Nintendo NX Will Use A Pascal Nvidia Tegra Chip

  • 167 results
  • 1
  • 2
  • 3
  • 4
Avatar image for techhog89
Techhog89

5430

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#151  Edited By Techhog89
Member since 2015 • 5430 Posts

@ronvalencia said:
@techhog89 said:
@ronvalencia said:

Take a look at the official statement about the 16FF+ node from TSMC

http://www.tsmc.com/english/dedicatedFoundry/technology/16nm.htm

TSMC’s 16FF+ (FinFET Plus) technology can provide above 65 percent higher speed, around 2 times the density, or 70 percent less power than its 28HPM technology. Comparing with 20 (nm) SoC technology, 16FF+ provides extra 40% higher speed and 60% power saving.

40 percent of Shield TV's 21 watts is 8.4 watts which is OK for 8 to 9 inch tablet and still too high for candy bar smart phones.

5 to 6 inch candy bar smart phones are about 1 to 5 watts.

Did you just say "candy bar" phones in 2016? Uh....

M-moving on from that, the 10-inch Pixel C tablet uses an X1 with the GPU clocked down to 850MHz, and only consumes 14W, screen and all. NX would have a much a smaller, lower-resolution screen, and overall could probably match it.

Note why I stated 8 to 9 inch tablet not 10.2 inch Pixel C tablet.

Change candybar with Phablet.

http://www.notebookcheck.net/Google-Pixel-C-Tablet-Review.158868.0.html

The price for Pixel C is $499 USD for 32 GB model https://store.google.com/product/pixel_c

40 percent faster speed yields around 56,931 points.

Nexus 9 with K1 is only 35 percent slower.

For comparison

40 percent of 14 watts is 5.6 watts... getting close.

My estimate for NX, it's far from XBO.

You're forgetting about the decrease in power consumption from going to a screen with a much smaller size and resolution. Factoring in that, it gets under 5W for sure. Also, from what I've heard, the X1's GPU uses just 1.5W when clocked at 500MHz, so a Pasal-based one should be 700MHz for 1.5W.

And yes, it's obviously going to be much slower than XB1 in even the best case.

Avatar image for techhog89
Techhog89

5430

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#152  Edited By Techhog89
Member since 2015 • 5430 Posts
@emgesp said:

@ronvalencia: I'd be surprised if the NX handheld is bigger than a Vita. I'm just assuming somewhere around 400 - 500 Gflops in the handheld. Will be a huge upgrade over the 3DS and a decent improvement over the Wii U.

Since you're assuming that there won't be difference between it being docked and on battery, that's extremely optimistic. Think 150-300 GFLOPS.

Avatar image for ronvalencia
ronvalencia

29612

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#153  Edited By ronvalencia
Member since 2008 • 29612 Posts

@techhog89 said:
@ronvalencia said:
@techhog89 said:
@ronvalencia said:

Take a look at the official statement about the 16FF+ node from TSMC

http://www.tsmc.com/english/dedicatedFoundry/technology/16nm.htm

TSMC’s 16FF+ (FinFET Plus) technology can provide above 65 percent higher speed, around 2 times the density, or 70 percent less power than its 28HPM technology. Comparing with 20 (nm) SoC technology, 16FF+ provides extra 40% higher speed and 60% power saving.

40 percent of Shield TV's 21 watts is 8.4 watts which is OK for 8 to 9 inch tablet and still too high for candy bar smart phones.

5 to 6 inch candy bar smart phones are about 1 to 5 watts.

Did you just say "candy bar" phones in 2016? Uh....

M-moving on from that, the 10-inch Pixel C tablet uses an X1 with the GPU clocked down to 850MHz, and only consumes 14W, screen and all. NX would have a much a smaller, lower-resolution screen, and overall could probably match it.

Note why I stated 8 to 9 inch tablet not 10.2 inch Pixel C tablet.

