Support NeoGAF

[cormack12]

Source: https://nvidianews.nvidia.com/news/nvidia-microsoft-windows-pcs-agents-rtx-spark

The RTX Spark superchip features an NVIDIA Blackwell RTX GPU with 6,144 CUDA cores and fifth-generation Tensor Cores with FP4 precision, connected via the NVIDIA NVLink®-C2C chip-to-chip interconnect to a high-performance, 20-core NVIDIA Grace™ CPU

Designed for AI, creating and gaming, RTX Spark brings together 30 years of NVIDIA innovation — including NVIDIA CUDA®, NVIDIA RTX™, DLSS, FP4, NVIDIA TensorRT™, NVIDIA OptiX™, Reflex and G-SYNC® — to slim Windows laptops with all-day battery life and small, ultraefficient desktop PCs.

• RTX Spark lets creators, AI developers and gamers render ultralarge 90GB+ 3D scenes, edit 12K 4:2:2 video, generate 4K AI videos, run 120B-parameter LLMs with up to 1 million tokens context using agents locally, and play AAA games at 1440p and over 100 frames per second.
• NVIDIA and Microsoft collaborate to deliver a native Windows experience for personal agents, including new security primitives and NVIDIA OpenShell to run agents securely on primary devices.

"The PC is being reinvented," said Jensen Huang, founder and CEO of NVIDIA. "For forty years, you launched apps. Click. Type. With RTX Spark and Microsoft Windows, you ask — and the PC does the work. RTX Spark brings everything NVIDIA has built — CUDA, RTX, our AI platform — into a single superchip. Local agents. Frontier models. Creative workflows. RTX games. All on a laptop. This is the new PC. The personal AI computer."

 

[bad guy]

I don't want to ask/beg. I want to do.

Edit 12K videos, all day battery. Hard to believe.

 

[SJRB]

O god when does it stop

For forty years, you launched apps. Click. Type. With RTX Spark and Microsoft Windows, you ask — and the PC does the work.

This is literally how a PC has been operating since its inception: you give a command and the PC does the thing.

 

[yamaci17]

more unnecessary bloat
just corporates finding uncreative ways to cram AI where it's not necessary
the mountain has brought forth a mouse...

 

[Nocty]

they keep coming up with more and more shit to keep the bubble inflated.

Sooner or later it pops and valuations will come crashing back to earth

 

[MCH2024]

Source: https://nvidianews.nvidia.com/news/nvidia-microsoft-windows-pcs-agents-rtx-spark

The RTX Spark superchip features an NVIDIA Blackwell RTX GPU with 6,144 CUDA cores and fifth-generation Tensor Cores with FP4 precision, connected via the NVIDIA NVLink®-C2C chip-to-chip interconnect to a high-performance, 20-core NVIDIA Grace™ CPU

Designed for AI, creating and gaming, RTX Spark brings together 30 years of NVIDIA innovation — including NVIDIA CUDA®, NVIDIA RTX™, DLSS, FP4, NVIDIA TensorRT™, NVIDIA OptiX™, Reflex and G-SYNC® — to slim Windows laptops with all-day battery life and small, ultraefficient desktop PCs.

• RTX Spark lets creators, AI developers and gamers render ultralarge 90GB+ 3D scenes, edit 12K 4:2:2 video, generate 4K AI videos, run 120B-parameter LLMs with up to 1 million tokens context using agents locally, and play AAA games at 1440p and over 100 frames per second.
• NVIDIA and Microsoft collaborate to deliver a native Windows experience for personal agents, including new security primitives and NVIDIA OpenShell to run agents securely on primary devices.

"The PC is being reinvented," said Jensen Huang, founder and CEO of NVIDIA. "For forty years, you launched apps. Click. Type. With RTX Spark and Microsoft Windows, you ask — and the PC does the work. RTX Spark brings everything NVIDIA has built — CUDA, RTX, our AI platform — into a single superchip. Local agents. Frontier models. Creative workflows. RTX games. All on a laptop. This is the new PC. The personal AI computer."

Revolutionary SoC as expected. Strix Halo level CPU + Medusa Halo level iGPU. To think it was delayed 18 months. Nvidia is on another level.

This will put strix halo in the trash can where it belongs.

