you to talk about it. As a matter of fact, I'm not sure why you're talking about it if it's pointless.

Of course, none of the many things that are unknown to us are going to impact what AMD end up doing, so if that's the bar for discussion we don't really have to talk about ccx's and I/o chips and core counts or anything else that isn't officially stated either.

I'm not willing to pay much more for an 8 core. $270 to low $300s sounds about right and as I already stated : Intel doesn't set market norms anymore. AMD set the pricing norms.

That's really not true though. As long as Intel sells a large chunk of processors it too sets the market together with AMD.

I have no clue why people keep bringing up the 9900k as if its some benchmark staple of 8 core computing... It's not.

People are probably using the 9900K as an example because of how it performs. We can argue that it's overpriced, but it performs better than most Intel CPUs aimed at the same market (i.e. not "HEDT" necessarily).

Time to read through the thread and become acquainted with the current market norms.

If that's aimed at me I'll just let you know my question was rhetorical.

 

[ub4ty]

Perhaps we are coming to an end on what a single core x86 can do (on silicon), but no reason why parallel computing cannot continue to improve and become more efficient and cheaper.

And, of course, I still think ARM has a lot more left in the tank even with single core performance.

Single core performance stopped being a thing some time ago beyond a goofy niche Intel tried to carve out for obsessed people who game a little uncompetitively. Most proper software is multi-threaded, scales excellent, and is numa aware. In data centers and professional environments, the clocks are far lower on both CPUs and GPUs to lower power consumption and run the processor in a far more efficient range. In this realm, multi-core performance trumps all. People who tend to rant about single core performance are typically Gen-Z get off my lawn die hard intel fans. Stuff north of 3Ghz is just fine for DDR4 pairing. At such a point, more cores and things like great ram and nvme drives matter.

At 8 cores, you're literally at enterprise computing from some years ago. Most people have no use for this. Capture cards and dedicated encode/decode hardware is best for efficient video flows like streaming. No one cares about the difference between an Nvidia encode/decoded video stream vs a CPU encoded streams of a person's goofy Fortnite match.

All-in-all, it's just a bunch of noise that can be ignored. The new standard is moar cores and I/O capacity. Overclocking to insane levels is a niche hobby and always will be. Processors are put in a ridiculously inefficient mode of operation to satisfy someone's desire to feel like they are in ultimate control of their platform. No one does this in a professional setting. I'm all for someone proving this wrong but I just find the screeching surrounding single core performance to be absurd.

ARM and RISC have much more to go. AMD is topping out the traditional role of CPUs in the next two years. Innovation will come from elsewhere as it always does in computing over time. The action is currently in Storage/Memory and interconnects. No one in enterprise or professional settings is yapping about single core performance.

 

[mattiasnyc]

AMD's current fastest CPU on AM4 is the 2700x which retails around $250. With that said when they first released Ryzen the 1800x could be had for $500, AMD priced it at $500 because Intel had nothing in the regular desktop market that could touch it because Intel only gave consumers a max of 4 cores for almost a decade. Intel's equivalent 8c/16t was priced at almost $1000 and could only be had on HEDT and their $500 MB. So AMD pricing their product at half of what Intel charges isn't new.

No I think one of us is missing something;

The proposition was that the new AMD chip would outperform the 9900K and then be priced at half that at the same time. It wasn't that AMD sells some chips at half of what some of Intel's chips are priced at.

You're basically just reinforcing my point: They currently have a CPU that's roughly half of what the 9900K is, but does not perform better. If this logic was all correct the 2700x should be priced lower right now.

In addition you're saying that AMD upped the price for the 1800x because there was no competition for that market segment from Intel. Well if AMD's next CPU can outperform the 9900K and Intel has nothing else in that segment then that again gives the incentive for AMD to price it high.

And comparing the HEDT platform to AMD's AM4 makes zero sense. If you want to go down that road then compare AMD's 8-core x399 to Intel's offerings.

In other words: History shows us AMD won't release a CPU that outperforms the 9900K at half its price.

