• Hey everyone, staff have documented a list of banned content and subject matter that we feel are not consistent with site values, and don't make sense to host discussion of on Famiboards. This list (and the relevant reasoning per item) is viewable here.
  • Do you have audio editing experience and want to help out with the Famiboards Discussion Club Podcast? If so, we're looking for help and would love to have you on the team! Just let us know in the Podcast Thread if you are interested!

StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (Read the staff posts before commenting!)

Reader mode
Serif said:
I think a lot of us assumed Mochizuki's 2021 reporting was him conflating Drake with Switch OLED, and that his 11 devs report was about Drake. As far as I'm concerned he was 'vindicated' by the NVN2 leaks. I personally draw no connection between a 4K enabled Switch and whatever John Linneman was talking about, which I still think is a Mariko Pro.
Click to expand...
I agree with this, keep in mind that Nintendo's R&D department comes up with many crazy types of hardware, many of which never see the light of day. I do believe that Drake production is still ongoing, but I would not be surprised if features like 4K and DLSS would have been attempted to be implemented into cancelled, revamped, or altered hardware eariler in the Switch's life.
 
Just a quick point. Rapidus is not explicitly tied to IBM in the sense some people seem to think. Rapidus exists to be another source of state of the art chips in the world this one explicitly from Japan which is way behind in the Chip race. IBM is partnering with them to make 2nm chips. Sony, Toyota and others have all invested in Rapidus and I could easily see Rapidus and its allies approaching Nintendo to invest in Rapidus for the future.

Rapidus won't be producing anything until 2027 so anything happening today is long term plans.
 
oldpuck said:
I was only commenting on the ability to hit that specific power draw with that specific clock.

Drake should be able to comfortably hit the original Switch’s clocks (or slightly higher) at the original Switch’s power draws on N4, based on everything we know.

Those clocks would be consistent with PS4+ levels of performance.
Click to expand...

I would be just fine with that. My only interest in observing the Steam Deck and what it can do was to get a glimpse into what could be expected from Switch 2. The caveat for me was to understand that Valve is allowing there device to pull 25+ watts under any conditions, where Switch 2 will be limited to 7-10 watts portable and 12-15 watts docked. When you turn down the TDP on Deck to get 10 watts of total system draw things start to buckle a bit. It certainly isnt the PS4 in your hands that it is when you let it run full tilt. The Decks APU is very modern and on a fairly efficient TSMC 7nm process. Just like Switch, there are efficiencies on consoles compared to PC, but I think its relevant to observe what Deck can do when limited in power and understand that its simply very tough to deliver PS4 levels of performance on 7 watts. Deck has also shown that high end games still look good at 720p when on a 7" screen. Nintendo can carry over their 7" OLED screen to Switch 2. This not only keeps power consumption down compared to using a higher res screen, but allows them to underclock the memory significantly in portable mode since bandwidth shouldn't be an issue at 720p.

So for giggles if we apply the Switch Clock speeds to Drake, 304-460Mhz portable 1.1-1.4 Tflop and docked 768Mhz 2.3 Tflop. This is well under some of the high end estimates some of us have hoped for, but also positions it comfortably beyond last gen consoles, especially with the inclusion of DLSS.
 
Do you guys think we can see More AAA Third Party Games come to Next gen nintendo console? Such as Grand Theft Auto 6, Red dead, Payday 3, Stalker 2, etc.
Of cource is scaled back verison probaly.
 
At this point, I think the major bottleneck for third-parties will be the online infrastructure, rather than the hardware itself, at least for multiplayer-focused games. Engines are so scalable nowadays, and with tech like DLSS, you can get away with some blocky-looking image quality that you won't really bat an eye unless someone points it out.

Speaking of - I'm guessing not, but - do we have any advancements from the available data (or just general expectations) for improvements on wireless tech for Drake? I'm honestly illiterate in this topic, so I have no clue if this even is something that would show up on a leaked API source code or the likes 🥲

I remember back in the day, when we learned that the Switch would support 5Ghz connections, that was a source of relief for some, even though, in the end, the antenna placement became a much bigger issue...
 
NateDrake said:
Clarity will come.

In the meantime: just operate under the condition of a plan to bring something out in 2023 may or may not exist.

Don't think of it as a cancellation of an SoC. At best, think of it as a delay to bring the SoC to market via this specific product.

As said; many questions here will be answered.
Click to expand...

Maybe it's me not being a native speaking (but my English is good enough on a business level...). But I have no idea what this is supposed to mean. Like I can literally not comprehend the sentences. Why is a delay "at best"? What does the first sentence even mean? Just that we should not assume anything?
And what is he alluding to? Of course all questions will be answered at some point, as Nintendo will do more hardware in the future.

