Hacking Gamecube Games on Nvidia Shield Are Emulated.

[Deathscreton]

Ssssh, you can't use logic on here, the hardware fanboys who're dissatisfied with the Switch being "weak" will only bitch about how it's not as powerful as the Shield ad nauseum. Clearly, these are serious hardware engineers here and know how a console should and shouldn't be according to their implacable tastes.

Uh, okay? I wasn't directing that to you, it was a general statement. If I was directing it to you, I would've quoted you. There are people on the internet who have been saying it was weak, not up to their standards, etc. Check IGN, YouTube, etc if you must.

By saying "here", everyone (I assume) who is contributing to this conversation is going to assume you're talking about this thread, this conversation because social ques. You then further concrete this by saying "hardware fanboys who are dissatisfied". Literally, how else am I supposed to interpret this considering we're talking about the hardware?

 

[the_randomizer]

By saying "here", everyone (I assume) who is contributing to this conversation is going to assume you're talking about this thread, this conversation because social ques. You then further concrete this by saying "hardware fanboys who are dissatisfied". Literally, how else am I supposed to interpret this considering we're talking about the hardware?

Did I stutter? I wasn't talking to you per se, there are people who bitch about the Switch not being a home console or that "if it was a home console it'd be more powerful blah blah blah". But yeah, I'm done though, I'm done. I wasn't attacking you at all, but thanks for assuming the worst.

 

[Deathscreton]

The reasons you're getting as to why it might not be worth looking into are about as good as you'll get. It's speculation as to why. However, to assume they're asinine/irrelevant is no less ignorant than to say "this is probably why". Nobody here can tell you why. The safe route is to assume that the speeds we've got are the best we'll get without repercussions. The heat/temp test you're referring to doesn't account for potential long term damage.

In regards to the long term damage, I suppose that is fair. However, I think Nintendo would have done these tests while keeping in mind the damage that could have come from heats of upwards to 140 degrees F. If they didn't we have much bigger problems on our hands as their R&D is probably shit.

As for the reasons, you could say that in either direction. I'm just asking for relevant, supportive claims for the argument that they underclocked it for those reasons, just like I used claims on my part as for why they underclocked. And what part of pulling apart a system and learning the internals is "safe" in regards to the health of the system? Literally every system we have a board for has had to undergo dangerous testing and possibly destructive reverse engineering. I'm not here for the safe route. I'm here to talk about the potential the system has for GC emulation and clocking the internals to their factory speeds, along with all of the possible downsides of doing so. I know no one has concrete answers and everyone is just throwing around guesses, because at this point, that's all we can do until the Switch is cracked open completely.

 

[Tom Bombadildo]

I never said anything about going past the limits set by the OS. In fact, if you read the rest of my posts, I clearly say that we're technically not overclocking anything with the Vita and PSP by pushing them to their factory max. I would never suggest doing that because they aren't designed to cool past that point. Same with the Switch I would assume. Using the term overclocking is a poor choice of words that many misinterpret, case and point here.

My post was being written before the edits, so they weren't there. Irregardless, you're still mistaken about the Vita, the factory clock for the ARM SoC used is 2GHz, the Vita is severely underclocked to deal with battery and thermal limits. The same is said here for the Switch, it's underclocked to deal with the battery and thermal limits.

As for the temperature, I don't know where you've seen it get to 60c in a dock that wasn't the result of a defect and not proper care or a properly working unit. Can you share a link? does a quick temp check at 4 hours in BOTW and it shows max 107 degrees F which is only 40c. I feel there's plenty more room here. Provided, this is only one video. I'd love more information on this. As for the bending, a lot of the reports have said that the switches have come from the factory bent. Most just have not noticed this.

Measuring the case with a thermal gun isn't accurate enough, especially with the way he's measuring the temps by just measuring the top of the Switch. The below video, and this link if you'd like to read some more info, uses the following method to check actual component temperatures:

Our tests apply self-adhesive, 1/100th-inch thick (read: laser thin, does not cause "air gaps") K-type thermocouples directly to the components under test. The thermocouples used are flat and are self-adhesive (from Omega), as recommended by thermal engineers in the industry.

