Support NeoGAF

[wonderful103]

So the X1 doesn't have Gpgpu extra sauce like in the WiiU right if it was taken advantage of properly?

Every modern platform can take advantage of this, but the reason it wasn't and isn't some kind of panacea is because gpu's are still not designed for general purpose code. They are also a limited resource, and as you can clearly see most developers are using them to make more shiny.

 

[joesiv]

This fp16 thing that has potential to improve performance overall. Is it easy for devs to implement on ports or is it too difficult that most will not bother using it? According to this article it seems it can only be used in special cases as well. I wonder how many special cases there are when porting games across from other platforms. Just trying to guage if devs will bother with it or not. I'm guessing ground up development yes, they will try and use it as much as possible but 3rd party ports its too expensive. Like I always hear. Devs are lazy, especially when porting games over.

It's just a data type, you declare a datatype when you create the variable. You could search and replace all FP32 variables to FP16 in one search/replease command... of course the results would suck.

There are probably some very obvious variables in certain systems that don't need the precision of FP32, they'd obviously be easy to implement, and you'd get some gains in performance, as it frees up resources. But going through with a find comb brush will take some developer/qa testing time. Consider that in most cases FP16 will work fine, but you will see precision artifacts, since (duh), it's not as precise (by a HUGE margin).

 

[Pasedo]

Further reading about fp16 I'm getting that a benefit is better power efficiency. So perhaps on portable mode devs will probably use it more often than optimising code for docked mode which will probably be fp32 as it sounds more versatile. A challenge for fp16 as well is to ensure its not bottlenecked with poor bandwidth. Is Switch bandwidth pretty average?? Trying to sound smart here am I doing okay?

 

[z0m3le]

It's just a data type, you declare a datatype when you create the variable. You could search and replace all FP32 variables to FP16 in one search/replease command... of course the results would suck.

There are probably some very obvious variables in certain systems that don't need the precision of FP32, they'd obviously be easy to implement, and you'd get some gains in performance, as it frees up resources. But going through with a find comb brush will take some developer/qa testing time. Consider that in most cases FP16 will work fine, but you will see precision artifacts, since (duh), it's not as precise (by a HUGE margin).

Exactly this. You'll also likely see fp16 used more as time goes on in ports as even PCs will start using it, as well as Pro and Scorpio. Exclusives will push it a bit more, but the 70% we throw around (around 600gflops mixed precision) should not be expected in a general use as it is about what devs are asking the gpu to do.

Developers are lazy sure, but the laziest option is to test fp16 on tasks, rather than creating new assets or code to get your game to stable performance levels.

Further reading about fp16 I'm getting that a benefit is better power efficiency. So perhaps on portable mode devs will probably use it more often than optimising code for docked mode which will probably be fp32 as it sounds more versatile. A challenge for fp16 as well is to ensure its not bottlenecked with poor bandwidth. Is Switch bandwidth pretty average?? Trying to sound smart here am I doing okay?

If the code runs fine in portable, creating separate 32bit code offers only complications and disadvantages, they would use the same fp16 code in docked mode. Simply, what works, is what works.

 

[FN-2187]

No it's not. I don't know how many more times do we need to state that the Wii was just an overclocked GameCube with the same API and more RAM. It was like 50% more powerful. If you think Switch is 50% more powerful than Wii U... reconsider. Fast and Zelda already show improvements that weren't possible going from GC to Wii and they're launch ports.

I am really not talking about the hardware specs at all (I realize I am in a thread about Switch hardware specs), and I do realize there is a bigger power gap between Switch/Wii U and Wii/GameCube. I was trying to say that conceptually Nintendo is content with Wii U-level graphics with the upgrade being that it's now fully portable, similar to how they were content with GameCube-level graphics on the Wii with the upgrade being motion controls.

 

[sugarhigh4242]

As others have said, this simply isn't true. The Shield TV throttles after just a few minutes to the same levels we see in the docked Switch, so Shield TV performance should actually be the baseline for expectations for the docked Switch.

However, this is without taking into account the massive Android overhead the Shield TV has to use which substantially reduces performance. So we should be seeing games perform quite a bit better on the Switch than on the Shield TV.