I honestly don't get this post, I would like to but it's just super confusing to me, even more then the already confusing initial statements. If there's nothing clear to say, maybe not say anything until it changes? Or I just don't get it.
 
I'm simply saying to wait until I can present everything we have and answers to the question of 2023 hardware will come at that time.

That's all the "maybe or maybe not" mention was for. Covering both questions and saying I'll answer that question and more soon.

One can also use the information we present to form your own conclusions.
 
thegodsend said:
Maybe it's me not being a native speaking (but my English is good enough on a business level...). But I have no idea what this is supposed to mean. Like I can literally not comprehend the sentences. Why is a delay "at best"? What does the first sentence even mean? Just that we should not assume anything?
And what is he alluding to? Of course all questions will be answered at some point, as Nintendo will do more hardware in the future.

I honestly don't get this post, I would like to but it's just super confusing to me, even more then the already confusing initial statements. If there's nothing clear to say, maybe not say anything until it changes? Or I just don't get it.
Click to expand...

Key takeaways:
  • Nothing conclusive is being said about a device hitting this year. It could still happen.
  • That said, something was cancelled.
  • The cancelled thing wasn't the SoC itself. T239 could see a release in another device.
 
NateDrake said:
I'm simply saying to wait until I can present everything we have and answers to the question of 2023 hardware will come at that time.

That's all the "maybe or maybe not" mention was for. Covering both questions and saying I'll answer that question and more soon.

One can also use the information we present to form your own conclusions.
Click to expand...
Take your time Nate! I will gladly await your podcast!
 
Random tangents (not even remotely related to tech this time):
1. It just hit me that Famitsu didn't take a Xmas week off like the manga (Japanese comics) magazines do
2. The January 5 issue (which should be on store shelves in Japan at this moment then...) does say 1/19 on its cover; so Famitsu's taking a break next week? Guess that's their new year's break then?
3. Famitsu, you're a bit over double the price of an issue of Weekly Shounen Jump!? Hope that your paper quality explains that difference :p
4. Interesting... alright, so, these should be physical magazines shipped across Japan in order to hit the shelves in time. Read as: there is an exploitable supply chain here. For those familiar with the manga world, you may know about scanlations of chapters being released up to almost a week earlier than the official Japanese publication. Somewhere in that supply chain between printing and placing on shelf, someone can just slip away a copy of the magazine and scan it. Or, some stores break street date by like a day or two and someone just buys an issue early.
The part that has me going "Really now? :unsure:" is that this practice doesn't seem to be anywhere as prevalent with Famitsu? An effect of the bifurcation of gaming culture between Japan and the Anglosphere? In this particular instance: this issue hasn't already been scanned and the full article itself has yet to make its way over to the Anglosphere?
 
Foltzie said:
In the first scenario would you have a guess as to why the label is a Wattage if it's not the second scenario?
Click to expand...
The GPU power draw? The power draw of a totally unrelated device?

Foltzie said:
For example could it be the an expected total power draw of Drake, which would give you the reverse information from the second scenario, but would imply lower clocks since the CPU and other items haven't been accounted for yet.
Click to expand...
Sure. Just no way of knowing. Let me get super in the weeds here, and lay out what happened, and why it's hard to know what these numbers mean. Feel free to ignore, it's too damn long.