K-type thermocouples have a known range of approximately 2.2C. We calibrated our thermocouples by providing them an "ice bath," then providing them a boiling water bath. This provided us the information required to understand and adjust results appropriately.

The video is from the same article, but basically gives a bit of information on the SoC and such as well (quite an interesting read, too).

 

[Kioku]

In regards to the long term damage, I suppose that is fair. However, I think Nintendo would have done these tests while keeping in mind the damage that could have come from heats of upwards to 140 degrees F. If they didn't we have much bigger problems on our hands as their R&D is probably shit.

As for the reasons, you could say that in either direction. I'm just asking for relevant, supportive claims for the argument that they underclocked it for those reasons, just like I used claims on my part as for why they underclocked. And what part of pulling apart a system and learning the internals is "safe" in regards to the health of the system? Literally every system we have a board for has had to undergo dangerous testing and possibly destructive reverse engineering. I'm not here for the safe route. I'm here to talk about the potential the system has for GC emulation and clocking the internals to their factory speeds, along with all of the possible downsides of doing so. I know no one has concrete answers and everyone is just throwing around guesses, because at this point, that's all we can do until the Switch is cracked open completely.

The CPU is clocked at half of what the Shield is. In all fairness, considering how small THAT system is? I question why, unknowingly. I do agree, Nintendo has to have tested extreme conditions extensively. Which is why I believe that what we have now is as good as it gets without repercussions. That's speculation. I AM interested to see how far we can push the system. Not even specifically for GameCube emulation. It goes without saying, that "at your own risk" comes to mind. We'll just have to wait and see.

 

[Deathscreton]

My post was being written before the edits, so they weren't there. Irregardless, you're still mistaken about the Vita, the factory clock for the ARM SoC used is 2GHz, the Vita is severely underclocked to deal with battery and thermal limits. The same is said here for the Switch, it's underclocked to deal with the battery and thermal limits.


Measuring the case with a thermal gun isn't accurate enough, especially with the way he's measuring the temps by just measuring the top of the Switch. The below video, and this link if you'd like to read some more info, uses the following method to check actual component temperatures:




The video is from the same article, but basically gives a bit of information on the SoC and such as well (quite an interesting read, too).

I do apologize for the edits. Trying to carry three conversations at once leads to double posts, and I hate double posts, so edit was the answer. I'll be sure to wait a longer period before responding next time, while trying to not get lost in the conversation. My bad. As for the Vita, I was wrong to say factory maxes. It didn't occur to me that they may be using chips that are severely underclocked, even for it's OS based limits. I'll keep this in mind next time. Though, I think it's fair to say that the Vita is in a different ball game considering that the Switch has active cooling, something the Vita can't benefit from.

I do believe that Nintendo probably DID push the Switch to what they thought might be practical for a normal consumer in terms of battery life and thermals when in undocked mode, but I don't think they wouldn't leave some headroom for developers when it comes to software in docked mode. This leads me to believe that the Switch still has much more room to expand in terms of how hard it can be pushed.

You weren't kidding though, that WAS super interesting. 60c is fairly high for a pkg reading, at least for one of this size. Though I don't see it as a theoretical limit for the device. I wonder how much of that heat is due to poor paste and paste application. Do you think we'd get better readings with some artic silver on that bad boy?

 

[Tom Bombadildo]

I do apologize for the edits. Trying to carry three conversations at once leads to double posts, and I hate double posts, so edit was the answer. I'll be sure to wait a longer period before responding next time, while trying to not get lost in the conversation. My bad. As for the Vita, I was wrong to say factory maxes. It didn't occur to me that they may be using chips that are severely underclocked, even for it's OS based limits. I'll keep this in mind next time. Though, I think it's fair to say that the Vita is in a different ball game considering that the Switch has active cooling, something the Vita can't benefit from.

I do believe that Nintendo probably DID push the Switch to what they thought might be practical for a normal consumer in terms of battery life and thermals when in undocked move, but I don't think they wouldn't leave some headroom for developers when it comes to software in docked mode. This leads me to believe that the Switch still has much more room to expand in terms of how hard it can be pushed.

You weren't kidding. That WAS super interesting. 60c is fairly high for a pkg reading, though I don't see it as a theoretical limit for the device. I wonder how much of that heat is due to poor paste and paste application. Do you think we'd get better readings with some artic silver on that bad boy?

