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

[mariodk18]

This is how DeepL translated it:

“There are no announcements of the latest models that will be the highlight of this fiscal year.”

Edit: this translation sounds more like they’re simply stating things as is. More like a “there are no announcements at this time”.

 

[ShadowFox08]

Perhaps it's just semantics here. Didn't the same thing happen with New 3ds (or was it just 3ds?) "There will be no announcement" and then the hardware literally gets revealed like a week later.

This is different from a definitive, "we will not be releasing new hardware this fiscal year" as going against that would really upset Nintendo's investors.

I dunno. I mean I do think it will most likely be released with Botw 2 early next year, but how and what nintendo phrases their words, do matter.

 

[Skittzo]

Perhaps it's just semantics here. Didn't the same thing happen with New 3ds (or was it just 3ds?) "There will be no announcement" and then the hardware literally gets revealed like a week later.

This is different from a definitive, "we will not be releasing new hardware this fiscal year" as going against that would really upset Nintendo's investors.

I dunno. I mean I do think it will most likely be released with Botw 2 early next year, but how and what nintendo phrases their words, do matter.

This isn't even from Nintendo it seems. It seems like it was a giant misunderstanding based on a poor translation.

We haven't had enough of those lately.

 

[Look over there]

If it's as low as 8gb it won't be from Micron. 16gb is their lowest on lppdr5.

Edit: Correction - https://www.micron.com/about/newsroom/media-relations/media-kits/lpddr5-media-kit

Right, so the product flyer lists LPDDR5 densities as 16 to 128 Gigabit (or, 2 to 16 Gigabyte)
Then, when you look at the part catalog, all of the 16 Gb/2 GB modules are either 16 bit or 32 bit.

Since we have a fairly reasonable inference that the memory bus width of Drake will be 128 bit, that's our target to hit. From Micron's catalog, the lowest possible way to hit that with modules in Production would be a pair of 48 Gb/6 GB modules for a total of 12 GB. And if we look at parts in Sampling, it would be a pair of 32 Gb/4 GB modules (or, four 16 Gb/2 GB modules) for a total of 8 GB.
Samsung's LPDDR5 catalog seems... absurdly out of date. Still just that one 64 bit 6 GB module at 5500 MT/s. As for SK Hynix... are they seriously almost all 16 bit (save for that one... 8 bit part in the PC section?)? Or maybe I just don't know how to read SK Hynix's catalog.

Of course, I still have the thought that Drake-based Switches will outlast standard LPDDR5 (like how regular Switch outlasted standard LPDDR4). Even if Drake itself doesn't launch with 5X, then a later revised-Drake likely would. And like with the Erista->Mariko transition, I'm not expecting a change in quantity of RAM (that is, I expect it to be 'silent'). Ergo, I'm keeping tabs on 5X.

 

[Serif]

Yeah this would make all of those articles and youtube reactions look kinda silly.

Yup. Reading the translations in the last few posts, it's genuinely funny that Nikkei's statement seems to be to just be a reiteration of the status quo. "There also is no announcement of new hardware."

Before the translations I assumed it was Nikkei deducing "Because of component shortages, do not expect new hardware to be released this fiscal year." At least that establishes expectations. But they're not even saying that? Lol.

 

[manfromthedam]

I just want a meaty September Direct with some good f***** food. zelda come on

September has an outside chance for hardware announcements but I'm kinda letting that go - Nate ruling out 2022 is the only sign I take seriously. The Nikkei stuff was obviously overblown from the get-go.

 

[LiC]

Perhaps it's just semantics here. Didn't the same thing happen with New 3ds (or was it just 3ds?) "There will be no announcement" and then the hardware literally gets revealed like a week later.

This is different from a definitive, "we will not be releasing new hardware this fiscal year" as going against that would really upset Nintendo's investors.

I dunno. I mean I do think it will most likely be released with Botw 2 early next year, but how and what nintendo phrases their words, do matter.

Regardless of the tense issue, there's no indication that that part of the article is from Furukawa. You can see the Japanese text in fwd-bwd's post.