  1. In 2015, Nvidia starts development of NVN.
  2. Nvidia uses a process called "CI/CD" or "Continuous Integration/Continuous Deployment"
  3. CI/CD means that every time a developer makes a code change to NVN, an automatic process compiles the software, and then runs a test suite to make sure the change works.
  4. The software that does CI/CD is Jenkins. Jenkins doesn't run on Nintendo's custom OS, which isn't even ready yet.
  5. Jenkins also needs a development toolchain to actually build the software
  6. And it also needs hardware which can run in a server farm somewhere, not hardware built to run on a desktop
  7. So somewhere, in a datacenter, there are a group of machines that look nothing like Nintendo hardware that run NVN tests all day
  8. These are probably AMD based Windows Servers with 8 Nvidia GPUs in a cluster, possibly not even desktop GPUs, but server specific GPUs
  9. Every day, Jenkins builds and tests NVN over and over again
  10. Over the years of development the test suite gets huge
  11. There is a test for every bug ever found in the course of development
  12. There is a functional test created by developers during development to verify functionality
  13. There are integration tests to make sure NVN and all it's parts work together
  14. And there are performance tests which show over time if NVN performance is getting better or worse
  15. Some of these performance tests are KPIs - Key Performance Indicators.
  16. Somewhere at Nvidia there is a giant dashboard that shows not only the current status of all of these things but the history - the performance of KPIs over time. Meanwhile...
  17. DLSS 2.0 development begins
  18. Just like NVN, it uses Jenkins and has a set of tests and KPIs
  19. At some point a developer builds dlssDonutTest, as a tool to test performance of DLSS.
  20. DLSS 2 doesn't work on it's own. it needs to be integrated into various Graphics APIs
  21. And they all need to be tested
  22. dlssDonutTest gets updated to run on DirectX11, DirectX12, and Vulkan, all with a standard set of KPIs
  23. This not only allows each integration to be tested, it allows them to be compared to each other
  24. If all three get slow, then DLSS is the problem
  25. If just one gets slow, it's the integration
  26. But for all this comparison to work the hardware and the configuration of it all needs to be consistent for every single test
  27. 2019, NVN2 Development begins
  28. NVN2 = NVN1 + DLSS + RT + Drake Specific Optimizations
  29. Step one for NVN2? Take a copy of NVN1, and copy the DLSS development code into it, begin integrating
  30. This is kinda similar to the integration of DirectX of Vulkan, in fact
  31. NVN2 inherits all of NVN1's tests, benchmarks, and KPIs. They may change over time, but that's the starting point, it's a huge resource
  32. It also inherits the DLSS test suite and benchmarks, including the dlssDonutTest.
  33. Just like NVN before it, it cannot run on Drake, as Drake doesn't even exist. It needs to run in the server farm, probably on windows.
  34. It also needs to inherit NVN's testing hardware
  35. Partially just because it already exists, but unless you're running benchmarks under the same setup you can't know if NVN2 is getting slower or faster than the original NVN1.
  36. You also want to keep the tests named the same so that you can see trends over time, even if those names don't reflect the current state of affairs. This is one of the most annoying parts of my job as a perf engineer, honestly.
  37. The CI/CD process takes some time. Developers want to quickly check if their changes will work before checking their code in and firing off Jenkins
  38. So a developer makes a script that runs some KPI benchmarks, but on his local machine simulating the testing machine which only has a vague relationship to the final hardware.
  39. This is the script being discussed
I realize this was a lot but it's important context. Everything in that list we basically know happened because it's either in the leak, or is absolute bog standard practice for the industry. Some version of this absolutely happened internally at Nvidia. But it leaves us with a bunch of questions

  • Where did this KPI come from? New for NVN2? Inherited from NVN1? Inherited from DLSS2?
    • This test was probably inherited from DLSS2. Earlier versions of this test script checked Windows Vulkan performance as well as NVN2. It seems highly likely that the NVN integration took the dlssDonutTest, and ran it using the same parameters as the other integrations. Early on, the Vulkan numbers were compared side by side with the NVN2 numbers to verify that the performance matched and NVN2 didn't have any issues
  • Do the clock speeds mean anything? Are they Drake clock speeds?
    • Almost definitely they are not. This script sets GPU parameters on a developer's machine in order to make their workstation (probably running a consumer GPU) act like the testing hardware in the server farm (running a server GPU) while Drake was in very early days of development, and was inherited from the early NVN1 development environment. The odds that the clocks on the workstation match the server clocks match the in development hardware, match the final numbers that Nintendo selects balancing battery life, performance, and developer requests is vanishingly small.
    • It is entirely possible - even likely - that these clocks were set as a testing baseline for DLSS and have nothing to do with NVN2 development at all.
  • What do the wattage numbers mean? Surely these are for Drake?
    • Maybe? But what part of Drake? The whole SOC? The GPU? Or even the whole console, including RAM, screen and Joy-Con draw?
    • Or they could have been inherited from Original Switch NVN development. After all, NVN probably ran it's tests under multiple performance profiles. It is totally reasonable to assume that a low/middle/high power configuration were picked, based on TX1's power draw... and the labels just stuck. As I said before, renaming/altering your performance tests is a general no-no, because it breaks your historical trends and comparisons. Labels and configs tend to stick around long after their original use case is gone.
    • Or they could refer to one of the various pieces of testing/development hardware. After all, Drake doesn't exist at the time these scripts are being used. These power numbers could refer to an Xavier testing board, or a virtual GPU partition on an A100 in the server farm.
    • Or these labels come from DLSS development. This is very unlikely, but it's not impossible. Nvidia was touting power savings for DLSS early on, these labels might be to match DLSS's test environment, with wattages refer to tensor core draw.
    • Or a combination of these things - Like the power draw referring to the testing profiles for early NVN1 development, which was done on TK1s instead of TX1s
  • Okay, what does Occam's razor say? Surely the simplest explanation is these are Drake clocks and Drake power draws?
    • No. That's only the simplest explanation if you ignore all the data above. Occam's razor says that the simplest thing that explains all the data is probably the true one. It doesn't mean the explanation that ignores some of the data ;)
    • The simplest explanation is these clocks have nothing to do with Drake. The odds that the clocks a developer using a desktop Turing based RTX 2060 is using to simulate tests running on an Ampere based server A100 GPU will match the final clocks of an in development die shrunk Drake, all of which have wildly different numbers of SMs, is vanishingly small. And all of that assumes that this test is, in fact, trying to simulate Drake at all, rather than just being a modified DLSS test.
    • These wattages probably do have to do with Drake in some capacity. The wattages only appear in NVN2 tests that we can see, and NVN2 is clearly optimized for Drake. It's not a slam dunk, but it is the most likely guess.
    • But we can't predict anything from these numbers without further context. Since the clock numbers likely have nothing to do with Drake's clocks, we can't guess at the final clocks. Since we have no idea whether or not this is just the GPU or the whole SOC, we can't guess at battery life.
 