And the Switch has an extra 1GB of RAM, which people (including me just now) tend to forget.

I don't see anything online talking about cpu throttling on the Shield to justify what you're saying.

Digital Foundry has the Switch CPU running at 1Ghz in both docked and portable mode, which is half the speed of the Shield TV. The GPU difference is even more profound, running at 307Mhz in portable mode, down from 1Ghz on the Shield.

You got a point with Android overhead, but I certainly would not call it "massive". The Switch OS is pretty much a shell at this point, so it's pointless to talk about the system overhead until the Switch system actually does half of what a gaming OS is supposed to do.

 

[th3sicknness]

I don't see anything online talking about cpu throttling on the Shield to justify what you're saying.

Digital Foundry has the Switch CPU running at 1Ghz in both docked and portable mode, which is half the speed of the Shield TV. The GPU difference is even more profound, running at 307Mhz in portable mode, down from 1Ghz on the Shield.

You got a point with Android overhead, but I certainly would not call it "massive". The Switch OS is pretty much a shell at this point, so it's pointless to talk about the system overhead until the Switch system actually does half of what a gaming OS is supposed to do.

It is pretty much common knowledge that x1 throttles down during intensive tasks.. Most products with tegra in it like the pixel tablet will throttle its performance down.

 

[z0m3le]

I don't see anything online talking about cpu throttling on the Shield to justify what you're saying.

Digital Foundry has the Switch CPU running at 1Ghz in both docked and portable mode, which is half the speed of the Shield TV. The GPU difference is even more profound, running at 307Mhz in portable mode, down from 1Ghz on the Shield.

You got a point with Android overhead, but I certainly would not call it "massive". The Switch OS is pretty much a shell at this point, so it's pointless to talk about the system overhead until the Switch system actually does half of what a gaming OS is supposed to do.

It's just about what they can cool and power in the handheld mode, maybe they could have pushed the cpu as high as 1.4ghz, but then the fan would likely be louder, and it would probably need to stay at 307mhz on the gpu rather than 384mhz that they have it running at now. The issue is that whatever the cpu runs in handheld is what it needs to run docked, and because the cpu is clocked low, they can have a higher gpu clock.

Finally, there is probably still room for a 1.2ghz clock for the cpu, half a watt isn't much and would benefit ports greatly.

 

[sugarhigh4242]

It is pretty much common knowledge that x1 throttles down during intensive tasks.. Most products with tegra in it like the pixel tablet will throttle its performance down.

Nothing about the X1 platform is "common knowledge". Please link.

 

[z0m3le]

Nothing about the X1 platform is "common knowledge". Please link.

Look up the user mdave on here, he did some tests a month or two ago in the switch speculation there with his shield tv

 

[Skittzo0413]

Nothing about the X1 platform is "common knowledge". Please link.

Start from here and keep going. MDave shows how the Shield TV throttles after very little time to about Switch levels.

EDIT: I think this post basically shows his conclusion:

Oh man, this is interesting. The CPU is definitely throttling the GPU when its at 2GHz. And it looks like the other way around is true too.

Here are the results when I don't lock the CPU to any frequency, the kernel / governor manages it all:

These are the best results. This is also when the GPU went as low as 768MHz when it was, I suspect, it being throttled.

https://puu.sh/tfUTW/8bb1f4b217.jpg

Here are the results of locked 2GHz CPU benchmarks. Notice it has actually performed worse then a non-locked CPU frequency! The GPU sometimes goes as low as 537MHz.

http://puu.sh/tgLof/19520ff241.jpg
http://puu.sh/tgLJs/dd2570df40.png

And this is the results when the CPU is locked at 1GHz. The GPU is able to stay much closer to 1GHz. But it looks like the GPU might be thermal throttling the CPU, as the CPU can't keep a lock at 1GHz for some reason.

http://puu.sh/tgLGp/0c42f6c426.jpg
http://puu.sh/tgLLw/deb8b6bbe1.png

It's looking like the CPU and GPU cannot operate at their maximum frequencies when both are pushed as hard as they can at the same time. So to get the most out of it, best to develop a game that doesn't push the CPU too hard to get the most out of the GPU, and vice versa

Lastly, it's not simple to turn off vsync after all. Looks like Android relies on it at a core level.