 

[Dakhil]

Doesn't Drake support 2x NVDLA v2 though, because that it's only supported on the AGX models, and not NX.

I don't think there has been any mention whether or not NVDLAs (NVIDIA Deep Learning Accelerators) are present in Drake.

And 2x NVDLAs are present in the Jetson Orin NX 16 GB module.

 

[Hermii]

I don't think there has been any mention whether or not NVDLAs (NVIDIA Deep Learning Accelerators) are present in Drake.

And 2x NVDLAs are present in the Jetson Orin NX 16 GB module.

This is a separate dedicated hardware than tensor cores?

I assume it won’t be supported if it’s not in desktop ampere.

 

[Dakhil]

This is a separate dedicated hardware than tensor cores?

Going by an analysis of the die shot provided by Nvidia of Jetson AGX Xavier, yes.

 

[daase ko]

So things got scary with that article but in the end I can get back expecting a march release with zelda

 

[Skittzo]

So things got scary with that article but in the end I can get back expecting a march release with zelda Metroid Prime 4k

FTFY.

But really, it's August, Metroid Prime's anniversary month, and the game remains unannounced. Announcing them both together would do wonders for the game's sales.

 

[Serif]

I prefer a world where Switch Ultra comes before BotW2, so I have time to (a) actually obtain the console and (b) enjoy the BotW1 4K patch before diving into the game.

I really think they could market a 4K patch for the game as a part of the sequel's marketing cycle and that alone would push units.

 

[manfromthedam]

I really think they could market a 4K patch for the game as a part of the sequel's marketing cycle and that alone would push units.

That'd be a smart move for sure! Honestly, at this point I'm not even sure they'll show Zelda in September though. I'd say they'd kinda have to but... yeah - I've been saying that for at least a year at this point xD .

 

[LiC]

FTFY.

But really, it's August, Metroid Prime's anniversary month, and the game remains unannounced. Announcing them both together would do wonders for the game's sales.

Maybe I'm misunderstanding, but Metroid Prime's release date was November 18, 2002.

 

[Hermii]

Going by an analysis of the die shot provided by Nvidia of Jetson AGX Xavier, yes.

Is there any reason Nintendo would not cut it?

 

[ReddDreadtheLead]

Right, so the product flyer lists LPDDR5 densities as 16 to 128 Gigabit (or, 2 to 16 Gigabyte)
Then, when you look at the part catalog, all of the 16 Gb/2 GB modules are either 16 bit or 32 bit.

Since we have a fairly reasonable inference that the memory bus width of Drake will be 128 bit, that's our target to hit. From Micron's catalog, the lowest possible way to hit that with modules in Production would be a pair of 48 Gb/6 GB modules for a total of 12 GB. And if we look at parts in Sampling, it would be a pair of 32 Gb/4 GB modules (or, four 16 Gb/2 GB modules) for a total of 8 GB.
Samsung's LPDDR5 catalog seems... absurdly out of date. Still just that one 64 bit 6 GB module at 5500 MT/s. As for SK Hynix... are they seriously almost all 16 bit (save for that one... 8 bit part in the PC section?)? Or maybe I just don't know how to read SK Hynix's catalog.

Of course, I still have the thought that Drake-based Switches will outlast standard LPDDR5 (like how regular Switch outlasted standard LPDDR4). Even if Drake itself doesn't launch with 5X, then a later revised-Drake likely would. And like with the Erista->Mariko transition, I'm not expecting a change in quantity of RAM (that is, I expect it to be 'silent'). Ergo, I'm keeping tabs on 5X.

I’m unsure if they’ll even swap to the 5X later per se, as it’s more of an alteration to the silicon itself to change from the 5 memory controller to the updated 5X memory controller.

Mariko for example was more than just a die shrink itself, changing the 4 memory controller for 4X.

Though that should be a pretty trivial change.

If my understanding of it is in fact correct. Like how the 6800U doesn’t have the memory controllers for 5X but for LPDDR5, a 5X wouldn’t work and a refresh that addressed the memory controller to work for 5X (ie the 5X memory controller inside the chip). But this in turn doesn’t let the chip use LPDDR5 memory.