 

[ub4ty]

Ok, fine. I didn't say you cared and nobody's forcing you to talk about it. As a matter of fact, I'm not sure why you're talking about it if it's pointless.

Of course, none of the many things that are unknown to us are going to impact what AMD end up doing, so if that's the bar for discussion we don't really have to talk about ccx's and I/o chips and core counts or anything else that isn't officially stated either.

People are probably using the 9900K as an example because of how it performs. We can argue that it's overpriced, but it performs better than most Intel CPUs aimed at the same market (i.e. not "HEDT" necessarily).

If that's aimed at me I'll just let you know my question was rhetorical.

Well, if you read through the thread, there's no reason to continue the same tired discussion about price over and over where you're nowhere near close to an actual value. The discussions about CCX's and I/O chips occurred and it brought up interesting technical reflections. A discussion on price has occurred as well. No need to beat a dead horse.

That's really not true though. As long as Intel sells a large chunk of processors it too sets the market together with AMD.

I suggest you go back some pages where I posted the sales numbers for AMD/Intel going into the holidays. No one's buying Intel's overpriced junk. AMD is outselling them in a major way across the board. Why? Because people eventually wake up to the far better value. In such a way, Intel doesn't impact the market. AMD has 2 prior series of dominance to compete/price against. They are the market leader. When I look at Ryzen 3 performance/pricing, it will be against Ryzen 1st gen and 2nd gen and pricing therein. I could care less what Intel is doing and their pricing because they're irrelevant and aren't the market leader. The logic is quite easy to follow : I'm not going to pay $540 for a Ryzen 3 series 8 core when I can buy a Ryzen 1950x (16 core) for $540. See how ez that is? What Intel prices their goofy 9900k processor at has no bearing on my decision/evaluation.

People are probably using the 9900K as an example because of how it performs. We can argue that it's overpriced, but it performs better than most Intel CPUs aimed at the same market (i.e. not "HEDT" necessarily).

Performance is not a stand-alone topic. Performance/Value. If I can buy a 16 core with double the I/O, why in the world would I buy some dated 8 core Intel processor for more money that they tossed out of the back of a truck in desperation using 14nm process on a dead socket/platform? People bring this processor up because its seemingly the only thing one can reference as to any relevancy intel has left. It's a dead discussion because HEDT beats it in performance... And if someone claims they're all about performance outside of a leisure activity like playing vidya, they don't reference an 8-core when there is 12/16/24/32 core available. If you walk into a professional computing center, do you think they're going to be talkinga bout a 9900k? No they wont... because its a niche underperforming power sucking hotrod processor used by Intel to remain relevant that serves the most niche of niche people who like to have overpowered gaming rigs who don't care about power utilization/heat/ or performance value.

 

[ub4ty]

No I think one of us is missing something;

The proposition was that the new AMD chip would outperform the 9900K and then be priced at half that at the same time. It wasn't that AMD sells some chips at half of what some of Intel's chips are priced at.

You're basically just reinforcing my point: They currently have a CPU that's roughly half of what the 9900K is, but does not perform better. If this logic was all correct the 2700x should be priced lower right now.

In addition you're saying that AMD upped the price for the 1800x because there was no competition for that market segment from Intel. Well if AMD's next CPU can outperform the 9900K and Intel has nothing else in that segment then that again gives the incentive for AMD to price it high.

And comparing the HEDT platform to AMD's AM4 makes zero sense. If you want to go down that road then compare AMD's 8-core x399 to Intel's offerings.

In other words: History shows us AMD won't release a CPU that outperforms the 9900K at half its price.

History shows...
Current pricing
8 core - 1700 w/ heatsink fan $170
8 core - 2700 w/ heatsink fan $270
16 core - 1950x $540
16 core - 2950x $700

Now, try and help me figure out what Intel has to do with a clear window of pricing that AMD has already established in their own product line that they have to "compete" against. Explain to me where a $5xx+ 8 core processor from AMD fits in the above pricing scheme. No one knows whether or not the newgen will beat the 9900k and no one cares tbqh. People care about how much better it will be vs 2 generations of AMD's own product line.