And remember that a game console (Switch) cannot throttle at all, so it would need to be clocked at the lowest clock rate that the TX1 in the Shield TV typically operates, in order to completely avoid the need to throttle.

 

[joesiv]

You'll also likely see fp16 used more as time goes on in ports as even PCs will start using it, as well as Pro and Scorpio.

People keep saying this, but I have my doubts to be honest. Pro and Scorpio already have such a performance boost over the vanilla consoles, that why put in the extra effort, especially when the vanilla consoles can't take advantage of it, and they're the ones that need it the most.

I think the low hanging fruit will be done, and maybe as they refine the engine they'll take into account some FP optimizations that are obivous, but I don't think it'll affect much in the long run.

I could just see some diligent SE going in and changing a bunch of FP32's to FP16's, a week later QA complains that some minor aspect of image quality has artifacts, and logs a bug. For that 1/1600th of a boost in performance, the SE just reverts his change, and never makes such a change again... lol

 

[Pasedo]

If at some point 3P's were serious about releasing their games for the Switch, would it be easier/faster/cheaper for them to develop the game first on the Tegra X1 and then porting it over to console/PC code, rather than the other way around? It also makes more sense to start with the lowest common denominator and then scale up from there. I hope 3P's seeing the initial success of Switch sales have started to do this with games which have just started development. How great would it be for Switch owners if the game was coded for Switch first

 

[Polygonal_Sprite]

I tried Splatoon 2 docked tonight. Looks like it's 720p. Very similar image quality to the original. Disappointing.

 

[orioto]

About Zelda people are sooo underestimating how crazy demanding the game is. I sometimes wonder how it can even run on this type of hardware.

This is not realized enough how everything has to be calculated, not only graphic wise, at an insane distance, and the whole game is built on that aspect. When you're doing crazy shits with enemies and physics etc.. the game knows 3 miles away from that, there is a giant dragon flying around. It knows the arrow you thrown 2 minutes ago from a mountain fell near a river friggin' far from where you are etc..

I mean i'm pretty sure 99% games don't bother with this at all. It must make everything so difficult to optimize and to run. Not only the size of the game but the surface that is actively calculated and tested for many things is insane.

 

[Not]

I tried Splatoon 2 docked tonight. Looks like it's 720p. Very similar image quality to the original. Disappointing.

It's gotta be the online factor. Ninty don't have that together yet, for all their technical wizardry in other areas.

 

[Pasedo]

Another thought is that most 3rd party PC games run better on Nvidia hardware..or is at least the general consensus in the PC world. Seeing as Switch uses Nvidia based on the same PC Maxwell architecture, would this give it some extra performance advantage for 3P game ports?

 

[z0m3le]

People keep saying this, but I have my doubts to be honest. Pro and Scorpio already have such a performance boost over the vanilla consoles, that why put in the extra effort, especially when the vanilla consoles can't take advantage of it, and they're the ones that need it the most.

I think the low hanging fruit will be done, and maybe as they refine the engine they'll take into account some FP optimizations that are obivous, but I don't think it'll affect much in the long run.

I could just see some diligent SE going in and changing a bunch of FP32's to FP16's, a week later QA complains that some minor aspect of image quality has artifacts, and logs a bug. For that 1/1600th of a boost in performance, the SE just reverts his change, and never makes such a change again... lol

Yep, but eventually ps5 will come out, and there is this idea in my head at least that developers will gain knowledge on what can be done in FP16, because it is universal right? Like if you find that "xcode" can be done in FP16, that becomes something that should always be able to be fp16. For instance, a really good dof technique, can always be done in FP16 if they first find that it works correct? In fact isn't post processing techniques largely good targets for fp16?

Another thought is that most 3rd party PC games run better on Nvidia hardware..or is at least the general consensus in the PC world. Seeing as Switch uses Nvidia based on the same PC Maxwell architecture, would this give it some extra performance advantage for 3P game ports?