And since the Tegra is a doc that incorporates all the CPU, GPU, Northbridge, Southbridge and Memory Controller in one, the chance of them “switching” is really more of a “will they make a new SoC that just has this Memory controller?”


I wouldn’t rule out the possibility of 5X in the model, just have it as low chance. I’m still banking on 5 though.

But in an ideal world, 12GB 5X memory for 137GB/s memory bandwidth would be good enough I think. Would be close to the PS4 who managed fine and that’s woefully inferior to the later uArch types.

It's not automatic, that mode clocks the GPU down to 10% of docked speed, which would obviously be catastrophic if devs couldn't choose when to use it.

Ah, ok thank you. I read somewhere that other games were taking advantage of this for boosted loading.

 

[Skittzo]

Maybe I'm misunderstanding, but Metroid Prime's release date was November 18, 2002.

Whoops, I meant original Metroid. I added the prime by habit.

 

[Stinky Horse]

This is a separate dedicated hardware than tensor cores?

I assume it won’t be supported if it’s not in desktop ampere.

Yeah, they make up for like 1/3rd of the quoted TOPs figures too.
Per clock, 2xDLA are equivalent to 8 Orin SMs worth of tensor cores, in theoretical TOPs at least, I have no idea about implementation.

 

[Stinky Horse]

Ah, ok thank you. I read somewhere that other games were taking advantage of this for boosted loading.

Yep, I guess it is available for anyone to patch in now but it reportedly came about from a request of the Dragon Quest XI team originally.

 

[Shiba]

One of the biggest things I hope Nintendo improves on is the overall online experience. Revamp NSO and make it a lot better in its offerings/perks and don’t rely on peer to peer for all of their multiplayer games(except for Smash which needs rollback). With how important online is it’s definitely one of the pillars they need to improve with their next system besides ergonomics and hardware quality.

 

[ShadowFox08]

This isn't even from Nintendo it seems. It seems like it was a giant misunderstanding based on a poor translation.

We haven't had enough of those lately.

Right..I'm trying to make sense of the quotation marks. Thought it was president of Nintendo that was being quoted.

Right, so the product flyer lists LPDDR5 densities as 16 to 128 Gigabit (or, 2 to 16 Gigabyte)
Then, when you look at the part catalog, all of the 16 Gb/2 GB modules are either 16 bit or 32 bit.

Since we have a fairly reasonable inference that the memory bus width of Drake will be 128 bit, that's our target to hit. From Micron's catalog, the lowest possible way to hit that with modules in Production would be a pair of 48 Gb/6 GB modules for a total of 12 GB. And if we look at parts in Sampling, it would be a pair of 32 Gb/4 GB modules (or, four 16 Gb/2 GB modules) for a total of 8 GB.
Samsung's LPDDR5 catalog seems... absurdly out of date. Still just that one 64 bit 6 GB module at 5500 MT/s. As for SK Hynix... are they seriously almost all 16 bit (save for that one... 8 bit part in the PC section?)? Or maybe I just don't know how to read SK Hynix's catalog.

Of course, I still have the thought that Drake-based Switches will outlast standard LPDDR5 (like how regular Switch outlasted standard LPDDR4). Even if Drake itself doesn't launch with 5X, then a later revised-Drake likely would. And like with the Erista->Mariko transition, I'm not expecting a change in quantity of RAM (that is, I expect it to be 'silent'). Ergo, I'm keeping tabs on 5X.

is Switch's RAM from micron?

yeah I do feel like a repeat of lpddr5x on the revision could potentially happen for power draw saving purposes like for Mariko as well.

I don't think there has been any mention whether or not NVDLAs (NVIDIA Deep Learning Accelerators) are present in Drake.

And 2x NVDLAs are present in the Jetson Orin NX 16 GB module.

Ah you're right. I misread!

 

[Skittzo]

Right..I'm trying to make sense of the quotation marks. Thought it was president of Nintendo that was being quoted.

Furukawa only said it was going to be difficult to sell the remaining 18m units needed to meet their forecast. That's all he was quoted on there.