Change candybar with Phablet.

http://www.notebookcheck.net/Google-Pixel-C-Tablet-Review.158868.0.html

The price for Pixel C is $499 USD for 32 GB model https://store.google.com/product/pixel_c

40 percent faster speed yields around 56,931 points.

Nexus 9 with K1 is only 35 percent slower.

For comparison

40 percent of 14 watts is 5.6 watts... getting close.

My estimate for NX, it's far from XBO.

You're forgetting about the decrease in power consumption from going to a screen with a much smaller size and resolution. Factoring in that, it gets under 5W for sure. Also, from what I've heard, the X1's GPU uses just 1.5W when clocked at 500MHz, so a Pasal-based one should be 700MHz for 1.5W.

And yes, it's obviously going to be much slower than XB1 in even the best case.

1.5Watts at 500 Mhz? Where's it's 5 to 6 inch mobile phone solution?

Multiple mobile phone vendors' avoidance of TX1 indicates inferior lower power handling.

Avatar image for techhog89
Techhog89

5430

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#154 Techhog89
Member since 2015 • 5430 Posts

@ronvalencia said:
@techhog89 said:
@ronvalencia said:
@techhog89 said:
@ronvalencia said:

Take a look at the official statement about the 16FF+ node from TSMC

http://www.tsmc.com/english/dedicatedFoundry/technology/16nm.htm

TSMC’s 16FF+ (FinFET Plus) technology can provide above 65 percent higher speed, around 2 times the density, or 70 percent less power than its 28HPM technology. Comparing with 20 (nm) SoC technology, 16FF+ provides extra 40% higher speed and 60% power saving.

40 percent of Shield TV's 21 watts is 8.4 watts which is OK for 8 to 9 inch tablet and still too high for candy bar smart phones.

5 to 6 inch candy bar smart phones are about 1 to 5 watts.

Did you just say "candy bar" phones in 2016? Uh....

M-moving on from that, the 10-inch Pixel C tablet uses an X1 with the GPU clocked down to 850MHz, and only consumes 14W, screen and all. NX would have a much a smaller, lower-resolution screen, and overall could probably match it.

Note why I stated 8 to 9 inch tablet not 10.2 inch Pixel C tablet.

Change candybar with Phablet.

http://www.notebookcheck.net/Google-Pixel-C-Tablet-Review.158868.0.html

The price for Pixel C is $499 USD for 32 GB model https://store.google.com/product/pixel_c

40 percent faster speed yields around 56,931 points.

Nexus 9 with K1 is only 35 percent slower.

For comparison

40 percent of 14 watts is 5.6 watts... getting close.

My estimate for NX, it's far from XBO.

You're forgetting about the decrease in power consumption from going to a screen with a much smaller size and resolution. Factoring in that, it gets under 5W for sure. Also, from what I've heard, the X1's GPU uses just 1.5W when clocked at 500MHz, so a Pasal-based one should be 700MHz for 1.5W.

And yes, it's obviously going to be much slower than XB1 in even the best case.

1.5Watts? Where's it's 5 to 6 inch mobile phone solution?

Please tell me you're not going to use a logical fallacy to prove a point. You're better than that.

Avatar image for ronvalencia
ronvalencia

29612

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#155  Edited By ronvalencia
Member since 2008 • 29612 Posts

@techhog89 said:
@ronvalencia said:
@techhog89 said:
@ronvalencia said:

Note why I stated 8 to 9 inch tablet not 10.2 inch Pixel C tablet.

Change candybar with Phablet.

http://www.notebookcheck.net/Google-Pixel-C-Tablet-Review.158868.0.html

The price for Pixel C is $499 USD for 32 GB model https://store.google.com/product/pixel_c

40 percent faster speed yields around 56,931 points.

Nexus 9 with K1 is only 35 percent slower.

For comparison

40 percent of 14 watts is 5.6 watts... getting close.

My estimate for NX, it's far from XBO.

You're forgetting about the decrease in power consumption from going to a screen with a much smaller size and resolution. Factoring in that, it gets under 5W for sure. Also, from what I've heard, the X1's GPU uses just 1.5W when clocked at 500MHz, so a Pasal-based one should be 700MHz for 1.5W.

