Speculation: Ryzen 3000 series

[DrMrLordX]

. We as consumers will profit as prices will innevitably go down so anyone can take a pick and chose the best chip for their needs (be it intel or AMD).

Prices will only go down once Intel is able to compete again.

@Asterox

That's really impressive! I can't wait to see what overclocked 3600s can do.

 

[inf64]

Looking good. T minus 11 days and counting.

I wonder what is going on with write BW in AIDA test on Zen2. Looks like 1/2 of what it should be.

 

[IEC]

5.275GHz on LN2 and > 1.7V

I think 5GHz on air all cores is probably... optimistic, to say the least.

 

[fleshconsumed]

Stock R5 3600 X265 Benchmark vs i7 8700K.

https://www.reddit.com/r/Amd/comments/c5n8r9

That is awesome. I'm doing a lot of x265 encoding and it's good to see Zen2 outperforming intel. I am so glad I switched to AM4 platform as soon as Zen came out.

 

[Fanatical Meat]

Prices will only go down once Intel is able to compete again.

@Asterox

That's really impressive! I can't wait to see what overclocked 3600s can do.

I’m still thinking intel has a core2duo up its sleeve again.
I’m guessing next year around this time we’ll hear about some little known engineering team intel employs in god knows where that has a “new” processor that performs at twice the speed or power or efficiency and miraculously it will appear on the market shortly after.
No facts just a feeling.

 

[PotatoWithEarsOnSide]

Stock R5 3600 X265 Benchmark vs i7 8700K.

https://www.reddit.com/r/Amd/comments/c5n8r9

What is the impact of Realtime versus Normal and HPET Active versus HPET Not Active?

 

[BigDaveX]

So AMD will finally match or exceed contemporary iteration of Core processors in IPC since the Core 2 Duo days? It has been over 12 years. Holy dear god.

The funny thing is, if Intel's 10nm process hadn't been such an utter, catastrophic failure, Zen 2 would most have likely fit into the historical pattern since Phenom I (disregarding the Bulldozer era) of AMD's latest and greatest being a generation behind its Intel counterpart in single-thread performance, assuming the early reports about Ice Lake's performance have any substance to them. Possibly it would have just ended up being a repeat of the original Zen's launch, with Intel ahead in gaming and scenarios that required a low-to-mid thread count, and AMD ahead on content creation.

 

[Hans Gruber]

I’m still thinking intel has a core2duo up its sleeve again.
I’m guessing next year around this time we’ll hear about some little known engineering team intel employs in god knows where that has a “new” processor that performs at twice the speed or power or efficiency and miraculously it will appear on the market shortly after.
No facts just a feeling.

I think you are right. When that happens Zen 3 will be out. You guys realize they dropped hyper threading not because intel processors didn't need it. All the security issues went away when hyper threading was removed. Not exactly a good outcome for intel.

 

[DrMrLordX]

5.275GHz on LN2 and > 1.7V

1.7v? That doesn't seem right somehow. I don't think even 14LPP took that much voltage to hit 5.2 GHz under LN2.

I think 5GHz on air all cores is probably... optimistic, to say the least.

Air can barely handle the heat output from my 1800x. It won't cut it on a 3950x, certainly not at those clockspeeds.

 

[inf64]

I think you are right. When that happens Zen 3 will be out. You guys realize they dropped hyper threading not because intel processors didn't need it. All the security issues went away when hyper threading was removed. Not exactly a good outcome for intel.

AMD plans to intersect Sunny Cove core with Zen3. It should negate any small IPC advantage Icelake parts might have over Zen2.

 

[lixlax]

I’m still thinking intel has a core2duo up its sleeve again.
I’m guessing next year around this time we’ll hear about some little known engineering team intel employs in god knows where that has a “new” processor that performs at twice the speed or power or efficiency and miraculously it will appear on the market shortly after.
No facts just a feeling.

Conroe was a further developed core of the mobile ones (Banias, Dothan) from the A64 vs P4 era so it could've been predicted to come in a desktop form at some point. At the moment we know about Icelake and this is by no means a second coming of Conroe.

The netburst architecture was a terrible one to begin with (mainly regarding IPC) which helped them to get that ~2x IPC increase, I think it was "only" about 20% higher than K8. The same goes for Zen since the architecture it replaced had subpar IPC.
They'd really need something groundbreaking to pull something like this off again. Guess we'll find out in ~1-2 years.