Nikkei seemingly added the point about new hardware not yet being announced simply as another reason why it may be hard to hit that goal.

 

[ReddDreadtheLead]

I’d hope that if they reserve 1-1.5GB of RAM for the OS, they’d manage to make the UX more… pretty.


Even the Wii U or 3DS with less memory managed to have a nicer UX but were slow for those systems.

Switch 2 with 8-12GB of RAM can have 1-1.5GB and still have a pretty snappy and interesting UX, no?

 

[ShadowFox08]

I’d hope that if they reserve 1-1.5GB of RAM for the OS, they’d manage to make the UX more… pretty.


Even the Wii U or 3DS with less memory managed to have a nicer UX.

Switch 2 with 8-12GB of RAM can have 1-1.5GB and still have a pretty snappy by interesting UX, no?

Yeah a meatier OS along with more playback for videos would be nice, as long as doesn't sacrifice speed to boot up.

Not directly related to RAM, but could they bring an NVMe for storage (the NX Orion models are compatible with an external one, but weirdly enough don't seem to come with storage.. for saving them cost I bet )? Its better/faster than eMMc, but more power consuming, right? I really want a fast storage.. Being to able to suspend even two games at once would me amazing.

 

[ReddDreadtheLead]

Yeah a meatier OS along with more playback for videos would be nice, as long as doesn't sacrifice speed to boot up.

Not directly related to RAM, but could they bring an NVMe for storage (the NX Orion models are compatible with an external one, but weirdly enough don't seem to come with storage.. for saving them cost I bet )? Its better/faster than eMMc, but more power consuming, right? I really want a fast storage.. Being to able to suspend even two games at once would me amazing.

We’ve been entertaining the idea of UFS which is the pretty fast.

But still expect eMMC.

 

[Hermii]

Yeah a meatier OS along with more playback for videos would be nice, as long as doesn't sacrifice speed to boot up.

Not directly related to RAM, but could they bring an NVMe for storage (the NX Orion models are compatible with an external one, but weirdly enough don't seem to come with storage.. for saving them cost I bet )? Its better/faster than eMMc, but more power consuming, right? I really want a fast storage.. Being to able to suspend even two games at once would me amazing.

I think UFS is ideal, as it is created with low power consumption in mind.. UFS 3 would be the best possible scenario.

 

[Josh5890]

kinda worried about the whole China Taiwan USA relationship too..

You mean that missile stunt that China pulled off that has approx. 150 vessels in the area on alert at the moment? Yea, I'm sure that won't be a big deal at all.......

 

[Serif]

That'd be a smart move for sure! Honestly, at this point I'm not even sure they'll show Zelda in September though. I'd say they'd kinda have to but... yeah - I've been saying that for at least a year at this point xD .

I was in the same line of thinking, "there will be a June direct, they got the delay announcement out of the way and even showed new footage there, they must be gearing up for a new trailer", but 'twas not meant to be.

Honestly the only thing I have some confidence in is that they'll give us the title of the game this year. I empathize with the developers working on this for years and years through a pandemic, I want to at least call their work by its proper name, considering it's slated for less than half a year away.

Prior to BotW's release there was some cross-promotion with TPHD, with the Wolf Link amiibo. The best game for cross-promoting BotW2 would be its literal prequel that millions of Switch owners already own, that they still sell at full price, and would be a perfect showcase for a DLSS patch.

Lot of folks say BotW2 + Switch Ultra would be good synergy, and while that's true there's no guarantee their timelines would sync up perfectly - so fuck it, use the same game the Switch launched with, alongside Prime HD, Xenoblade 3, and Pokemon.

Show a trailer with BotW's graphics patch that seamlessly transitions into new BotW2 footage running on Switch Ultra. Like this but better.

 

[manfromthedam]

Honestly the only thing I have some confidence in is that they'll give us the title of the game this year. I empathize with the developers working on this for years and years through a pandemic

Yeah, I really feel for them (and all devs having to squeeze any workload through the pandemic).

Also, I hadn't seen that GIF before, that must've been crazy hype! BotW is such a crazy technological and artistic leap from the Wii games.

 

[Dakhil]

Its better/faster than eMMc, but more power consuming, right?

