Speculation: Ryzen 4000 series/Zen 3

[jpiniero]

And the right way to do it. I still cant get how intel is run by the accountant (ceo).

Well, BK was a "fab guy" and look where it got them.

 

[Saylick]

Well, BK was a "fab guy" and look where it got them.

BK being a "fab" guy is like how Apple is run by a "supply chain" guy. Neither are true technical/design types; they're both bean counter types that are great at milking a successful business and padding the bottom line, not for driving innovation and pushing the envelope. The difference being Apple just has too much clout and a lack of true competitors for them to flounder with Tim at the helm.

 

[Richie Rich]

As a software architect driving semi large projects I dont have a single developer on my team which role I cant stand in myself. Having a tech management role requires you to be bleeding edge yourself.

Exactly. The point is to know all important stuff which creates 80% of impact from each segment. Those last 20% is often huge amount of less important details. That's Pareto rule. More segments you can absorb means bigger picture you can see. This is main difference between visionary type (Musk, Jobs, Keller etc.) and specialist type. This is reason why even great specialist (or CFO->CEO) promoted into leading position could be a disaster (like Bulldozer or Sculley in Apple).

Anyway, every person who went from engineer position up to vice-president with significant industry impact should be valuated as positive example worth to follow. If not there is always something wrong with those people.

 

[soresu]

Musk, Jobs, Keller

Keller does NOT belong in the same breath as Musk and Jobs.

Jobs was mostly a business 'genius' who took the lions share of the credit for ideas realised by others like Wozniak - like a frontman for a band who takes the public credit for songs he sings, but did not write.

Musk certainly has a greater overall technical grounding than Jobs did, but don't be fooled - he's hardly some contemporary Einstein among engineers, he's still just doing what Jobs did.

Keller on the other hand is an actual expert/specialist, more Wozniak than Jobs - he just tends to get more credit than those around him due to being a more visible management fixture in several notable CPU development groups since DEC Alpha.

Even at the university level, often the true author of a big discovery can be overshadowed by the name of their mentor and head of any research effort.

I can't count the number of times I read the name of a specific professor called James Tour in relation to materials research work (graphene, ReRAM) at Rice University, and even though he often singled out the true author of the paper under his mentorship, those names slip my mind because his name keeps coming up afterward over something else - I suspect a similar fog lies over Keller's many collaborators.

 

[soresu]

The difference being Apple just has too much clout and a lack of true competitors for them to flounder with Tim at the helm.

That and Samsung stopped really pushing the envelope too some time ago - I don't think folding phones really count to be honest.

I'm hoping that their licensing RDNA from AMD could indicate a drive for standalone VR to replace Gear VR in the future, at least then they would be back in a field they had pioneered in the early days of the modern VR push, as it is Apple is clearly dipping their toes in with software before a hardware push in the future.

 

[kapulek]

Any idea what is this? Is it for apple or someone else?

https://www.reddit.com/r/Amd/comments/euq7b0

 

[uzzi38]

Any idea what is this? Is it for apple or someone else?

https://www.reddit.com/r/Amd/comments/euq7b0

There's one thing _rogame forgot to mention there - the memory is most likely GDDR6. GT2 in Time Spy is predominantly memory bound, so if it's beating the GDDR5 based Subor Z+ by such a large margin.. well I'm sure you can guess.

Aside form that, everything we know is there.

Honestly, I think he's on point with his guess that it's the Xbox Series S - or in other words, the rumoured 'Lockhart'.

Also, it's probably not for Apple. This benchmark suggests that only GDDR6 memory is used, there's no DDR4 etc at all. Shared memory between the CPU and GPU only - like the Subor Z+. I don't think Apple would request for such a device, a large number of workloads in Apple's ecosystem are quite dependent on memory latency, the iGPU isn't good enough for this to be a device based around gaming when compared to upcoming consoles.

 

[kapulek]

If Lockhart is 8c/8t and Xbox Series X is 8c/16t developers must be "happy". Apart from PS5, XsX and PC they will have one more console to develop for.

uzzi38 have you heard anything about next gen Nintendo console with AMD hardware?

 

[Veradun]

If Lockhart is 8c/8t and Xbox Series X is 8c/16t developers must be "happy". Apart from PS5, XsX and PC they will have one more console to develop for.

uzzi38 have you heard anything about next gen Nintendo console with AMD hardware?

I am an early band wagoner for 8c/8t in XsX

Yeah, should or shouldn't my point is an 8c/8t setup is perfectly in line with what Devs have been using the past generation and doesn't bear any additional baggage for additional optimization steps.

I second this and further add that I think the CPU part will be 8c/8t (no SMT) to have better efficiency and also a strict parallelism with jaguar for perfect portability of whatever code one wants to port from the one x

 

[uzzi38]

uzzi38 have you heard anything about next gen Nintendo console with AMD hardware?

Nope, and I doubt there will be one.

 

[NostaSeronx]

I have a small expectation of SmachZ team to announce another delay and a move to Renoir+LPDDR4X.