 

[Tobimacoss]

Wow, that's powerful.

 

[twilo99]

The price point will be hilarious

 

[peish]

Revolutionary SoC as expected. Strix Halo level CPU + Medusa Halo level iGPU. To think it was delayed 18 months. Nvidia is on another level.

This will put strix halo in the trash can where it belongs.

Amd will renew with a 192GB strix halo. lol it would work better if ram prices arent crazy...which they themselves caused it.

yeah, they dont really give a fuck now.

Even this RTX spark is a trickle leftover of DGX Spark, without the Nvidia network card. For some reasons, network card can cost a lot of $$$

https://www.tomshardware.com/pc-components/gpus/nvidia-dgx-spark-review

 

[Dirk Benedict]

This is for the bed rot gen, that's coming up.

 

[IDKFA]

more unnecessary bloat
just corporates finding uncreative ways to cram AI where it's not necessary
the mountain has brought forth a mouse...

they keep coming up with more and more shit to keep the bubble inflated.

Sooner or later it pops and valuations will come crashing back to earth

Agreed. It's a way to justify the obscene money being spent on AI.

What is the major "revolution" here? That I can bark commands at my laptop and it does all the work for me?

It just sounds pointless. I'd rather do the work myself than get closer to humans from the film Wall-E.

 

[Pedro Motta]

well the DGX Spark price is ridiculous given it's specs, this will be even more expensive? I'm actually interested in the GB300 platform...but it really depends on the price...

 

[Brucey]

Revolutionary SoC as expected. Strix Halo level CPU + Medusa Halo level iGPU. To think it was delayed 18 months. Nvidia is on another level.

This will put strix halo in the trash can where it belongs.

"

• "We worked closely with NVIDIA to support a great gaming experience and are excited to expand access to XBOX on RTX Spark devices, making it easy for players to discover and play with XBOX on PC," said Jason Ronald, vice president of Next Generation at XBOX.
• "RTX Spark allows even more gamers to experience NetEase titles like 'NARAKA: BLADEPOINT' on ultrathin, high-performance laptops the way the developers intended," said Long Cheng, senior vice president of Thunderfire BU at NetEase.
• "Remedy is looking forward to bringing its games together with NVIDIA to these stunning new RTX Spark laptops," said Mika Vehkala, chief technology officer of Remedy Entertainment."

Perhaps more exciting than helix and 2 years sooner assuming 2028 helix launch?

 

[General Lee]

AMD's going to need to step up. When's the Strix Halo successor coming?

I do expect these to be hideously expensive though, given how much Strix Halo systems cost already. Not going to be useful for anyone not really into local models.

 

[Prodyhig]

I ask from a place of complete ignorance:

When are we likely to see desktop PCs released with this technology?

 

[spons]

I don't get why NVIDIA partnered with Microsoft on this. They don't need them. Microsoft needs NVIDIA. So I hope their paycheck for this is rather exorbitant. You can market games all you want, but nobody is going to buy these Arm laptops for that. Windows is dead weight.

 

[nkarafo]

Go bankrupt pls.

 

[Shubh_C63]

For forty years, you launched apps. Click. Type. With RTX Spark and Microsoft Windows, you ask — and the PC does the work.

This is the first step towards the Mother Computer interface like in Aliens. No interactivity or changes can be made, just yes and no.
"Personal" computer hell.

 

[TrebleShot]

Very cool and id definitely be interested , running a 100b model on a local machine right now would need 4 5090s.

Put into perspective I run a Gemma 4 model locally with a 28b model and it just about runs anything close to or above your vram crawls to a halt and you'll hear your GPU crunching through it. (you can get quantised models)

100b is very impressive if true and yes this will be extortionate.

Local AI is getting better and better and is much better security and compliance wise.

This is an incredible advancement and will be keeping a close eye on it.

 

[Hadji*]

So all that new era of PC is just AI nonsense..

I'm Good, The new PC era for me would be when SteamOS becomes stable enough for me to switch to.

 

[ShirAhava]

AI-RTX Copilot laptop $3999.99 coming soon

Slop. It. Up.

 

[Hudo]

Nvidia have successfully conned Microsoft into a deal where Nvidia can sell more of their shit, while Microsoft can show the world again why they are the king of Amerijank.

That'll be $5000.

