NVIDIA and SEGA Reunite After 30 Years, Bringing Classic Franchises to RTX Spark

David_Schroth

Administrator
Staff member
Joined
Apr 23, 2019
Messages
1,613
Reaction score
2,143
Thirty years ago, SEGA wrote NVIDIA a $5 million check and saved the company from going under. Jensen Huang flew to Tokyo to say thank you in person, and the two companies announced that SEGA’s game lineup is coming to NVIDIA RTX Spark. The first confirmed title is Virtua Fighter Crossroads, the new installment in […]

See full article...
 
I'd like to see some tecmo games on the spark... but that's different... I just like some of their games.... well seeing them at least.
 
First off - ok, so RTX Spark is a CPU. There kinda needs to be an entire machine, with Ram and storage and stuff...

Second - it's ARM... so not exactly groundbreaking for anything to be ported to an ARM CPU these days.

And lastly - people have to buy these machines, and then the games, for this announcement to mean anything at all. And today is the day of $1k+ Steam Machines and console manufacturers firing their entire workforce because hardware is basically non-existent.

Would be better for Sega to say they are backporting their entire lineup to Apple or Android or something (as much as I would cringe at that, it makes a lot more business sense).
 
First off - ok, so RTX Spark is a CPU
It's an SoC with a 20-core Grace ARM CPU and some kind of Blackwell GPU I believe.

There kinda needs to be an entire machine, with Ram and storage and stuff...
Exactly what I was thinking. I was not even familiar with RTX Spark until now. So are there systems coming with these SoCs in them? The article says "RTX Spark systems are expected 'this Fall.'" Is that so? I hadn't heard sh1t about that, but okay. And what exactly are these, handheld or mini systems? Laptops too?

But yes, I too am confused by this nVidia + Sega announcement. Kinda weird. Seems strange. I dunno.
 
Also. Seems that even though nVidia is announcing this, when you read about Spark, it sure seems like they are pushing it towards AI use too.

And if it can be used for AI, the voracious appetite of that monster consumes all.
 
Honestly reads more like a stab at Apple.
That said may be all about price in the end I guess.
The chip reads pretty muscular.
Honestly with 128 gig of ram meaning 96 gig for ai/inference and 32ish left for io and os... of you can daosy chain these like the Mac mini with the custom os to get 360+ gig of effective ram to natively run LLMs with obscene amounts of parameters without dropping 200k on a aerver.. that's pretty huge.

While you cant use this to train an AI effectively its still a great way to expiramnet.
 
Become a Patron!
Back
Top