 

[DisEnchantment]

I think going forward we can be sure those leaks will be drying up. Zen1 and Zen2 benchmarks were leaked months ahead of launch.
AMD will be more reluctant to share details going forward to avoid Osborning themselves.
Its 2 Q away and there is nothing on the internet for RDNA2/Zen3 in contrast to Zen benchmarks leaks of almost 9+ months before launch and 7+ months for Zen2

 

[amrnuke]

I think going forward we can be sure those leaks will be drying up. Zen1 and Zen2 benchmarks were leaked months ahead of launch.
AMD will be more reluctant to share details going forward to avoid Osborning themselves.
Its 2 Q away and there is nothing on the internet for RDNA2/Zen3 in contrast to Zen benchmarks leaks of almost 9+ months before launch and 7+ months for Zen2

It always seemed silly that they would benchmark their unreleased CPUs (or permit benchmarking their unreleased CPUs) on programs that they know communicate with a server. I'm glad it's locked down. I actually think it may tamp down on the rampant speculation, since there won't be "Oh, that's an ES and probably we can expect xx% higher clocks from the final product" and so on.

 

[joesiv]

Nope, and I doubt there will be one.

I also doubt it too. Nvidia was very smart with their agreement with Nintendo for the switch. Nvidia is doing the developement of the Switch SDK for developers. This would abviously make it much harder for Nintendo to switch vendors and have to create all the toolsets again from scratch, especially if they want backwards compatibility with the next gen switch (switch pro?) and the current software.

I don't know if Nintendo is stilll planing on being more mobile like (IOS/Android) in terms of decoupling hardware/software release cycles, not locking software to a specific hardware generation. If so, it would seem that Nintendo would stick with Nvidia for several generations, and most definitely arm for the CPU.

 

[coercitiv]

I actually think it may tamp down on the rampant speculation, since there won't be "Oh, that's an ES and probably we can expect xx% higher clocks from the final product" and so on.

The worst rampant speculation on Zen 2 had nothing to do with the leaks. Remember the 6c/12t @ $99 or 16c/32t @ 5.1Ghz claims?

 

[scannall]

The worst rampant speculation on Zen 2 had nothing to do with the leaks. Remember the 6c/12t @ $99 or 16c/32t @ 5.1Ghz claims?

None of those ever made any sense at all.

 

[coercitiv]

None of those ever made any sense at all.

All I'm saying is people invent things out of thin air, no ES bench leaks necessary.

PS: on the flip side I welcome the relative silence too, gives us heaps of time to endlessly debate SMT4 and 6xALU for ALL. /s

 

[guachi]

None of those ever made any sense at all.

The worst rampant speculation on Zen 2 had nothing to do with the leaks. Remember the 6c/12t @ $99 or 16c/32t @ 5.1Ghz claims?

Kinda hit one with the 6c/12t with the 1600AF at $85. Not exactly the same, but still crazy!

 

[scannall]

All I'm saying is people invent things out of thin air, no ES bench leaks necessary.

PS: on the flip side I welcome the relative silence too, gives us heaps of time to endlessly debate SMT4 and 6xALU for ALL. /s

At some point they will have to go wider. Silicon won't clock any higher. And alternative materials, while possible are a lot more expensive. And just slapping 2 more ALU's on by themselves wouldn't add much throughput. Need to keep all that fed as well, which is even more transistors. When will it happen? I have no idea, my crystal ball is defective.

 

[Ajay]

All I'm saying is people invent things out of thin air, no ES bench leaks necessary.

PS: on the flip side I welcome the relative silence too, gives us heaps of time to endlessly debate SMT4 and 6xALU for ALL. /s

I need a filter that blocks posts with with the acronyms SMT4 or 6xALU in them. Not sure if I’m being sarcastic.

 

[DrMrLordX]

I need a filter that blocks posts with with the acronyms SMT4 or 6xALU in them. Not sure if I’m being sarcastic.

There's an ignore button for that.

 

[coercitiv]

Kinda hit one with the 6c/12t with the 1600AF at $85. Not exactly the same, but still crazy!

The 1600 AF does indeed call for a moment of reflection on the past few years. The jump in performance and value in the $100 and $200 price brackets was quite astounding.

[Edit] I wanted to reference the 4C decade in some way, but couldn't find the right words until I stumbled upon this old article: The Intel Core i3-7350K (60W) Review: Almost a Core i7-2600K. This article title is the perfect embodiment of the past era and exactly what Zen helped resetin terms of consumer choice.

 

[NTMBK]

I am an early band wagoner for 8c/8t in XsX

It does make sense for a game console. SMT increases variability due to resource sharing, making it harder to guarantee you will hit a locked 60fps.

 

[Thunder 57]

It does make sense for a game console. SMT increases variability due to resource sharing, making it harder to guarantee you will hit a locked 60fps.

I agree. I would not be surprised at all if the next gen consoles were both 8C/8T.

 

[darkswordsman17]

I agree. I would not be surprised at all if the next gen consoles were both 8C/8T.

I think both Sony and Microsoft have already stated 8C/16T. Perhaps that's still just rumor, but I believe they've outright said as much (Microsoft with regards to the Series X, and Sony the next Playstation). There was also a report that Microsoft had Xbox dev units that could do 3 threads per core via some special code path. Sounds like there's a push to make games much more multi-threaded.