thegodsend said:
I honestly don't get this post, I would like to but it's just super confusing to me, even more then the already confusing initial statements. If there's nothing clear to say, maybe not say anything until it changes? Or I just don't get it.
Click to expand...
I mean, on the flip side someone said something pretty clear and they seem to have been wrong because "plans change," which is understandable haha. It's weird to see how the differing reactions to both which only tells me that some want to be told something with confidence and not fact.

But I do agree with "maybe not say anything." I think it would be better if nobody said a thing and we just waited patiently(lol) for Nintendo to announce whatever it is.
 
0
NateDrake said:
I'm simply saying to wait until I can present everything we have and answers to the question of 2023 hardware will come at that time.

That's all the "maybe or maybe not" mention was for. Covering both questions and saying I'll answer that question and more soon.

One can also use the information we present to form your own conclusions.
Click to expand...
Thank you! I personally respect you waiting until you have all your ducks in a row to give your definitive perspective. It reaffirms how thorough and careful you are, which is very important in a source of information like this.

I know these are all leaks and rumors and you don't owe us anything, but as a victim of overwhelming hype, I really appreciate it.
 
NateDrake said:
I'm simply saying to wait until I can present everything we have and answers to the question of 2023 hardware will come at that time.

That's all the "maybe or maybe not" mention was for. Covering both questions and saying I'll answer that question and more soon.

One can also use the information we present to form your own conclusions.
Click to expand...


Nate we both know damn well click bait will come from these statements. The irony is you had a podcast about this issue.
 
NateDrake said:
I'm simply saying to wait until I can present everything we have and answers to the question of 2023 hardware will come at that time.

That's all the "maybe or maybe not" mention was for. Covering both questions and saying I'll answer that question and more soon.

One can also use the information we present to form your own conclusions.
Click to expand...
Considering we can "form our own conclusions" the answer to the question "will new hardware come out in 2023?" is "maybe"?
 
kimbo99 said:
Nate we both know damn well click bait will come from these statements. The irony is you had a podcast about this issue.
Click to expand...
What is there to take from the post to make into clickbait? I'm not offering up any details yet. I'm simply saying I'm going to have information that will answer many of the questions this thread has had.
 
What I don't get is that regarding 2023, it would be an easy answer already.
Nate did say that what he reported on (regardless if it was Drake or not) was planned for end of 22/ early 23 and was shelved.

He could now easily say if there's another hardware (regardless if it's Drake or not) expected for this year or not.

This is not about the SoC. If he knows more about it, fine, give details in a podcast.

Bit the answer to new hardware that is not shelved in 2023 ist just yes, no or he doesn't know and there's no reason not to say apart from being vague on a purpose.

And if that's it, being purposely vague, I think not saying anything would be the better option.

Myain complaint here once again is not about sharing info or not, it's about t the style of communication, that is confusing and vague and I'm not the only one here complaining about it.
 
oldpuck said:
The GPU power draw? The power draw of a totally unrelated device?


Sure. Just no way of knowing. Let me get super in the weeds here, and lay out what happened, and why it's hard to know what these numbers mean. Feel free to ignore, it's too damn long.