Ryzen is a platform from dual core all the way up to 32 core and now 64 core.
HEDT and beyond is relevant when someone begins carrying on incessantly about 8 core performance or silly single core performance. Why? because it clearly separates out someone whose just carrying on about spec sheets from someone who has an actual professional and serious compute workload that requires real performance. Relative 8 core performance is irrelevant because you can walk out and buy a 16 core or 24 core or 32 core right now if you need "so much performance". The big question is not performance anymore. It's performance/value and price. You have options all the way out to 64 cores coming soon. The question is the performance/value and price. How much do I have to pay for x,y,z performance?

Anytime I hear the 9900k brought up as some type of benchmark, I know I'm talking to someone who has no serious compute needs and is just blowing hot air. Single core performance is even a more laughable concept... it's a hold over from the bulldozer fiasco.

 

[moinmoin]

Not really, even if AMD beats the crap out of Intel for two years they’ll get lazy and intel will comeback hard. That’s the way things work.
Good to see new processors with actual new technology in them instead of the steady intel tweaks we’ve been seeing for the last 5 years.

Let me use some dreaded analogy, this time from sports using the league system (those without a separate knockout stage at the end). A season is long, it's hard to keep concentration and motivation up for every single match day even though one bad match day can make the difference for the final standings. There are two popular ways to achieve that concentration and motivation: 1) Looking at the competition and trying to keep them at distance/to catch up at any given point. 2) Ignore the competition, just looking from match day to match day and try the best each. The former allows one to preserve energy (=perceived laziness) but can backfire. The latter can burn out but is at the best untouchable for the competition.

Now with Intel and AMD the competition is a funny one, they have to try to compete not with the competition of today but plan ahead to be able to compete with the suspected competition in 3-5 years. Intel apparently planned their future based on being untouchable due to their slow but steady improvement both in chip design and process nodes. They took a risk with the latter without an equivalent plan B, while having locked their chip design to the process nodes. That meant they were hard locked to an old chip design on an old process node that they, as a plan B on a short notice, kept improving. As they were still significantly ahead there was no urgency to actually restore the previously steady improvement both in chip design and process nodes. AMD on the other hand was way behind and had to think of different tricks to catch up, to close the huge gap. One obvious area of attack Intel left them was the core count. AMD already tried this with the construction cores, but the whole setup with the modules was seen as inferior by the public and failed spectacularly at the performance as well. Now, considering an advantage with the process node and frequency was very unlikely, they had to at the very least create true cores, better SMT, and better scalability to significantly more cores to make at least a tiny dent in Intel's market domination mindshare wise. This is Zen 1/Zeppelin. For Zen 2 the projected competition was Ice Lake still at a better node than AMD would have access to.

So in the end Intel tried approach 2) but failed to keep it up at the end, took 1) and as a result lacked the urgency to restore 2) quickly enough. AMD in their effort catching up to an Intel capable of keeping up 2) tried all possible tricks they could afford with the little money and man months they still had, and ended up catching up quicker than planed. Now with the changed circumstance AMD's best way ahead is keeping up their current modus operandi while no longer looking at the competition as the goal to reach, so switching from approach 1) to 2).

 

[ub4ty]

Let me use some dreaded analogy, this time from sports using the league system (those without a separate knockout stage at the end). A season is long, it's hard to keep concentration and motivation up for every single match day even though one bad match day can make the difference for the final standings. There are two popular ways to achieve that concentration and motivation: 1) Looking at the competition and trying to keep them at distance/to catch up at any given point. 2) Ignore the competition, just looking from match day to match day and try the best each. The former allows one to preserve energy (=perceived laziness) but can backfire. The latter can burn out but is at the best untouchable for the competition.

No reason to reference the convoluted sports industry.
The tech industry is just fine. In tech, you survive by innovating.
Tech is cyclical. Sometimes hardware is driving software innovation. Sometimes software is driving hardware innovation. Material science and physics dominates progression capacity in hardware. Architectures are settled on many years in advance. Software has lead hardware for some time. Now hardware will lead software.