Correct, comparing my rough active power calculations for the eMMC 5.1 chips used on the Nintendo Switch family of systems, with the active power measurements of the Toshiba (Kioxia) BG4 M.2 2230.

Fortunately, UFS exists, with performance not too far off from NVMe SSDs (e.g. up to 2.1 GB/s in sequential read speeds & 1.2 GB/s in sequential write speeds for Samsung's eUFS 3.1 chips vs up to 2.4 GB/s in sequential read speeds and 1.95 GB/s for the sequential write speeds for Western Digital's 1 TB SN530 NVMe SSD [Xbox Series X's NVMe SSD]), but have much lower power consumption compared to NVMe SSDs (e.g. Toshiba's (Kioxia's) THGJFAT0T44BAIL (128 GB UFS 3.1) chip seems to have an active power draw of roughly 793.8 mW (1.26 V [VCCQ] * 630 mA [ICCQ]), in comparison to Western Digital's 1 TB SN530 NVMe SSD, which seems to have an active power draw of 3500 mW for sustained sequential read, and 2900 mW for sustained sequential write).

 

[fwd-bwd]

Heh, don`t go to far in the other direction.
It`s in a paragraph where the writer is implying that because there is no announcement of a new system, Nintendo making their sales targets for hardware is going to be difficult.

It is almost a throw-away line, I would not actually read anything into it either way regarding some insider knowledge from Nikkei.

Thanks for adding to the discussion. As you pointed out, that sentence itself didn’t make clear of its tense, therefore we have to rely on the context to translate it. Since it’s sandwiched between the Furukawa quote about what could/would have been, and the Nikkei’s speculation of sales shortfall for the next 9 months, one may lump the “no new hardware” sentence with the former (has been no announcement) or the latter (will be no announcement).

I chose to translate it using the future tense, because the core message of the paragraph is a warning to the investors about the full FY performance. But it really can be read both ways. As I suggested in my previous post, fundamentally this isn’t a translation issue but bad writing/editing to begin with. It left the readers—us and Mochizuki too—befuddled.

 

[Serif]

I hadn't seen that GIF before, that must've been crazy hype! BotW is such a crazy technological and artistic leap from the Wii games.

It's from the November 2015 direct, I was in my senior year of high school () and it blew my mind to see Zelda Wii U again and so suddenly. I love that they were flexing with actual gameplay too.

TPHD tided us over until BotW. They should save WW and TP remasters after BotW2 is out, and use BotW 4K as this year's (free) remaster. One can dream.

 

[Nintendojitsu]

Right..I'm trying to make sense of the quotation marks. Thought it was president of Nintendo that was being quoted.

is Switch's RAM from micron?

yeah I do feel like a repeat of lpddr5x on the revision could potentially happen for power draw saving purposes like for Mariko as well.

Ah you're right. I misread!

Yes. That's why I linked to it initially.

 

[Dakhil]

is Switch's RAM from micron?

Yes, for the OLED model.

 

[Look over there]

I’m unsure if they’ll even swap to the 5X later per se, as it’s more of an alteration to the silicon itself to change from the 5 memory controller to the updated 5X memory controller.

Mariko for example was more than just a die shrink itself, changing the 4 memory controller for 4X.

Though that should be a pretty trivial change.

If my understanding of it is in fact correct. Like how the 6800U doesn’t have the memory controllers for 5X but for LPDDR5, a 5X wouldn’t work and a refresh that addressed the memory controller to work for 5X (ie the 5X memory controller inside the chip). But this in turn doesn’t let the chip use LPDDR5 memory.

And since the Tegra is a doc that incorporates all the CPU, GPU, Northbridge, Southbridge and Memory Controller in one, the chance of them “switching” is really more of a “will they make a new SoC that just has this Memory controller?”


I wouldn’t rule out the possibility of 5X in the model, just have it as low chance. I’m still banking on 5 though.

But in an ideal world, 12GB 5X memory for 137GB/s memory bandwidth would be good enough I think. Would be close to the PS4 who managed fine and that’s woefully inferior to the later uArch types.