  1. In 2015, Nvidia starts development of NVN.
  2. Nvidia uses a process called "CI/CD" or "Continuous Integration/Continuous Deployment"
  3. CI/CD means that every time a developer makes a code change to NVN, an automatic process compiles the software, and then runs a test suite to make sure the change works.
  4. The software that does CI/CD is Jenkins. Jenkins doesn't run on Nintendo's custom OS, which isn't even ready yet.
  5. Jenkins also needs a development toolchain to actually build the software
  6. And it also needs hardware which can run in a server farm somewhere, not hardware built to run on a desktop
  7. So somewhere, in a datacenter, there are a group of machines that look nothing like Nintendo hardware that run NVN tests all day
  8. These are probably AMD based Windows Servers with 8 Nvidia GPUs in a cluster, possibly not even desktop GPUs, but server specific GPUs
  9. Every day, Jenkins builds and tests NVN over and over again
  10. Over the years of development the test suite gets huge
  11. There is a test for every bug ever found in the course of development
  12. There is a functional test created by developers during development to verify functionality
  13. There are integration tests to make sure NVN and all it's parts work together
  14. And there are performance tests which show over time if NVN performance is getting better or worse
  15. Some of these performance tests are KPIs - Key Performance Indicators.
  16. Somewhere at Nvidia there is a giant dashboard that shows not only the current status of all of these things but the history - the performance of KPIs over time. Meanwhile...
  17. DLSS 2.0 development begins
  18. Just like NVN, it uses Jenkins and has a set of tests and KPIs
  19. At some point a developer builds dlssDonutTest, as a tool to test performance of DLSS.
  20. DLSS 2 doesn't work on it's own. it needs to be integrated into various Graphics APIs
  21. And they all need to be tested
  22. dlssDonutTest gets updated to run on DirectX11, DirectX12, and Vulkan, all with a standard set of KPIs
  23. This not only allows each integration to be tested, it allows them to be compared to each other
  24. If all three get slow, then DLSS is the problem
  25. If just one gets slow, it's the integration
  26. But for all this comparison to work the hardware and the configuration of it all needs to be consistent for every single test
  27. 2019, NVN2 Development begins
  28. NVN2 = NVN1 + DLSS + RT + Drake Specific Optimizations
  29. Step one for NVN2? Take a copy of NVN1, and copy the DLSS development code into it, begin integrating
  30. This is kinda similar to the integration of DirectX of Vulkan, in fact
  31. NVN2 inherits all of NVN1's tests, benchmarks, and KPIs. They may change over time, but that's the starting point, it's a huge resource
  32. It also inherits the DLSS test suite and benchmarks, including the dlssDonutTest.
  33. Just like NVN before it, it cannot run on Drake, as Drake doesn't even exist. It needs to run in the server farm, probably on windows.
  34. It also needs to inherit NVN's testing hardware
  35. Partially just because it already exists, but unless you're running benchmarks under the same setup you can't know if NVN2 is getting slower or faster than the original NVN1.
  36. You also want to keep the tests named the same so that you can see trends over time, even if those names don't reflect the current state of affairs. This is one of the most annoying parts of my job as a perf engineer, honestly.
  37. The CI/CD process takes some time. Developers want to quickly check if their changes will work before checking their code in and firing off Jenkins
  38. So a developer makes a script that runs some KPI benchmarks, but on his local machine simulating the testing machine which only has a vague relationship to the final hardware.
  39. This is the script being discussed
I realize this was a lot but it's important context. Everything in that list we basically know happened because it's either in the leak, or is absolute bog standard practice for the industry. Some version of this absolutely happened internally at Nvidia. But it leaves us with a bunch of questions

  • Where did this KPI come from? New for NVN2? Inherited from NVN1? Inherited from DLSS2?
    • This test was probably inherited from DLSS2. Earlier versions of this test script checked Windows Vulkan performance as well as NVN2. It seems highly likely that the NVN integration took the dlssDonutTest, and ran it using the same parameters as the other integrations. Early on, the Vulkan numbers were compared side by side with the NVN2 numbers to verify that the performance matched and NVN2 didn't have any issues
  • Do the clock speeds mean anything? Are they Drake clock speeds?
    • Almost definitely they are not. This script sets GPU parameters on a developer's machine in order to make their workstation (probably running a consumer GPU) act like the testing hardware in the server farm (running a server GPU) while Drake was in very early days of development, and was inherited from the early NVN1 development environment. The odds that the clocks on the workstation match the server clocks match the in development hardware, match the final numbers that Nintendo selects balancing battery life, performance, and developer requests is vanishingly small.
    • It is entirely possible - even likely - that these clocks were set as a testing baseline for DLSS and have nothing to do with NVN2 development at all.
  • What do the wattage numbers mean? Surely these are for Drake?
    • Maybe? But what part of Drake? The whole SOC? The GPU? Or even the whole console, including RAM, screen and Joy-Con draw?
    • Or they could have been inherited from Original Switch NVN development. After all, NVN probably ran it's tests under multiple performance profiles. It is totally reasonable to assume that a low/middle/high power configuration were picked, based on TX1's power draw... and the labels just stuck. As I said before, renaming/altering your performance tests is a general no-no, because it breaks your historical trends and comparisons. Labels and configs tend to stick around long after their original use case is gone.
    • Or they could refer to one of the various pieces of testing/development hardware. After all, Drake doesn't exist at the time these scripts are being used. These power numbers could refer to an Xavier testing board, or a virtual GPU partition on an A100 in the server farm.
    • Or these labels come from DLSS development. This is very unlikely, but it's not impossible. Nvidia was touting power savings for DLSS early on, these labels might be to match DLSS's test environment, with wattages refer to tensor core draw.
    • Or a combination of these things - Like the power draw referring to the testing profiles for early NVN1 development, which was done on TK1s instead of TX1s
  • Okay, what does Occam's razor say? Surely the simplest explanation is these are Drake clocks and Drake power draws?
    • No. That's only the simplest explanation if you ignore all the data above. Occam's razor says that the simplest thing that explains all the data is probably the true one. It doesn't mean the explanation that ignores some of the data ;)
    • The simplest explanation is these clocks have nothing to do with Drake. The odds that the clocks a developer using a desktop Turing based RTX 2060 is using to simulate tests running on an Ampere based server A100 GPU will match the final clocks of an in development die shrunk Drake, all of which have wildly different numbers of SMs, is vanishingly small. And all of that assumes that this test is, in fact, trying to simulate Drake at all, rather than just being a modified DLSS test.
    • These wattages probably do have to do with Drake in some capacity. The wattages only appear in NVN2 tests that we can see, and NVN2 is clearly optimized for Drake. It's not a slam dunk, but it is the most likely guess.
    • But we can't predict anything from these numbers without further context. Since the clock numbers likely have nothing to do with Drake's clocks, we can't guess at the final clocks. Since we have no idea whether or not this is just the GPU or the whole SOC, we can't guess at battery life.
Click to expand...
Hidden content is only available for registered users. Sharing it outside of Famiboards is subject to moderation.
 