Well as they are making the PC version as the lead platform instead, they can just make sure it works on Intel IGP and amd APUs (this one is still the best for them) switch is cross compatible with PC development so looking at what issues they come across on AMD APUs with 19.2GB/s memory bandwidth at best and working that stuff out, allows them to move over that game to switch and see it run similarly, I actually don't think ports are going to require herculean effort, but just some reasonable adjustments to the render pipelines and optimizing cpu code. That is not to say that they won't find games that take much more effort, but I don't think switch is magically going to start getting every third party, even if it did end up selling like the Wii.

 

[Seik]

I tried Splatoon 2 docked tonight. Looks like it's 720p. Very similar image quality to the original. Disappointing.

True, but I expect it to have a higher res at release though.

At least 900p, it wouldn't make sense to have absolutely no gains by docking the console to me.

 

[orioto]

It's gotta be the online factor. Ninty don't have that together yet, for all their technical wizardry in other areas.

I'm really not getting why the game is not 720p (it seems it's sub native) undocked. ONline factor ? But MK8 runs perfectly and the what is the game doing that the WiiU wasn't ?

And it sets a terrible terrible trend for the Switch, of subnative games undocked and 720p docked..

True, but I expect it to have a higher res at release though.

At least 900p, it wouldn't make sense to have absolutely no gains by docking the console to me.

There is a gain actually, that's the worst part


Edit: Well actually false alarm the game IS 720p and native undocked.

 

[VanWinkle]

About Zelda people are sooo underestimating how crazy demanding the game is. I sometimes wonder how it can even run on this type of hardware.

This is not realized enough how everything has to be calculated, not only graphic wise, at an insane distance, and the whole game is built on that aspect. When you're doing crazy shits with enemies and physics etc.. the game knows 3 miles away from that, there is a giant dragon flying around. It knows the arrow you thrown 2 minutes ago from a mountain fell near a river friggin' far from where you are etc..

I mean i'm pretty sure 99% games don't bother with this at all. It must make everything so difficult to optimize and to run. Not only the size of the game but the surface that is actively calculated and tested for many things is insane.

It IS a demanding game...that runs just as well on their last console and is almost exactly the same from a tech standpoint (save for a resolution bump). People were expecting a pretty big bump up from the Wii U version.

 

[KingSnake]

At least 900p, it wouldn't make sense to have absolutely no gains by docking the console to me.

Depends where the bottleneck is. Which applies for anything, including Zelda.

Splatoon must be 60 fps both in handheld mode and console mode. The environment is continuously changing in this game. I wonder if that doesn't hit the same limitations of memory bandwidth like Zelda does, according to DF's speculation.

 

[krumble]

It's gotta be the online factor. Ninty don't have that together yet, for all their technical wizardry in other areas.

Yet the online play tonight during the test fire was top notch from what I played, wireless and wired (at least in my case)

 

[icecold1983]

Another thought is that most 3rd party PC games run better on Nvidia hardware..or is at least the general consensus in the PC world. Seeing as Switch uses Nvidia based on the same PC Maxwell architecture, would this give it some extra performance advantage for 3P game ports?

not really the case anymore for big titles

 

[ultrazilla]

I tried Splatoon 2 docked tonight. Looks like it's 720p. Very similar image quality to the original. Disappointing.

I agree. Here's my post from the Splatoon 2 thread:

I hate the motion controls so I turned that off quick. Can't find a Pro Controller anywhere which I think would be perfect for this game. The joyconns are ok with it. AFK players seemed to be in every game I played and it wasn't fun getting destroyed.

I played the Wii U version and honestly, I'm not seeing much difference in graphical quality here to be honest. A tad better *perhaps* but nothing remotely superior to the first game.

So far, the Switch appears to basically be a portable Wii U in terms of actual graphics.

We need a developer out there that can really push the damn chip and give us a good looking game that can stand up to a PS4 or Xbox One title. I think it's possible.

 

[Hermii]

Docked splatoon is probably still not optimized enough for perfect framerate

 

[Pasedo]

Well as they are making the PC version as the lead platform instead, they can just make sure it works on Intel IGP and amd APUs (this one is still the best for them) switch is cross compatible with PC development so looking at what issues they come across on AMD APUs with 19.2GB/s memory bandwidth at best and working that stuff out, allows them to move over that game to switch and see it run similarly, I actually don't think ports are going to require herculean effort, but just some reasonable adjustments to the render pipelines and optimizing cpu code. That is not to say that they won't find games that take much more effort, but I don't think switch is magically going to start getting every third party, even if it did end up selling like the Wii.