Ah, ok thank you. I read somewhere that other games were taking advantage of this for boosted loading.

Thing is, while it is work to update the memory controller to support 5X, I think that work is unavoidable in the long term. I think that within the lifetime of Drake/revised-Drake Switches, regular LPDDR5 will be phased out. Ergo, for sourcing reasons, I think that 5X support is inevitable.

Right..I'm trying to make sense of the quotation marks. Thought it was president of Nintendo that was being quoted.

is Switch's RAM from micron?

yeah I do feel like a repeat of lpddr5x on the revision could potentially happen for power draw saving purposes like for Mariko as well.

Ah you're right. I misread!

According to ifixit's teardowns, while the original used RAM from Samsung, the OLED uses Micron's.

Correct, comparing my rough active power calculations for the eMMC 5.1 chips used on the Nintendo Switch family of systems, with the active power measurements of the Toshiba (Kioxia) BG4 M.2 2230.

Fortunately, UFS exists, with performance not too far off from NVMe SSDs (e.g. up to 2.1 GB/s in sequential read speeds & 1.2 GB/s in sequential write speeds for Samsung's eUFS 3.1 chips vs up to 2.4 GB/s in sequential read speeds and 1.95 GB/s for the sequential write speeds for Western Digital's 1 TB SN530 NVMe SSD [Xbox Series X's NVMe SSD]), but have much lower power consumption compared to NVMe SSDs (e.g. Toshiba's (Kioxia's) THGJFAT0T44BAIL (128 GB UFS 3.1) chip seems to have an active power draw of roughly 793.8 mW (1.26 V [VCCQ] * 630 mA [ICCQ]), in comparison to Western Digital's 1 TB SN530 NVMe SSD, which seems to have an active power draw of 3500 mW for sustained sequential read, and 2900 mW for sustained sequential write).

Oh neat, thanks for finding those pdf files.
I do have to correct the power draw for Toshiba's UFS 3.1 chip. I think that you need to add both VCC and VCCQ. Looking at the diagram, I think that VCC is for the NAND while VCCQ is for everything else.
So, for 2 lanes, that's 3.6V*365mA + 1.26V*630mA = 1,314 + 793.8 = 2,107.8 mW. That seems like a lot, right? However, these are maximum values. There's a note specifying that this for 85 Celsius. There is no place in the Switch that should hit anywhere near 85C. According to Digital Foundry's review of the OG model, the hottest spot is around the air vent and peaked at about 52C. And even this worst case scenario which isn't realistic whatsoever for the Switch is still more preferable than the NVMe drive. And the NVMe drive's numbers are given at 25C. At that temperature, eUFS is way ahead in efficiency. Annoyingly, typical operating current still isn't given. But at the least, one can look at the values for sleep mode and notice the drastic difference between the voltages and currents given for 25C and for 85C. And the PDF says that this particular chip is rated for 1,850 MB/s sequential read when using both lanes.
Anyway, for 1 lane, the worst case is 3.6V*225mA + 1.26V*450mA = 810 + 567 = 1,377mW. Sequential read speed-wise, single lane's rated for 1,060 MB/s.

Edit: Oh yea, while looking at ifixit's teardown of the OG Switch: the fan's rated for 5V and 0.33A? It's potentially up to 1.65W on its own?
Edit 2: Wow, did I flub that single lane max power draw.

 

[ShadowFox08]

We’ve been entertaining the idea of UFS which is the pretty fast.

But still expect eMMC.

I remember people talking about ufs 3.0. Didn't cross my mind when I brought that question. Was just thinking about what the orion's offered between the AGX and NX models. Good to know speeds are comparable to NVMe.

Are there much faster eMMC storage now though?

 

[ILikeFeet]

Correct, comparing my rough active power calculations for the eMMC 5.1 chips used on the Nintendo Switch family of systems, with the active power measurements of the Toshiba (Kioxia) BG4 M.2 2230.