 

[Black_Stride]

Users can render ultralarge 90GB 3D scenes with OptiX and DLSS
NVIDIA to bring OTOY Octane with Render Network support to RTX Spark will deliver a new class of portable systems for creators.

^This doesnt mean shit, if its a 90GB scene you are using DLSS and OptiX over the Render Network........mate, the PC is barely doing any work.
Who the fuck is loading a 90GB scene in RAM anyway?




<---Jealous of being able to load more than 16GB into Unified Memory.......Out-Of-Core Rendering sucks!
<---Uses OptiX and DLSS

 

[Zathalus]

AMD's going to need to step up. When's the Strix Halo successor coming?

I do expect these to be hideously expensive though, given how much Strix Halo systems cost already. Not going to be useful for anyone not really into local models.

Strix Halo came out a year ago, Medusa Halo is apparently coming out next year and based on leaks will probably have faster GPU performance compared to this (double the memory bandwidth for one thing). Nvidia will probably counter in 2028 with something even faster again.

 

[T-Cake]

This is ARM right so will need CPU translation for games?

 

[An Corp]

they keep coming up with more and more shit to keep the bubble inflated.

Sooner or later it pops and valuations will come crashing back to earth

They sure do. Honestly I'm surprised most of the investors haven't pulled out of ai.

 

[Gamezone]

Have they introduced some ARM support for gaming that we missed? I thought you had to use PC emulators on Android to run them.

 

[Larxia]

Things I'm thinking:
- At what price?
- If the price is the same (or more expensive) than gpu, ram, cpu parts combined, I don't think it really is worth it for desktop usage, because you won't be able to upgrade components individually anymore. If you want to upgrade something, purchase a new chip that costs 4000$, instead of just upgrading one specific part.
- They mention 1440p 100 fps on games like Forza and 007, but in what conditions? Because I remember when they announced the RTX 5000 series and claimed the 5070 (non ti) had 4090 performances. If it's with DLSS performance and X6 frame generation...

 

[BloodFree]

 

[Sintoid]

Leatherjacket deliver shit again

 

[MCH2024]

Perhaps more exciting than helix and 2 years sooner assuming 2028 helix launch?

Definitely

This is forecast to be 10M units sold / 2 years. Helix probably won't sell that whole life time.

The fact RTX Spark Grace Blackwell was literally delayed 18 months and it's still leading by this much is an indictment of X86 complacency.

And RTX Spark Grace Blackwell is just a little kitten.

The Serial Mogger is RTX Spark Vera Rubin.

ARM C2 Ultra "Canyon" Cores
Graphics Rubin Architecture
TSMC 2N Process Node

AMD's Medusa Halo is as a result Dead on Arrival.

One needs to understand Zen6 actually needs 7.0 GHz to match 4.5 GHz Canyon LMAO. Canyon is a monster 12 wide core with IPC approaching 2 x Golden Cove.

It's called a SUPER Core (and not a performance core) for a reason.

 

[Topher]

I ask from a place of complete ignorance:

When are we likely to see desktop PCs released with this technology?

This fall

 

[MCH2024]

Amd will renew with a 192GB strix halo. lol it would work better if ram prices arent crazy...which they themselves caused it.

Strix Halo is in 3 laptops and the 3 laptops are selling like shit. And that's before Nvidia released this beast that destroys Strix Halo in Graphics and has CUDA.

Strix Halo was DOA as a laptop chip and now it's been buried 6 feet under so bad I am surprised they aren't just playing dead.

I am not surprised they are turning to Bozo commodity maxing when their product is so bad. They always do this.

But 192GB LPDDR5X will cost a kidney. Why would you pay for so much premium memory when it's placed alongside a budget brand failed SoC with sketchy software support?

 

[diffusionx]

The chip could be good or not good but unless Microsoft fixes Windows to not be a piece of shit it doesn't matter. Apple's M chips are so good in part because of the OS integration.

 

[Topher]

Have they introduced some ARM support for gaming that we missed? I thought you had to use PC emulators on Android to run them.

Games are a secondary concern here. This primarily about running local AI.

 

[MCH2024]

- At what price?