NateDrake said:
What is there to take from the post to make into clickbait? I'm not offering up any details yet. I'm simply saying I'm going to have information that will answer many of the questions this thread has had.
Click to expand...

I agree, you're basically saying be patience and wait and see. But, we know some people will take that an infer from it that 2023 will be the launch window for Switch 2. Or something along those lines. You're stating all facts but people will look at everything you say with a magnifying glass to try and extract any but of info. It sucks, but it's the reality. Some people will latch onto anything.
 
kimbo99 said:
Nate we both know damn well click bait will come from these statements. The irony is you had a podcast about this issue.
Click to expand...
eye0h said:
Considering we can "form our own conclusions" the answer to the question "will new hardware come out in 2023?" is "maybe"?
Click to expand...
I would suggest you both chill out. He's recording a podcast soon. This is video games, nothing worth being this snarky/hostile about.
 
kimbo99 said:
I agree, you're basically saying be patience and wait and see. But, we know some people will take that an infer from it that 2023 will be the launch window for Switch 2. Or something along those lines. You're stating all facts but people will look at everything you say with a magnifying glass to try and extract any but of info. It sucks, but it's the reality. Some people will latch onto anything.
Click to expand...
I'm in general agreement. If one wants to infer something that was not said, that's on them. The inclusion of "maybe or maybe not" was done so such inference could not be made; but it may not have been clear enough.
 
Dmax3901 said:
I would suggest you both chill out. He's recording a podcast soon. This is video games, nothing worth being this snarky/hostile about.
Click to expand...


This is not an attack on Nate. It's the wack YouTubers and bloggers using anything as clickbait.
 
LiC said:
* Hidden text: cannot be quoted. *
Click to expand...
Hidden content is only available for registered users. Sharing it outside of Famiboards is subject to moderation.


One of the good ways to check in on this kind of predicting is "what would it look like if the most likely, reasonable conclusions were all correct." In the case of Drake you wind up with something like:

1) Drake's final GPU looks like it does in the hack: 12 Ampere SMs
2) The CPU is the same as the Linux driver, 8 A78Cs in a single cluster
3) It runs in a Switch form factor
4) It maintains Switch battery life
5) It is made on Samsung 8nm
6) These tests reflect target Drake power draws
7) And why not, target Drake GPU clock speeds

All of these are the most reasonable, middle of the road interpretations of all the data - and they cannot all be true at the same time. No way Drake is that big, running at those clocks, in that form factor, on that node, with those power draws.

No matter how smart this thread, or how much detective work we do, at least one "extremely unlikely" possibility is going to come true. Probably several. If you don't keep your mind at least slightly open to insane possibilities, then you will find yourself constantly trying to make new data fit the old "reasonable" narrative.
 
0
NateDrake said:
I'm simply saying to wait until I can present everything we have and answers to the question of 2023 hardware will come at that time.

That's all the "maybe or maybe not" mention was for. Covering both questions and saying I'll answer that question and more soon.