Indeed, that tiny little fan the Switch has makes a ton of difference, temp wise, but unfortunately it's still not the superman of cooling, it has it's limits. It helps that the ARM SoC is a relatively low-power chip in general, but there's only so much you can do before you simply don't have enough cooling power to keep up with temperatures. I wouldn't say 60C is the hard limit (the Tegra X1 does have a max temp of 105C before it'll shut down, so it's got plenty of wiggle room ), but the OS is setup so it'll throttle at 60+C. Whether that's because Nintendo doesn't want the SoC to work itself too hard and lower the life expectancy of the chip or if it's because of some other asinine reason, that's simply the way it was designed. Whether Nintendo would consider removing it is completely up to them, but I don't see that happening IMO just based on how "safe" Nintendo generally goes with their hardware designs.

And Nintendo did leave more room for devs in docked mode, that's why the GPU clocks up to 768mhz from 384Mhz in portable mode. See, here's the thing, gaming (when not hindered by OS overhead and such, like say on a game console ) doesn't really need all that much CPU power in general (obviously not counting those few genres with inherently high CPU usage, like RTS, that have tons of CPU-based calculations to perform). That's why people can still have competent gaming rigs combining some old 2nd gen Core i5/i7 with like a GTX 1xxx series card, most games aren't really CPU-heavy so it's not much of a bottleneck. I imagine the main reason the Switch is clocked down all the time, besides obvious battery/thermal issues, is because there's really no reason to have the Switch's CPU clocked up when the vast majority of software for the thing wouldn't take advantage of it. That's why the PS4 and Xboner can get away with using AMD's APUs (which have awful CPU components), it's not really "necessary" to stick a big ol' honking CPU in the thing when most software is designed to use GPUs.

While that's fine for just native Switch gaming, it's not for emulation. Emulation is very CPU heavy, especially when we start talking Sixth gen consoles and up, and I highly doubt Nintendo's R&D team made Gamecube emulation a high priority when they were duct taping the Switch together. Especially when they decided to opt for an ARM SoC with a relatively low IPC, something that's totally unsuited for GC emulation in general.

 

[sansnumen]

Do we really need another Xbox 360 RROD fiasco with the Switch? All the major components inside the Switch are BGA mounted. Over time, thermal stress can bend the PCB, crack the solder under the chips, or even worse desoldering chips from the board. Heat is the enemy of all electronics. The Switch CPU does not run at 1.9GHz for reasons, very likely thermal reasons.

 

[Deathscreton]

Do we really need another Xbox 360 RROD fiasco with the Switch? All the major components inside the Switch are BGA mounted. Over time, thermal stress can bend the PCB, crack the solder under the chips, or even worse desoldering chips from the board. Heat is the enemy of all electronics. The Switch CPU does not run at 1.9GHz for reasons, very likely thermal reasons.

Again, this is assuming that Nintendo didn't do extensive testing of their PCBs and plastics used to house the electronics. I really fail to see them not going to rigorous amounts of stress testing to ensure the components functioned properly well above whatever heat the Switch is going to put out. So unless we just all of a sudden lost faith in their R&D department, I don't think thermal problems are going to occur.

Btw, using the RROD issues as a point of reference isn't exactly strong considering MS has always been shit at their designs. It took them almost three iterations of the 360 to fix this problem. That says something about them, not the BGA mounts or their inherent defects. Cold solder can happen to any type of joint, by the way, not just BGA type mounts (I'm assuming cold solder/voiding is what you're referring to when you brought up the BGA, if not, I apologize. lmao).

EDIT: Do we know if the package is underfilled? I would assume with a device being mobile like this, they'd glue the damn thing to the PCB.

 

[DiscostewSM]

But the switch is weaker than the shield tv...the shield has eight cores, 4 clocked at 2.2Ghz and 4 clocked at 2.0Ghz, wherever the switch only has 4 clocked at 1.0Ghz, not to mention the underclocked tegra x1 gpu...

An gamecube virtual console could still be possible, but dont expect anything greatter than 720p on both docked and/or undocked mode, even more, i highly doubt undocked mode would be able to run gamecube games at 720p at all, considering it underclocks the gpu even more.