 

[DrMrLordX]

They'd really need something groundbreaking to pull something like this off again. Guess we'll find out in ~1-2 years.

There's always Atom. Keep your eyes on those Tremont cores.

 

[IEC]

1.7v? That doesn't seem right somehow. I don't think even 14LPP took that much voltage to hit 5.2 GHz under LN2.

Those electrons just wouldn't move themselves, they had to be pushed while screaming the whole way.

Source:
https://pbs.twimg.com/media/D9-K1ZQXUAApTYn.jpg

 

[Fanatical Meat]

Conroe was a further developed core of the mobile ones (Banias, Dothan) from the A64 vs P4 era so it could've been predicted to come in a desktop form at some point. At the moment we know about Icelake and this is by no means a second coming of Conroe.

The netburst architecture was a terrible one to begin with (mainly regarding IPC) which helped them to get that ~2x IPC increase, I think it was "only" about 20% higher than K8. The same goes for Zen since the architecture it replaced had subpar IPC.
They'd really need something groundbreaking to pull something like this off again. Guess we'll find out in ~1-2 years.

My point is intel hasn’t had a fire under their ass to do anything for a long time. Ryzen 2000s were good but not good enough to be a great competitor.
Intel has been doing what all big companies do, pocket the money, I’m sure when they start throwing billions of dollars at a problem something that was impossible to do last year suddenly becomes possible.

 

[coercitiv]

My point is intel hasn’t had a fire under their ass to do anything for a long time.

Intel has had their ass ignited many timed in the recent past, or have you forgotten about their mobile endeavor fiasco?

I’m sure when they start throwing billions of dollars at a problem something that was impossible to do last year suddenly becomes possible.

They threw billions of dollars at mobile too, and those little ARM cores chewed them up like candy. Contra revenue was sweet. You can make predictions based on feelings and faith all day long, the harsh reality is this industry works on the foundations of science, engineering and most importantly building and planning years ahead. Money works wonders if you already spent them years ago, not starting today for next year. Apple dumped billions into a Google Maps competitor and it took them 7 years to even get close to their competition, let alone Conroe them.

And let this sink in: Zen 2 isn't about AMD R&D budget only, it's surprising strength comes from TSMC R&D too, and TSMC has worked like crazy in the last decade.

This will not end with some magic unicorn move. It will be bloody, sweaty and tearful.

 

[DrMrLordX]

Those electrons just wouldn't move themselves, they had to be pushed while screaming the whole way.

Source:
https://pbs.twimg.com/media/D9-K1ZQXUAApTYn.jpg

Hmm, looks like 2700x LN2 overclocks took over 1.75v as well.

 

[Topweasel]

My point is intel hasn’t had a fire under their ass to do anything for a long time. Ryzen 2000s were good but not good enough to be a great competitor.
Intel has been doing what all big companies do, pocket the money, I’m sure when they start throwing billions of dollars at a problem something that was impossible to do last year suddenly becomes possible.

The problem is money doesn't solve a problem the reality of CPU design is time heals all wounds. Give the right people the right amount of time and they can come up with something great. Zen isn't just great because it brought AMD up to parity, but because it was whole philosophical change in approach to how to produce CPU's from the bottom all the way up to the top. Zen 2 being the first full realization of that design commitment. Intel needs to have some of those eureka moments and you can't throw money at that. Sunnycove for example is a mediocre boost (10%) in a core design they have been working on for 12-13 years. It will be a kudos for eaking that out but its another 3-4 years before they have anything different. They aren't just going to chuck that out right away.

Perfect example of throwing billions of dollars at a problem. 10nm, what did that result in, Intel throwing several Billion to expand 14nm.

This isn't about Sky is falling and the end of Intel. It is about the end of an era at Intel and that era is going to end with them losing their superiority in about all CPU markets they are in but maybe HPC. Who knows what it means market wise. Who knows when and if they will retake their normal leads. Who knows how big AMD might get in the meantime. But just throwing money at the problem won't solve it.

 

[cytg111]

Ill be looking for benches showing them toe to toe, before and after security patches.

 

[Zucker2k]

Intel needs to have some of those eureka moments and you can't throw money at that. Sunnycove for example is a mediocre boost (10%) in a core design they have been working on for 12-13 years. It will be a kudos for eaking that out but its another 3-4 years before they have anything different. They aren't just going to chuck that out right away.