One can also use the information we present to form your own conclusions.
Click to expand...

Is it safe to assume that all of this will be the primary subject of conversation on your next podcast?
 
NateDrake said:
I'm simply saying to wait until I can present everything we have and answers to the question of 2023 hardware will come at that time.

That's all the "maybe or maybe not" mention was for. Covering both questions and saying I'll answer that question and more soon.

One can also use the information we present to form your own conclusions.
Click to expand...
Nate, will your upcoming podcast share only details up to that point where you announced the canned version or have you already collected more info to tell us more if a launch is possible in 2023?

And when do you think you will release this the furthest?

Thanks, as always!
 
Truno said:
I'm completely writing off H1, and I doubt Nintendo has any intention of hurting Switch's holiday sales. As of now, I'm expecting 2024
Click to expand...
We are 4 days into 2023 and we’re already going “bubububuBUT THE HOLIDAY SALES” with no indication of how the year will play out for Switch sales?
It begs the question of what hardware sales level in the holiday period would be low enough to end this frankly baffling level of pearl-clutching on the subject. Or are we waiting until we’re getting out the shovels to bury the Switch after it’s fully dead and gone?
Skittzo said:
Here's a question for you. Not trying to force you to believe anything one way or the other:

If we forget all rumors, everything from any journalist on the subject, what are we left with? We're left with leaked data from an illegal hack that Nvidia confirmed was real, and publicly accessible information from the Linux kernel.

All of the information we can glean from those sources suggests that a chip for a Nintendo product has been completed and likely produced physically, at least for samples, as of April 2022. People with experience following chip development say that you likely aren't going to be waiting 2 years from when samples were produced for the chip to be in a released product.

So again, if you ignore every rumor on the subject the actual hard, verified data heavily suggests a 2023 launch.
Click to expand...
You also have the financial data to rely on that gives a relatively clear indicator of when R&D for this device started and can extrapolate from that when it can be expected to end, which greatly supports the 2023 hypothesis.
Gotdatmoneyy said:
Launching a successor doesn't mean that hardware sales for current hardware automatically go to 0 lol. If Nintendo wants the OG Switch to sell well during the Holidays they can reduce the price, bundle, introduce cheaper software etc. The only way OG Switch sales go extremely low is if Nintendo does not produce it or does nothing to make it attractive.

I don't get this desire to ride the OG Switch into oblivion before launching a successor. This is a bad way to optimize profits. If Nintendo immediately sunsets the OG Seitch in favour of the Switch 2 I will also call that a bad practice. You want your new product riding high and turning the profit corner as your old product is reaching end of life. You don't want sharp partitions. What Sony did with the PS4 to PS5 transition is the wrong approach.
Click to expand...
I’ve been banging this drum for months, it never seems to get traction, but you’re welcome to join me in the drumline.
MrPenguin said:
Do you guys think we can see More AAA Third Party Games come to Next gen nintendo console? Such as Grand Theft Auto 6, Red dead, Payday 3, Stalker 2, etc.
Of cource is scaled back verison probaly.
Click to expand...
The PS4 cross-gen games (that are continuing on into 2024 as of right now) have felt like a lock to me for some time now, we should be able to rely on a large number of them making the jump, at least.
 
Insider Culture Discussion 1/4/23
  • Index
    • Hey everyone, this is just a heads-up to let you all know that the moderation team is currently in the process of prepping a thread for an open community discussion surrounding insider culture. We'd like the conversation to encompass things such as how members respond to insider information, how the community would prefer we handle the often impassioned sentiments that emerge, and so on.

    We're open to hearing from both supporters and skeptics. All we ask is that the dialogue remain civil, with all of us working together toward a productive framework for how these matters are approached by the community, and moderated by staff. Speaking candidly, the team finds the reports that emerge during periods of new info circulating to be some of the most challenging to handle. We simply wish to ensure we're doing right by the community, which is to be the purpose of the thread. We'll aim to have it out within the next couple days.

    - Aurc, Red Monster, Derachi
     
  • Index
    • NateDrake said:
    With that being said: in the closing days of 2022, I'm going to reevaluate my time & participation on this site. Attempting to answer questions and partaking in the community is becoming an unpleasant affair &, frankly, one I can do without.
    Click to expand...
    Late reply, but I really appreciate your participation in this forum and input. It helps to drive some discussion and make the place more fun.

    I would venture to say the vast majority of forum members (and lurkers) would agree.

    Sometimes folks get too wound up over a topic, for whatever reason specific to their situation. They temporarily forget that this place is all about sharing good times around a common passion.

    At times like these I think back to what a videogame wiseman(guy?) once said.

    My block button is bigger than ever!
    Click to expand...