Fortunately, UFS exists, with performance not too far off from NVMe SSDs (e.g. up to 2.1 GB/s in sequential read speeds & 1.2 GB/s in sequential write speeds for Samsung's eUFS 3.1 chips vs up to 2.4 GB/s in sequential read speeds and 1.95 GB/s for the sequential write speeds for Western Digital's 1 TB SN530 NVMe SSD [Xbox Series X's NVMe SSD]), but have much lower power consumption compared to NVMe SSDs (e.g. Toshiba's (Kioxia's) THGJFAT0T44BAIL (128 GB UFS 3.1) chip seems to have an active power draw of roughly 793.8 mW (1.26 V [VCCQ] * 630 mA [ICCQ]), in comparison to Western Digital's 1 TB SN530 NVMe SSD, which seems to have an active power draw of 3500 mW for sustained sequential read, and 2900 mW for sustained sequential write).

do you know what emmc's power measure is in a like for like scenario? I don't think I've seen a rating for a sustained sequential load like I have for m.2 drives

 

[Look over there]

eMMC specifically? Not by much really; eMMC peaked with version 5.1 in 2014 with... ostensibly a max of 330 MB/s sequential read, 150 MB/s sequential write, 14,500 random read IOPS, and 14,400 random write IOPS*. Or so a Samsung brochure says, anyway. UFS is effectively the successor and is what's actively developed.

*if you've seen other brochures/datasheets, this number for random write is not believable. For that matter, the 13,000 random write IOPS for 5.0? Hell no, the other sources I've seen point 5.0's random write IOPs to be more in the 3 digit to low 4 digit area; forget about 5 digits.

 

[Skittzo]

Thanks for adding to the discussion. As you pointed out, that sentence itself didn’t make clear of its tense, therefore we have to rely on the context to translate it. Since it’s sandwiched between the Furukawa quote about what could/would have been, and the Nikkei’s speculation of sales shortfall for the next 9 months, one may lump the “no new hardware” sentence with the former (has been no announcement) or the latter (will be no announcement).

I chose to translate it using the future tense, because the core message of the paragraph is a warning to the investors about the full FY performance. But it really can be read both ways. As I suggested in my previous post, fundamentally this isn’t a translation issue but bad writing/editing to begin with. It left the readers—us and Mochizuki too—befuddled.

The part of the initial interpretation that never made sense to me was the "heavily featured" or whatever that was supposed to be. The past interpretation makes that a lot more logical, basically saying the new hardware would be heavily featured and help them achieve the remaining 18M units but so far this FY it hasn't been announced.

 

[fwd-bwd]

The part of the initial interpretation that never made sense to me was the "heavily featured" or whatever that was supposed to be. The past interpretation makes that a lot more logical, basically saying the new hardware would be heavily featured and help them achieve the remaining 18M units but so far this FY it hasn't been announced.

I believe the phrase that you referred to is "目玉". It literally means things that attract eyeballs. Based on the context (and how poetic you want to be), it may be translated to centerpiece, flagship model, featured product, heavy hitter, crown jewel, and so on. In this particular case, however, the item has not been revealed or even acknowledged officially, therefore I went with "highly anticipated new model" instead, considering that Nikkei's own past reporting contributed to that anticipation. Regardless of the translation, the phrase doesn't seem to carry any deeper meaning or hints, but rather a literary flair (which tends to trip up machine translations).

 

[Skittzo]

I believe the phrase that you referred to is "目玉". It literally means things that attract eyeballs. Based on the context (and how poetic you want to be), it may be translated to centerpiece, flagship model, featured product, heavy hitter, crown jewel, and so on. In this particular case, however, the item has not been revealed or even acknowledged officially, therefore I went with "highly anticipated new model" instead, considering that Nikkei's own past reporting contributed to that anticipation. Regardless of the translation, the phrase doesn't seem to carry any deeper meaning or hints, but rather a literary flair (which tends to trip up machine translations).

What I mean is that it seems like that phrase is entirely unnecessary and even potentially confusing in the context of "Nintendo will struggle to meet their forecast and the new model will not be released this fiscal year" but in the context of "Nintendo will struggle to meet demand and they have not announced a new model yet this fiscal year" it does make sense and adds further to the overall point, which is that they may see lower demand without said model.