3P's after all are businesses and if you show them the money they should jump on board for every game. If porting is cheap enough for them and so long as there's an install base to turn around a profit, then why not make some extra cash. Every dollar counts

Rough and I think conservative numbers. Say it takes 4 months to Port a game across to Switch's Tegra X1 with a team of 10 who earn Avg $50 p/h and 40h week. That's 320k cost approx. Say bump that up to 500k for licensing and any other costs. I think Switch has sold 1.5m units in its first few weeks. At this rate it will be 6m units in a few months. If install rate is even only 10% at $60 bucks a pop that's a revenue of 36m for the publisher. Say 20% of that is pure profit. Than they've still banked 6.7m after porting costs. And I think the numbers used are fairly conservative.

Seems from a business sense it's sound, so perhaps the main decider for 3P's then is if the visuals and frame rates would need to be compromised so much it would look and play horribly on the Switch. So probably more of a branding/PR concern. Personally I'm willing for a compromise of at least 480p and 720p docked/undocked respectively with locked frame rates of 30fps. That's their min 10% install base right there lol. PS. Best to check my calculations are right as I'm still half asleep lol. Based in USD as well.

 

[icecold1983]

3P's after all are businesses and if you show them the money they should jump on board for every game. If porting is cheap enough for them and so long as there's an install base to turn around a profit, then why not make some extra cash. Every dollar counts

Rough and I think conservative numbers. Say it takes 4 months to Port a game across to Switch's Tegra X1 with a team of 10 who earn Avg $50 p/h and 40h week. That's 320k cost approx. Say bump that up to 500k for licensing and any other costs. I think Switch has sold 1.5m units in its first few weeks. At this rate it will be 6m units in a few months. If install rate is even only 10% at $60 bucks a pop that's a revenue of 36m for the publisher. Say 20% of that is pure profit. Than they've still banked 6.7m after porting costs. And I think the numbers used are fairly conservative.

Seems from a business sense it's sound, so perhaps the main decider for 3P's then is if the visuals and frame rates would need to be compromised so much it would look and play horribly on the Switch. So probably more of a branding/PR concern. Personally I'm willing for a compromise of at least 480p and 720p docked/undocked respectively with locked frame rates of 30fps. That's their min 10% install base right there lol. PS. Best to check my calculations are right as I'm still half asleep lol. Based in USD as well.

Wasnt it nvidia who said just under a year was the estimated time to port a game to switch?

 

[Seik]

Wasnt it nvidia who said just under a year was the estimated time to port a game to switch?

It was Miyamoto IIRC.

 

[Instro]

I tried Splatoon 2 docked tonight. Looks like it's 720p. Very similar image quality to the original. Disappointing.

Would be pretty terrible if the game actually launches that way. Yeesh.

 

[Polygonal_Sprite]

True, but I expect it to have a higher res at release though.

At least 900p, it wouldn't make sense to have absolutely no gains by docking the console to me.

It could have been 900p but if it is then it's an extremely jaggy 900p. It definitely looked like an upscaled image to me, could have even been a dynamic resolution to hold 60fps like Fast Racing Remix.

I don't see why the test should be running at a lower resolution than the final game at this stage of development. The game is due out in the next three months not a year.

I don't understand why resolution is an issue for Splatoon 2 tbh, it really does look very similar to the previous Wii U Splatoon so it should be hitting native 1080p when docked like MK8 is although at this stage it really wouldn't surprise me to find out MK8 is native 900p with the PR saying it's 1080p.

Like I said before if developers are going after games that look a lot better than Wii U games then I wouldn't expect anymore than 900p (maybe even 720p). 1080p for ports of Wii U games.

 

[icecold1983]

It could have been 900p but if it is then it's an extremely jaggy 900p. It definitely looked like an upscaled image to me, could have even been a dynamic resolution to hold 60fps like Fast Racing Remix.