    ===

    Looking forward to 2023! 🎊

    Would be stoked for powerful new hardware 🙏

    But we nevertheless on the cusp of a new Fire Emblem, along with the assurance of a date with Zelda this spring. They took their damn time, but it's nearly here and its probably worth the wait 🤞
     
    I know the thread isn't up yet but I think the ideal way to handle insiders and their information is just to tag post if the claim was ever correct or false at the given time. Gives users transparency if they want to rely on their info on the site.
     
    oldpuck said:
    We're all onprem github, and it's no better :(
    Click to expand...
    karmitt said:
    Atlassian suite is my friend
    Click to expand...
    Honestly, I mostly hate Jenkins because of my company's implementation, rather than Jenkins itself (although Jenkins is quite slow).
    Hidden content is only available for registered users. Sharing it outside of Famiboards is subject to moderation.
     
    the_marmolade said:
    I know the thread isn't up yet but I think the ideal way to handle insiders and their information is just to tag post if the claim was ever correct or false at the given time. Gives users transparency if they want to rely on their info on the site.
    Click to expand...
    This isn't a team correspondence, just my own, personal impression: the pitfall I see with this is that it turns moderators into fact checkers for the forum, when we're primarily responsible for ensuring members are engaging within the boundaries of site guidelines. Shouldn't the community draw their own conclusions, regarding insider info? You also have to factor in the fact that so much of the info is subject to change, that unless we're tagging old claims with an official "This claim was proven true" message upon full-on, unambiguous confirmation from Nintendo (which then makes our verification pointless), we can't really be expected to know what's true, and what's false, what may come out, what's been pushed back, etc.

    Folks are free to float their ideas, and I'm not pushing this one off the table, but I don't really see how it could work in a satisfactory way.
     
    Some people are just being nasty for no reason besides it is the internet and you can say any bullshit to someone without real repurcussions. People like NateDrake have enough credibility and correctly shared info that there shouldn't be interogation and name calling anytime something they say isn't 100% accurate or what people want to hear.

    I don't think this is difficult. Some of the posts callin out insiders are wack as fuck and should just garner heavy handed bans because people are not being cautious, they're being fucking assholes. It is an internet forum about video games. None of this, especially info about hardware is of such importance that some of the behaviour towards NateDrake or others is acceptable.

    You have to be responsible for your own hype and your own emotions. If you can't read some shared info without a critical eye and understanding that things change in business all the time, then you really just need to go outside and touch grass.

    Most of us are adults. There aint no way that if you shared NDA information from your job with someone, you'd give them so much detail as to out yourself and there is no way anyone who has worked even moderately high up in a big company can say that plans don't change, all the fucking time.

    I mean come on yall.
     
    Aurc said:
    This isn't a team correspondence, just my own, personal impression: the pitfall I see with this is that it turns moderators into fact checkers for the forum, when we're primarily responsible for ensuring members are engaging within the boundaries of site guidelines. Shouldn't the community draw their own conclusions, regarding insider info? You also have to factor in the fact that so much of the info is subject to change, that unless we're tagging old claims with an official "This claim was proven true" message upon full-on, unambiguous confirmation from Nintendo (which then makes our verification pointless), we can't really be expected to know what's true, and what's false, what may come out, what's been pushed back, etc.

    Folks are free to float their ideas, and I'm not pushing this one off the table, but I don't really see how it could work in a satisfactory way.
    Click to expand...

    It’s also impossible to know what is true or false when plans do literally change. Something leaked may not happen, but it doesn’t make the leak any less true when it was stated. It’s just the nature of these things.
     
    I don't think there ever was a Pro hardware. NVN2 = Drake and that's all there ever was.

    Nintendo is too cheap to pay for R&D on a modern chipset design and then not use it, modern chip designs are not cheap and Nvidia doesn't make like casually upgraded versions of the Tegra X1 or X2 ... you can't just order one with more CUDA cores or something like adding extra fries to a fast food order, they don't work like that. And I mean rightfully so, if Yamauchi was alive today and saw how much modern games and hardware cost to make, there's no way in hell they'd just casually have an upgraded hardware model that they don't release to recoup cost on.

    I think what more than likely happened is COVID had a much larger impact on Nintendo's hardware plans than people may understand, it completely altered time lines by 1-2 years I think, we're talking about a massive global event, frankly the biggest thing to happen to global supply chains since World War II, far bigger than 9/11 or anything like that. Switch Next-Gen will be the first (major) new hardware that launches post-COVID, I think we'll learn in the future that COVID had a massive impact on it's development and release.
     
    Reader mode
    You must log in or register to reply here.
    Please read this staff post before posting.

    Furthermore, according to this follow-up post, all off-topic chat will be moderated.
    Last edited:


    Back
    Top Bottom