The sentence as a whole reads much more logically imo if it's talking in the past tense there.

 

[fwd-bwd]

What I mean is that it seems like that phrase is entirely unnecessary and even potentially confusing in the context of "Nintendo will struggle to meet their forecast and the new model will not be released this fiscal year" but in the context of "Nintendo will struggle to meet demand and they have not announced a new model yet this fiscal year" it does make sense and adds further to the overall point, which is that they may see lower demand without said model.

The sentence as a whole reads much more logically imo if it's talking in the past tense there.

Gotcha, that makes sense. But gosh, how many people does it take to change a light bulb decipher a Nikkei sentence? I maintain that it was not a good piece of journalistic writing.

 

[LiC]

Thanks for adding to the discussion. As you pointed out, that sentence itself didn’t make clear of its tense, therefore we have to rely on the context to translate it. Since it’s sandwiched between the Furukawa quote about what could/would have been, and the Nikkei’s speculation of sales shortfall for the next 9 months, one may lump the “no new hardware” sentence with the former (has been no announcement) or the latter (will be no announcement).

I chose to translate it using the future tense, because the core message of the paragraph is a warning to the investors about the full FY performance. But it really can be read both ways. As I suggested in my previous post, fundamentally this isn’t a translation issue but bad writing/editing to begin with. It left the readers—us and Mochizuki too—befuddled.

Especially since the context is an article about how the signs are pointing for the rest of Nintendo's FY, I think throwing in a offhanded reference to unannounced plans (and it's new information not reported anywhere else), without so much as a "sources say," is implausible. There's a reason why articles with such reporting use familiar phrases like "according to people familiar with the matter" so the reader can distinguish what is sourced reporting and what's not. But if all you're doing is noting that nothing has been publicly announced -- something that is relevant to the question of how Nintendo's FY will play out, basically stating that there's no sign of this much-discussed thing that could make a big difference -- that's fine to just throw in there in the voice of the article.

I also think the idea of claiming a quite long (239 days! 8 months!) period where Nintendo isn't going to release hardware would be a weird thing to do as sourced reporting, context of this article aside. Surely the only way you can claim Nintendo isn't releasing hardware such a long ways in advance is if you have information that Nintendo is releasing hardware after March? And if that were the case I think Nikkei would have reported it that way, and probably in its own article.

 

[NintenDuvo]

Before i head to bed, has there been any new tidbits from our Funcles from the factories?

 

[CypressFX]

@LiC @Skittzo @fwd-bwd

So Am I getting you guys right? The Nikkei article is way to unclear - meaning in can mean anything? Especially, that neither Furukawa denied a model nor did Nikkei say that its not coming this FY?

 

[fwd-bwd]

@LiC @Skittzo @fwd-bwd

So Am I getting you guys right? The Nikkei article is way to unclear - meaning in can mean anything? Especially, that neither Furukawa denied a model nor did Nikkei say that its not coming this FY?

We don't have a consensus on the tense of that "no announcement" statement; it could mean that there hasn't been one, or there won't be one. No apparent attribution either. It's very unclear, therefore even Mochizuki, a native speaker, asked for clarification.

Before i head to bed, has there been any new tidbits from our Funcles from the factories?

Nothing relevant to a new hardware from them as of late.

 

[Thugstas]

We don't have a consensus on the tense of that "no announcement" statement; it could mean that there hasn't been one, or there won't be one. No apparent attribution either. It's very unclear, therefore even Mochizuki, a native speaker, asked for clarification.


Nothing relevant to a new hardware from them as of late.

September will be a good crossing point if we will get something before the end of the FY.
Anyway, 2023 will be the max.

 

[CypressFX]

September will be a good crossing point if we will get something before the end of the FY.
Anyway, 2023 will be the max.

Why? DO you think they will announce it in September or something similiar if it launches by March?

 

[Thugstas]

Why? DO you think they will announce it in September or something similiar if it launches by March?

Just wait and stop being stressful… nobody died yet because of non switch announcement but lots of died yesterday everywhere ( Yemen … Palestine… thousands in Africa ..)
Calm down and patience means reward in some religion and/or philosophy