I don't see why the test should be running at a lower resolution than the final game at this stage of development. The game is due out in the next three months not a year.

I don't understand why resolution is an issue for Splatoon 2 tbh, it really does look very similar to the previous Wii U Splatoon so it should be hitting native 1080p when docked like MK8 is although at this stage it really wouldn't surprise me to find out MK8 is native 900p with the PR saying it's 1080p.

Like I said before if developers are going after games that look a lot better than Wii U games then I wouldn't expect anymore than 900p (maybe even 720p). 1080p for ports of Wii U games.

Was splatoon 60 fps on wiiu?

 

[lwilliams3]

It could have been 900p but if it is then it's an extremely jaggy 900p. It definitely looked like an upscaled image to me, could have even been a dynamic resolution to hold 60fps like Fast Racing Remix.

I don't see why the test should be running at a lower resolution than the final game at this stage of development. The game is due out in the next three months not a year.

I don't understand why resolution is an issue for Splatoon 2 tbh, it really does look very similar to the previous Wii U Splatoon so it should be hitting native 1080p when docked like MK8 is although at this stage it really wouldn't surprise me to find out MK8 is native 900p with the PR saying it's 1080p.

Like I said before if developers are going after games that look a lot better than Wii U games then I wouldn't expect anymore than 900p (maybe even 720p). 1080p for ports of Wii U games.

That wasn't PR. Additional sites, including DF, played the game and deducted that it was truly running 1080p.

Also, the devs for Fast Racing Remix stated that the resolution should be a solid 1080p after Nintendo patches up a bug in which the OS saps GPU power when the game is running.

As for Splatoon, I would deduct that they are ensuring that the game can properly run in HH settings before they try to upgrade anything for console-mode.

Was splatoon 60 fps on wiiu?

Looks like it did.

 

[DeuceGamer]

It could have been 900p but if it is then it's an extremely jaggy 900p. It definitely looked like an upscaled image to me, could have even been a dynamic resolution to hold 60fps like Fast Racing Remix.

I don't see why the test should be running at a lower resolution than the final game at this stage of development. The game is due out in the next three months not a year.

I don't understand why resolution is an issue for Splatoon 2 tbh, it really does look very similar to the previous Wii U Splatoon so it should be hitting native 1080p when docked like MK8 is although at this stage it really wouldn't surprise me to find out MK8 is native 900p with the PR saying it's 1080p.

Like I said before if developers are going after games that look a lot better than Wii U games then I wouldn't expect anymore than 900p (maybe even 720p). 1080p for ports of Wii U games.

Digital Foundry tested Mario Kart 8 Deluxe and it was 1080p, IIRC.

Edit: beaten and apparently correct

 

[Polygonal_Sprite]

Fair enough guys. I have MK8 Deluxe pre ordered so I'm hoping for 1080 when docked.

Interesting about the Fast Racing resolution being a bug. The game is a great technical achievement especially from four developers and extremely fun but it's really not pushing detailed textures or large scale course design. It should be1080p without a doubt on docked Switch.

Yeah the original Splatoon was 60fps. If there was a game that shouldn't be struggling for resolution on Switch it should be Splatoon 2. Low player counts, small basic maps with some impressive paint shaders.

 

[Pasedo]

Wasnt it nvidia who said just under a year was the estimated time to port a game to switch?

Even better for the publisher then. 18m install base if rate maintains with lower proportionate costs. Get on to it 3rd parties! Lol. Should be some interesting announcements at E3 with potential Jan - March 2018 quarter 3rd party launches if we stick to this time frame

 

[icecold1983]

Even better for the publisher then. 18m install base if rate maintains with lower proportionate costs. Get on to it 3rd parties! Lol. Should be some interesting announcements at E3 with potential Jan - March 2018 quarter 3rd party launches if we stick to this time frame

18 mil install base? Not gonna happen for a while

 

[Pasedo]

Doesn't seem like it's slowing down from 1.5m sold so far and usually the more momentum you build earlier the more hype, word of mouth etc starts to spread, which drives sales even further. I don't think its unachievable.