Now with Intel and AMD the competition is a funny one, they have to try to compete not with the competition of today but plan ahead to be able to compete with the suspected competition in 3-5 years.

All major tech companies have to plan 3-5 years out as a standard practice.
There are too many factors to name regarding what influences this. The major thing is to ensure you're innovating internally and in tune with consumption/market. Your competitors are important but they aren't going to come and do development for you.

Intel apparently planned their future based on being untouchable due to their slow but steady improvement both in chip design and process nodes. They took a risk with the latter without an equivalent plan B, while having locked their chip design to the process nodes. That meant they were hard locked to an old chip design on an old process node that, as a plan B on a short notice, kept improving. As they were still significantly ahead there was no urgency to actually restore the previously steady improvement both in chip design and process nodes.

Intel got lazy and greedy. It happens to all well established big companies. This will never change. It's as natural as birth and death. In tech, this allows for someone conducting true innovation/progress to eat their lunch and rightfully so.

AMD on the other hand was way behind and had to think of different tricks to catch up, to close the huge gap. One obvious area of attack Intel left them was the core count. AMD already tried this with the construction cores, but the whole setup with the modules was seen as inferior by the public and failed spectacularly at the performance as well. Now, considering an advantage with the process node and frequency was very unlikely, they had to at the very least create true cores, better SMT, and better scalability to significantly more cores to make at least a tiny dent in Intel's market domination mindshare wise. This is Zen 1/Zeppelin. For Zen 2 the projected competition was Ice Lake still at a better node than AMD would have access to.

AMD saw what anyone should have seen in the industry.. PHY is used everywhere to scale compute at high speeds and low latency. So, why not with Core complexes. It brings cost down significantly and allows for huge scaling. There was no cheap trick conducted. It's a natural progression given the limits of die cuts. AMDs plans have always been ambitious. They just ran out of road temporarily enroute to HSA and unified GPU/CPU computing. If Intel had a longer term plan like this and executed on it as opposed to sitting back and milking the market, they would be in far better shape. Alas.

So in the end Intel tried approach 2) but failed to keep it up at the end, took 1) and as a result lacked the urgency to restore 2) quickly enough. AMD in their effort catching up to an Intel capable of keeping up 2) tried all possible tricks they could afford with the little money and man months they still had, and ended up catching up quicker than planed. Now with the changed circumstance AMD's best way ahead is keeping up their current modus operandi while no longer looking at the competition at the goal to reach, so switching from approach 1) to 2).

Intel got lazy and greedy like most big established companies do and a smaller more agile company ate their lunch. It happens all the time and thankfully so for progress's sake.

There's Option 3) Innovate.. Innovate .. Innovate and don't become a lazy greedy barrier to progress and you will always thrive. No corporation should last forever or dominate nor can they , so eventually you die. Such is the circle of life. Birth->death->Birth.

Live, make your mark, and then get out of the way so others can make theirs.
Try and be a brick wall in the way of innovation and progress and eventually a train will come plowing through you.

 

[mattiasnyc]

When I look at Ryzen 3 performance/pricing, it will be against Ryzen 1st gen and 2nd gen and pricing therein. I could care less what Intel is doing and their pricing because they're irrelevant and aren't the market leader. The logic is quite easy to follow : I'm not going to pay $540 for a Ryzen 3 series 8 core when I can buy a Ryzen 1950x (16 core) for $540. See how ez that is? What Intel prices their goofy 9900k processor at has no bearing on my decision/evaluation.

So what? I don't think this is about you. It's about the market.

Performance is not a stand-alone topic. Performance/Value. If I can buy a 16 core with double the I/O, why in the world would I buy some dated 8 core Intel processor for more money that they tossed out of the back of a truck in desperation using 14nm process on a dead socket/platform?

Which AMD CPU has 16 cores and costs below $540?