The Shield TV and Switch have the same SoC in them (Tegra X1), and both have 2 4-core CPUs in them. Switch only uses the A57s (A53s are disabled) while the Shield TV can use either the A57s or the A53s, but never both at the same time. The A53s are meant for when the system doesn't need to push as much with what it's doing. The A57s in the Shield only go up to 1.9Ghz with the A53s up to 1.3Ghz. Switch's A57s are tated at 1.0Ghz. The GPU in the Shield TV goes up to 1.0Ghz while it is 768Mhz in Switch's docked mode (307.2Mhz in undocked mode)

While one could claim the Shield TV is faster, that's not accounting for throttling. Supposedly people have tested the GPU and found that in the Shield TV under heavy load and thermals going up, it throttles down to 614Mhz, which is less than what the Switch's docked mode provides and is roughly twice what undocked mode provides (coincidence?). I haven't ready anything about throttling with the CPU, but I imagine than since it is on the same die as the GPU in its SoC package, it would also be throttled on Shield TV. Switch was confirmed not to throttle at all, but has in place a forced shutdown once it reaches a certain temperature, which takes a great deal of hindering the airflow (and even taking out the fan). Having throttling otherwise would be a destructive attribute to a closed-system, and I doubt Switch would have garnered nearly as many developers had it utilized throttling.

 

[sansnumen]

Again, this is assuming that Nintendo didn't do extensive testing of their PCBs and plastics used to house the electronics. I really fail to see them not going to rigorous amounts of stress testing to ensure the components functioned properly well above whatever heat the Switch is going to put out. So unless we just all of a sudden lost faith in their R&D department, I don't think thermal problems are going to occur.

Btw, using the RROD issues as a point of reference isn't exactly strong considering MS has always been shit at their designs. It took them almost three iterations of the 360 to fix this problem. That says something about them, not the BGA mounts or their inherent defects. Cold solder can happen to any type of joint, by the way, not just BGA type mounts (I'm assuming cold solder/voiding is what you're referring to when you brought up the BGA, if not, I apologize. lmao).

EDIT: Do we know if the package is underfilled? I would assume with a device being mobile like this, they'd glue the damn thing to the PCB.

From all the teardowns I've seen about the Switch it seems only the NAND chip is glued down. I do believe there was also thermal issues with the Iphone a few years ago where PCBs were bending. I would point out that past performance is no indicator of future performance. People are human and unexpected design issues pop up all the time. I'm sure the Switch was sufficiently tested, but it's hard to replicate real world conditions inside a lab. For example, here in my part of the country lead boils in the summer, and for an air cooled tablet like the Switch ambient temperature is very important when testing thermal stress.

 

[Tom Bombadildo]

The Shield TV and Switch have the same SoC in them (Tegra X1), and both have 2 4-core CPUs in them. Switch only uses the A57s (A53s are disabled) while the Shield TV can use either the A57s or the A53s, but never both at the same time. The A53s are meant for when the system doesn't need to push as much with what it's doing. The A57s in the Shield only go up to 1.9Ghz with the A53s up to 1.3Ghz. Switch's A57s are tated at 1.0Ghz. The GPU in the Shield TV goes up to 1.0Ghz while it is 768Mhz in Switch's docked mode (307.2Mhz in undocked mode)

While one could claim the Shield TV is faster, that's not accounting for throttling. Supposedly people have tested the GPU and found that in the Shield TV under heavy load and thermals going up, it throttles down to 614Mhz, which is less than what the Switch's docked mode provides and is roughly twice what undocked mode provides (coincidence?). I haven't ready anything about throttling with the CPU, but I imagine than since it is on the same die as the GPU in its SoC package, it would also be throttled on Shield TV. Switch was confirmed not to throttle at all, but has in place a forced shutdown once it reaches a certain temperature, which takes a great deal of hindering the airflow (and even taking out the fan). Having throttling otherwise would be a destructive attribute to a closed-system, and I doubt Switch would have garnered nearly as many developers had it utilized throttling.