And yes, it's obviously going to be much slower than XB1 in even the best case.

1.5Watts? Where's it's 5 to 6 inch mobile phone solution?

Please tell me you're not going to use a logical fallacy to prove a point. You're better than that.

My logic is backed by development patterns from multiple phone vendors.

Tegra X1 already has aggressive clock gating/clock speed throttling when it's idling or low compute usage.

At idle mode with max aggressive clock gating/clock speed throttling, Shield TV consumes around 3.2 watts without a screen.

Avatar image for GameboyTroy
GameboyTroy

9730

Forum Posts

0

Wiki Points

0

Followers

Reviews: 7

User Lists: 1

#156  Edited By GameboyTroy
Member since 2011 • 9730 Posts

The NX console will be released first in March 2017.

http://nintendonews.com/news/general/nintendo-nx-home-console/

Avatar image for techhog89
Techhog89

5430

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#157 Techhog89
Member since 2015 • 5430 Posts

@ronvalencia: The 1.5W is just for the GPU itself. It's not including the rest of the SOC.

Avatar image for ronvalencia
ronvalencia

29612

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#158  Edited By ronvalencia
Member since 2008 • 29612 Posts

@techhog89 said:

@ronvalencia: The 1.5W is just for the GPU itself. It's not including the rest of the SOC.

The system doesn't operate with just the GPU and system power consumption has double importance with handheld device form factors.

Qualcomm/Apple ARM SoC competitors reaches sub-1 watts. If TX1 is king/god for handheld devices, where TX1's 5 to 6 inch smart phone solution?

20 watts to 14 watts = 30 percent power consumption reduction.

http://www.futuremark.com/hardware/mobile/Google+Pixel+C/review

Pixel C's Ice Storm Graphics score is 52,457 (~14 watts, 10.2 screen tablet form factor)

Shield TV's Ice Storm Graphics score is 61,735 (~21 watts, lite desktop form factor)

14 percent reduction in performance.

Avatar image for techhog89
Techhog89

5430

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#159  Edited By Techhog89
Member since 2015 • 5430 Posts

@ronvalencia: The fact that no TX1 phone exists isn't proof that making a TX1-based phone is impossible. :/

Also, again, you're comparing apples to oranges. Out of that 14W, the screen is probably using close to half.

Avatar image for ronvalencia
ronvalencia

29612

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#160 ronvalencia
Member since 2008 • 29612 Posts

@techhog89 said:

@ronvalencia: The fact that no TX1 phone exists isn't proof that making a TX1-based phone is impossible. :/

Your arguments are not proof for 5 to 6 inch mobile phone considerations.

Avatar image for techhog89
Techhog89

5430

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#161 Techhog89
Member since 2015 • 5430 Posts

@ronvalencia said:
@techhog89 said:

@ronvalencia: The fact that no TX1 phone exists isn't proof that making a TX1-based phone is impossible. :/

Your arguments are not proof for 5 to 6 inch mobile phone considerations.

So, what are you guessing for NX then? 1 SM at 200MHz for 51 GFLOPS?

Avatar image for ronvalencia
ronvalencia

29612

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#162  Edited By ronvalencia
Member since 2008 • 29612 Posts

@techhog89 said:
@ronvalencia said:
@techhog89 said:

@ronvalencia: The fact that no TX1 phone exists isn't proof that making a TX1-based phone is impossible. :/

Your arguments are not proof for 5 to 6 inch mobile phone considerations.

So, what are you guessing for NX then? 1 SM at 200MHz for 51 GFLOPS?

My arguments are for the upper limit expectations i.e. NX has to beat Microsoft Surface Pro 4 before it beats XBO or PS4.

I'm not going to be box in with a specific estimate.

My view, NX is "Wii U 2.0" with a proper mobility like an 8 inch Tablet and Unreal Engine 4 support.