People bring this processor up because its seemingly the only thing one can reference as to any relevancy intel has left.

I don't think so. I think people bring it up because it's perceived as the best performing CPU for the consumer platform.

It's all pretty simple: You don't wanna talk about it, you're not going to buy an Intel CPU moving forward = stop arguing then.

PS: I'm not buying Intel either and am also sitting here typing on an R7 1700.

 

[mattiasnyc]

History shows...
Current pricing

Do you really not see how the above makes no sense all ready???

8 core - 1700 w/ heatsink fan $170
8 core - 2700 w/ heatsink fan $270
16 core - 1950x $540
16 core - 2950x $700

And which ones of the above:

a) beat the 9900K

and

b) aren't being phased out

?

Now, try and help me figure out what Intel has to do with a clear window of pricing that AMD has already established in their own product line that they have to "compete" against.

You're just arguing for the sake of arguing. I'll tell you what:

- Let's bookmark this and we can go back to it once AMD releases a brand spanking new Ryzen CPU that performs better than the 9900K and costs half of what the 9900K costs.

Sounds good?

Anytime I hear the 9900k brought up as some type of benchmark, I know I'm talking to someone who has no serious compute needs and is just blowing hot air. Single core performance is even a more laughable concept... it's a hold over from the bulldozer fiasco.

Who cares what you think about that. Someone brought the 9900K up and said a new AMD chip would be better and cost 50% of the 9900K. I never said anything about whether or not CPU X was good or bad or worth the money. What I said was that that's not likely to happen.

We can wait and see if I'm right about that or not.

 

[ub4ty]

So what? I don't think this is about you. It's about the market.

I defined the market : What issue do you take with it?
Ryzen 1st gen
Ryzen 2nd gen

Which AMD CPU has 16 cores and costs below $540?

1950x. So, again, why would I buy an 8 core processor from AMD for the same price when I can get their 16 core for $540? This is what I used to highlight why your reasoning and focus on Intel was wrong. AMD is driving and setting the prices. They have themselves to compete against and prior product line. I'm not buying an 8 core from AMD at the price of a 16 core from AMD.

I don't think so. I think people bring it up because it's perceived as the best performing CPU for the consumer platform.

And I just defined that it's a niche game centric processor. The best performing CPUs in the consumer market are found on the HEDT platforms. Gaming processors are for gaming and only a niche number of users have a platform that can exploit such minor differences... They largely go on and on about performance metrics that in all honesty have little application beyond 'muh fps'. If you need performance for a real compute task you increase cores. Core counts go up .. clocks go down. compute capability goes up. We could talk about the one hundred and one goofy pipeline combos a company makes to eek out performance or just focus on higher core counts.

It's all pretty simple: You don't wanna talk about it, you're not going to buy an Intel CPU moving forward = stop arguing then.

This is an AMD thread about Ryzen 3000 and AMD's platform.... AMD has produced on 14nm and 12nm... 7nm is on the horizon. Intel is on 14nm and is irrelevant.

PS: I'm not buying Intel either and am also sitting here typing on an R7 1700.

Says it all..

 

[ub4ty]

Do you really not see how the above makes no sense all ready???

Address it and highlight why it makes no sense.
A company has produced 2 generation of products at various price points.
They are the industry leader and have a new platform. Their platform pricing is more aggressive than their competitor. They are competing against themslves. If that makes no sense to you, I suggest you re-asses how businesses operate. Companies ignore conflicts of pricing within their own product lines? wew ... news to me.

And which ones of the above:
a) beat the 9900K

1950x by 50-70% for less.

and
b) aren't being phased out
?

The 9900k that is 2 generations behind on 14nm and on a discontinued socket.
I can drop a Ryzen 3 in all of my existing motherboards.

You're just arguing for the sake of arguing. I'll tell you what:

I addressed and refuted all of your assertions and answered all of your obvious questions. You have yet to even attempt the same. It's clear who your comment applies to.