The CPU doesn't throttle down in the Shield TV, it's just the GPU. Mainly because it's the GPU causing a majority of the heat. There were tests done with the CPU locked to 1ghz on the Shield TV, and the GPU would still throttle down after a certain point, with the CPU unaffected.

 

[AtlasFontaine]

The CPU doesn't throttle down in the Shield TV, it's just the GPU. Mainly because it's the GPU causing a majority of the heat. There were tests done with the CPU locked to 1ghz on the Shield TV, and the GPU would still throttle down after a certain point, with the CPU unaffected.

Do you mind providing source of those tests? I found this on Nvidia's forum (although it might be an outlier), since the Shield has pasive cooling and the CPU is close to the GPU (which throttles itself pretty hard) I think the CPU end up being affected too.

 

[Tom Bombadildo]

Do you mind providing source of those tests? I found this on Nvidia's forum (although it might be an outlier), since the Shield has pasive cooling and the GPU throttles itself that much I think the CPU might be affected too.

I don't think I have a link anymore, it was in a Neogaf thread IIRC but I can't seem to find it anymore (at least using their search ). If I find it I'll edit this post.

But that link you posted has inaccurate information regardless, the CPU in the Shield TV is clockable to 2GHz, and even if it were throttled it would never throttle all the way down to 300mhz. I think he's confusing CPU with GPU, since the GPU is clocked at 1Ghz and can throttle down a fair bit if the temps go too high.

 

[reminon]

My post was being written before the edits, so they weren't there. Irregardless, you're still mistaken about the Vita, the factory clock for the ARM SoC used is 2GHz, the Vita is severely underclocked to deal with battery and thermal limits. The same is said here for the Switch, it's underclocked to deal with the battery and thermal limits.


Measuring the case with a thermal gun isn't accurate enough, especially with the way he's measuring the temps by just measuring the top of the Switch. The below video, and this link if you'd like to read some more info, uses the following method to check actual component temperatures:




The video is from the same article, but basically gives a bit of information on the SoC and such as well (quite an interesting read, too).

Regardless.

 

[WhySayingMyName]

>what are your thoughts

wew here come the injections !

also how did they come up with up with a better emulator when dolphin has been in development forever

what do they have access to that makes them able to make a better emulator so quickly, yeah i know they designed the chip and it is probably optimized to only run on that one specific chip as opposed to dolphin which runs on any android with powerful enough hardware but is there anything else?

Dolphin has to be compatible with a lot of different configuration + approximative gamecube code, whereas Nvidia optimised the thing for the Tegra with access to the GC documentation.

Yet, there were rumors of a Dolphin developper involved in creating the Nvidia emulator, if that is the case then I highly think that Dolphin is gonna make some progress, especially for the Tegra hardware.

The odds of APKs being released into the wild and/or the emulator being reversed for injection are...?

Probably around 0, the APK has a licence checking system which is done every time you start the game, this is the only thing that render the apk useless to release yet. One team of motivated person can, crack the licence check

 

[vickdu31]

Probably around 0, the APK has a licence checking system which is done every time you start the game, this is the only thing that render the apk useless to release yet. One team of motivated person can, crack the licence check

This has been done so many times for android games with licence check.

 

[WhySayingMyName]

This has been done so many times for android games with licence check.

Then a team of dedicated person can make a licence check bypass, as well as a reverse enginered injector that will allow Shield owner to play Gamecube / Wii games

I do want the apk as well but I don't want to spend $250 on another Shield + import taxes + cryptic way to buy Zelda Twilight Princess alone without any Chinese aknowledges

Dolphin is cool, however the only game I played with minimal lags was Luigi's Mansion (with slowdowns here and there), this would be the ultimate alternative.
Let's dream and hope Nintendo and Nvidia announce Shield games for internationnal release

 

[KiiWii]

What about something like lucky patcher? Im sure I read there was a nonroot version out there, perhaps it would be useful to have the apk and try bruteforce or mod the apk.

Im kind of gutted they arent in wild yet.

 

[WhySayingMyName]

What about something like lucky patcher? Im sure I read there was a nonroot version out there, perhaps it would be useful to have the apk and try bruteforce or mod the apk.

Im kind of gutted they arent in wild yet.

Surprised as well that the apk wasn't leaked since it was dumped, apparently