 

[icecold1983]

Doesn't seem like it's slowing down from 1.5m sold so far and usually the more momentum you build earlier the more hype, word of mouth etc starts to spread, which drives sales even further. I don't think its unachievable.

Beyond the fact that i dont think sales will maintain, nintendo isnt even going to be producing that many of them for quite some time by their own reports

 

[bomblord1]

Beyond the fact that i dont think sales will maintain, nintendo isnt even going to be producing that many of them for quite some time by their own reports

They're manufacturing 16mil for the first year on the market

 

[z0m3le]

Beyond the fact that i dont think sales will maintain, nintendo isnt even going to be producing that many of them for quite some time by their own reports

Lol weird wording "fact" about what you think.

Also, they are producing 16 million from April til the end of March 2018, that doesn't include the 2.5m they have for this month, giving 18.5 million switch devices as the target for manufacturing over the next 12 months.

 

[Pasedo]

Beyond the fact that i dont think sales will maintain, nintendo isnt even going to be producing that many of them for quite some time by their own reports

Wall Street Journal reports they've amped up to 16m or more from 1st April year out. Sounds like it will cover most of this demand. They're also taking into account preorder data and perhaps even markets they are still yet to enter. PS4 sold about that in its 1st year. Still doable in today's market. Nvidia are doing okay out of this as they're getting paid upfront regardless of whether that '16m or more' sells or not.

https://www.google.com.au/amp/www.w...ouble-production-of-switch-console-1489728545

 

[qko]

Seems like people don't care if it's an X1 or X2. They just want to play Zelda, even if it's on last gen tech:

https://gamerant.com/zelda-breath-wild-wii-u-sales-increase-gamestop/

 

[icecold1983]

Wall Street Journal reports they've amped up to 16m or more from 1st April year out. Sounds like it will cover most of this demand. They're also taking into account preorder data and perhaps even markets they are still yet to enter. PS4 sold about that in its 1st year. Still doable in today's market. Nvidia are doing okay out of this as they're getting paid upfront regardless of whether that '16m or more' sells or not.

https://www.google.com.au/amp/www.w...ouble-production-of-switch-console-1489728545

Lol weird wording "fact" about what you think.

Also, they are producing 16 million from April til the end of March 2018, that doesn't include the 2.5m they have for this month, giving 18.5 million switch devices as the target for manufacturing over the next 12 months.

They're manufacturing 16mil for the first year on the market

I see. I thought i had read something about them aiming to produce a bit less than 10 mil for first year or so

 

[Instro]

I see. I thought i had read something about them aiming to produce a bit less than 10 mil for first year or so

That was prior to launch. They changed their production target a week or two ago.

 

[z0m3le]

That was prior to launch. They changed their production target a week or two ago.

In January after pre-orders

 

[Pasedo]

From their initial forecast of 8m it sounds like Nintendo were playing it safe using the cheaper off the shelf X1. Perhaps selling 8m Switchs at the cost of the standard X1 would allow them to hit their profit targets. They are innovators but I wish they took a greater business risk on this one. If they invested twice as much on the chip (R&D, manufacturing etc) and do end up selling double their forecast they would have still hit their profit dollars and we would have had a better chip in the Switch.

 

[kIdMuScLe]

From their initial forecast of 8m it sounds like Nintendo were playing it safe using the cheaper off the shelf X1. Perhaps selling 8m Switchs at the cost of the standard X1 would allow them to hit their profit targets. They are innovators but I wish they took a greater business risk on this one. If they invested twice as much on the chip (R&D, manufacturing etc) and do end up selling double their forecast they would have still hit their profit dollars and we would have had a better chip in the Switch.

Which better chip? The X1 was the only one available

 

[Pasedo]

Which better chip? The X1 was the only one available

As in worked with Nvidia to further improve the X1 making it truly semi custom and enhanced. Something Nvidia kinda eluded they did but now we know didn't lol. I'm not a tech guy but I'm sure the tech gaffers in here can think of how they could improve the chip with twice the budget. I'd say at least improve the memory bandwidth somehow. Perhaps adopt more Pascal architecture into it. Find ways to increase speed by reducing throttle etc...

 

[PdotMichael]

Which better chip? The X1 was the only one available

The best Nvidia's chip.