- Let's bookmark this and we can go back to it once AMD releases a brand spanking new Ryzen CPU that performs better than the 9900K and costs half of what the 9900K costs.
Sounds good?

Ummm, that's my assertion. You've been arguing with yourself.

Who cares what you think about that. Someone brought the 9900K up and said a new AMD chip would be better and cost 50% of the 9900K.

No. Someone brought up the 9900k and argued that AMD should have pricing inline with it if beats the processor in performance. I pointed out the absurdity of such a proposition given that AMD already makes a 16 core with 50-70% more performance than the 9900k that is priced at $540.

I never said anything about whether or not CPU X was good or bad or worth the money. What I said was that that's not likely to happen.
We can wait and see if I'm right about that or not.

I only care about official releases and specs.
Meanwhile, I have pointed out clear absurdities.
Predictions would only matter if I had advanced plans/decisions to make based on them. Beyond that, being right has no value. If the new line is above a certain price point, I'll buy the old line at reduced cost.

 

[mattiasnyc]

I only care about official releases and specs.

Clearly you care enough about other things to keep arguing.

If the new line is above a certain price point, I'll buy the old line at reduced cost.

This still isn't about you. I know you might be disappointed.


------------------------------


But whatever. You win:

AMD will release a CPU that will perform better than the 9900K, and it will cost half as much.

You convinced me.

 

[ub4ty]

Clearly you care enough about other things to keep arguing.

I discuss, critique, and logically progress through assertions. It keeps me sharp. Now that it's clear what can be concluded, our exchange si indeed over.

This still isn't about you. I know you might be disappointed.

Indeed which is why its good to focus on a particular subject/assertion and supporting arguments instead of making things personal. It's the only way you can have a constructive and intelligent exchange. I'd be disappointed if it was about anything else.

But whatever. You win:

AMD will release a CPU that will perform better than the 9900K, and it will cost half as much.

You convinced me.

It was never about you or me.
Logically as a consumer, if AMD doesn't price their Ryzen 3 processors correctly, it will make perfect sense to buy their 1st or 2nd generation for better value. Intel's value is horrible enough not to be considered. So, the only upcoming decision I have to make is whether or not to buy a Ryzen 1,2, or 3rd gen. AMD and any person with business sense understands this. So, i'd be pleasantly surprised if they do something as silly as price a 3rd generation 8 core at the same price as a 1st gen 16 core. Ryzen 2 provided no value or reason to upgrade. So, it's on them to capture me now and make a sale... tick/tock/tick. Or, I'll be buying a 16 core 1st gen for $400 bucks or less, consolidating, and selling off my Ryzen 1700s. Down the road, I'll consolidate to a 64 core Rome once the prices come back to reality. Really nothing more to consider as a consumer... I look for the best valued product in line w/ my needs (not wants or desires).


Other than that, there's pure excitement simply for the technological advancement Ryzen 3rd gen presents w/ 7nm/pcie 4.0/ and the I/O Chip. We can all agree on that. As is always, as an intelligent consumer, it comes down to price/performance after that.

Oh and Lastly.... If Intel wasn't enough of a case study of what happens when you become arrogant and greedy, check out Apple's free-fall that is occurring because they thought they could continue charging absurd amounts of money for tech that is nowhere in line with their production cost...
https://www.marketwatch.com/story/a...ars-with-massive-revenue-shortfall-2019-01-02
$158 share price to $145 (-$56Bn in market cap) back to 2017... for arrogantly thinking they can charge people almost a grand for a goofy same ol' same ol' iphone that cost about $300 to make.

Got greedy and lazy... and built a shiny new head quarters which brings about the famed valley curse. Then try to milk their consumers dry.. Nvidia did the same recently w/ RTX.... and a shinny new headquarters...
Share price Oct 2018 : $289
Current share price : $131

Essentially a train wreck back to reality... For the same classic reason I outlined earlier.
You live.. you die and make way for new life innovation/progress. Cycle of life.

AMD w/ Momma Su's leadership doesn't strike me as the kind of company that is going to make this mistake fresh off the heals of just recovering.