Preliminary pricing for a Lenovo Yoga Pro 7

2,699$ for N1-475 + 32GB + 1TB
3,099 for N1X-650 + 32GB + 1TB
3,899$ for N1X-675 + 64GB + 1TB

Nvidia is a premium brand. So the pricing is a bit less than MacBook Pro but not by much. There might be cheapo 16GB N1X laptops in the future. But for now, it leans expensive.

- If the price is the same (or more expensive) than gpu, ram, cpu parts combined, I don't think it really is worth it for desktop usage, because you won't be able to upgrade components individually anymore. If you want to upgrade something, purchase a new chip that costs 4000$, instead of just upgrading one specific part.

the reason AI developers want it is because it's a lot of memory connected to the GPU directly.

RTX 5090 is 4k and 32GB only. RTX 6000 PRO is like 7 grand and 96GB.

This works out great with 128GB.

- They mention 1440p 100 fps on games like Forza and 007, but in what conditions?

Remember the target audience. They aren't going to bore normies with average FPS graphs. They want memorable lines and easy to understand.

 

[Topher]

Not sure what the source here is

 

[VAVA Mk2]

This will be expensive

 

[diffusionx]

Not sure what the source here is

I still think the best option for these workflows is a Mac, but I get why they're doing this.

 

[Topher]

 

[RoboFu]

Microsoft reports are exposing AI's real cost problem: Using the tech is more expensive than paying human employees | Fortune

Companies are racing to incentivize employees to use AI. But as some companies are finding, the more employees that use the technology, the heavier the bill.

fortune.com

 

[Witchilich]

This will be another fail.

 

[fran.fernandez]

Keep going Nvidia , for my second house lol

 

[Topher]

Microsoft reports are exposing AI's real cost problem: Using the tech is more expensive than paying human employees | Fortune

Companies are racing to incentivize employees to use AI. But as some companies are finding, the more employees that use the technology, the heavier the bill.

fortune.com

Which is why local AI will become more and more an attractive option.

 

[Nocty]

Leatherjacket deliver shit again

Ineed. He's a predictable saussage if nothing else.

Lets be honest, as an Nvidia shareholder since like 2016, he's the best CEO ever in history.

 

[StereoVsn]

I am curious on performance of these for local AI models and the price.

Unified memory will go a long way for the above potentially. However if it's more expensive than equivalent Mac MBPro or Studio, then meh.

 

[Topher]

lol

 

[Trogdor1123]

That's a lot of words. What does it mean in layman terms?

 

[Hudo]

That's a lot of words. What does it mean in layman terms?

That'll be $5000.

 

[Wolzard]

Introducing a powerful new chapter for Windows PCs, accelerated by NVIDIA RTX Spark

Today at NVIDIA GTC, Microsoft and NVIDIA announced the world’s most powerful and efficient thin-and-light Windows PCs ever. Accelerated by NVIDIA RTX Spark

blogs.windows.com

About gaming:

Prism emulation enhancements

Prism, our emulator for running 32-bit and 64-bit x86 apps on Windows on Arm, will also be present and optimized for RTX Spark powered PCs. Prism ensures apps run well on these devices even if those apps haven't been built for the Arm architecture. We have continued to enhance the Prism emulator with additional performance and compatibility features, building on the Prism optimizations delivered last year that added support for the AVX/AVX2 instruction set extensions. Prism has been tuned for the microarchitecture of RTX Spark and when combined with the raw power of the silicon, unlocks great performance for developers, creators and gaming workloads running under emulation.

On Game Developers Support

Game developers have also laid a strong foundation for RTX Spark's arrival. Today, native anti-cheat solutions from partners like Epic's Easy Anti-Cheat and BattlEye, expanded Prism emulator compatibility, and XBOX PC app support means players will have access to a deep catalog of Windows PC games. RTX Spark will bring even higher levels of gaming performance to AAA titles on Arm. Riot Games, one of the world's leading game developers and publishers, has announced that League of Legends and VALORANT are coming to the platform. PUBG: Battlegrounds, the iconic battle royale title from KRAFTON, will also be joining the expansive catalog of compatible titles including Pragmata, Alan Wake 2*,* Naraka: Bladepoint, War Thunder and more.

---

This first generation should focus on creators and developers, but the idea seems to be to create a new market within the ARM architecture, especially since Nvidia has shown that it already has two more generations planned.