Avatar image for techhog89
Techhog89

5430

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#163 Techhog89
Member since 2015 • 5430 Posts

@ronvalencia said:
@techhog89 said:
@ronvalencia said:
@techhog89 said:

@ronvalencia: The fact that no TX1 phone exists isn't proof that making a TX1-based phone is impossible. :/

Your arguments are not proof for 5 to 6 inch mobile phone considerations.

So, what are you guessing for NX then? 1 SM at 200MHz for 51 GFLOPS?

My arguments are for the upper limit expectations i.e. NX has to beat Microsoft Surface Pro 4 before it beats XBO or PS4.

I'm not going to be box in with a specific estimate.

So, basically, you're just here to state the obvious and you don't actually know what you're talking about.

Avatar image for emgesp
emgesp

7848

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#164  Edited By emgesp
Member since 2004 • 7848 Posts

@techhog89:

Look's like I read Ron's post wrong.

Yeah, X2 at 3 watts would yield less than 400 Gflops. Well, at least its still a huge upgrade over the 3DS.

Avatar image for ronvalencia
ronvalencia

29612

Forum Posts

0

Wiki Points

0

Followers

Reviews: 1

User Lists: 0

#165  Edited By ronvalencia
Member since 2008 • 29612 Posts

@techhog89 said:
@ronvalencia said:
@techhog89 said:
@ronvalencia said:

Your arguments are not proof for 5 to 6 inch mobile phone considerations.

So, what are you guessing for NX then? 1 SM at 200MHz for 51 GFLOPS?

My arguments are for the upper limit expectations i.e. NX has to beat Microsoft Surface Pro 4 before it beats XBO or PS4.

I'm not going to be box in with a specific estimate.

So, basically, you're just here to state the obvious and you don't actually know what you're talking about.

I have already stated my TX1 16 nm FinFET estimate which is based from TSMC's guidelines.

I do not need to repeat myself.

Your 500 Mhz argument is amateurish i.e. mobile SoCs has very aggressive clock speed throttling. You think TX1 stays at 850 Mhz or 1 Ghz all the time?

I'm not Nintendo i.e. NX is either 5-to-6 inch smart phone or 7-to-8 inch tablet power consumption profiles. It depends on the size of the device and battery pack i.e. Pixel C's battery pack is not cheap.

A modern chip design has multiple p-states. A system integrator could limit the upper p-state to reach a certain power consumption and performance goals.

Avatar image for bunchanumbers
bunchanumbers

5709

Forum Posts

0

Wiki Points

0

Followers

Reviews: 11

User Lists: 0

#166 bunchanumbers
Member since 2013 • 5709 Posts

Guys... it doesn't matter what the specs are. If it runs Grid it will always have access to the latest and greatest in 3rd party content. Nintendo has already shown that they can make a game work on a potato. Between them and Grid they are fine. NX will be awesome.

Avatar image for superbuuman
superbuuman

6400

Forum Posts

0

Wiki Points

0

Followers

Reviews: 14

User Lists: 0

#167  Edited By superbuuman
Member since 2010 • 6400 Posts

@GameboyTroy said:

The NX console will be released first in March 2017.

One more month to go...till Nintendo release info on the NX..officially? :P

Avatar image for KungfuKitten
KungfuKitten

27389

Forum Posts

0

Wiki Points

0

Followers

Reviews: 42

User Lists: 0

#168 KungfuKitten
Member since 2006 • 27389 Posts

Hmm people are mentioning short battery life a lot... But pascal is very tiny, right? I thought it was pretty efficient compared to older tech.

Avatar image for scrollinglayers
ScrollingLayers

632

Forum Posts

0

Wiki Points

0

Followers

Reviews: 0

User Lists: 0

#169  Edited By ScrollingLayers
Member since 2015 • 632 Posts

Reminder, Nvidia has their Q2 earnings call in about an hour.

http://edge.media-server.com/m/p/n6phuy56

If they happen to mention a new design win, that would be almost a confirmation of Tegra in NX.