Intel's problem is not architectural design. It's funny that you think 10% (is that an average figure?) upon Skylake architecture is "paltry" when AMD is now getting to Skylake ipc level after how many tries? AMD's ONLY advantage at this time is TSMC's 7nm which translates to more performance per watt so they can cram more cores into a chip at the same tdp level.

I think Intel will learn a lot from the challenges of both 14nm and 10nm and could tap into lessons learned from these processes going forward. On the architectural front, I see the beginning of the cache wars cropping up on us. Intel has no reason not to make the L4 cache a mainstay on their desktop chips, especially with the introduction of progressively beefier igps.

 

[OTG]

[QUOTE="Zucker2k, post: 39855216, member: 189543" AMD's ONLY advantage at this time is TSMC's 7nm[/QUOTE]

Well, that's just not true- being the first to have a functioning chiplet design has all sorts of advantages. It's probably the biggest factor allowing AMD to have Zen2 this year, Zen3 next year, and Zen4 the year after.
It means that they can design new CPU chiplets, push them into (affordable) production earlier than a large monolithic design, and lets them shuffle around CPU, GPU and IO dies as needed for new products.
Being ahead on process is icing on the cake; they'd be competitive without that advantage, but they are indeed fortunate that it puts them comfortably in the lead (for now).

 

[guachi]

Those electrons just wouldn't move themselves, they had to be pushed while screaming the whole way.

Source:
https://pbs.twimg.com/media/D9-K1ZQXUAApTYn.jpg

Screaming Electrons is my new band name.

 

[lobz]

Intel's problem is not architectural design. It's funny that you think 10% (is that an average figure?) upon Skylake architecture is "paltry" when AMD is now getting to Skylake ipc level after how many tries?

After 2 tries

 

[inf64]

Intel's problem is not architectural design. It's funny that you think 10% (is that an average figure?) upon Skylake architecture is "paltry" when AMD is now getting to Skylake ipc level after how many tries? AMD's ONLY advantage at this time is TSMC's 7nm which translates to more performance per watt so they can cram more cores into a chip at the same tdp level.

I think Intel will learn a lot from the challenges of both 14nm and 10nm and could tap into lessons learned from these processes going forward. On the architectural front, I see the beginning of the cache wars cropping up on us. Intel has no reason not to make the L4 cache a mainstay on their desktop chips, especially with the introduction of progressively beefier igps.

That is all true but don't forget that AMD of today is a moving target. They have Zen 3 ready to intersect Sunny Cove core and Zen 4 one year after that to extend the lead. They are aware of the opportunity that was presented and intel knows that as well. What needs to happen in next two years for intel to regain the competitive edge is :

1) AMD to slip tremendously with Zen3/Zen4 and/or TSMC to fudge the 7nm+/6nm/5nm
2) intel to execute flawlessly with Sunny Cove(+) cores and their own process nodes.

Too many factors need to align so that intel can regain the lead in next 2-3 years. I'm of opinion that this will not happen and intel will bleed heavily their market share the next 2 years.
They are a massive company that can rebound from the slump their are in right now, but we need to be realistic. Behemoths move slowly and take a lot of time to take a swing. When they do they inflict a lot of damage. AMD is now inflicting a death of thousand small cuts to intel trying to bleed them before they can recuperate.

 

[maddie]

Intel's problem is not architectural design. It's funny that you think 10% (is that an average figure?) upon Skylake architecture is "paltry" when AMD is now getting to Skylake ipc level after how many tries? AMD's ONLY advantage at this time is TSMC's 7nm which translates to more performance per watt so they can cram more cores into a chip at the same tdp level.

I think Intel will learn a lot from the challenges of both 14nm and 10nm and could tap into lessons learned from these processes going forward. On the architectural front, I see the beginning of the cache wars cropping up on us. Intel has no reason not to make the L4 cache a mainstay on their desktop chips, especially with the introduction of progressively beefier igps.

You do realize that they're pretty much in the same place IPC wise, just having taken different paths to get there?

 

[Abwx]

AMD is now getting to Skylake ipc level after how many tries? .

10% better perf/clock in ST in CB, and likely more in MT, is not exactly what can be called the same IPC level...