Nvidia GRACE ARM CPU Confirmed & Ampere NEXT RTX 40 GPU In 2022 - Nvidia Roadmap & GTC Analysis

Nvidia GRACE ARM CPU Confirmed & Ampere NEXT RTX 40 GPU In 2022 - Nvidia Roadmap & GTC Analysis

Nvidia's GTC Event had the company confirm its roadmap until 2024 for both GPU and CPU products.

Nvidia will be releasing ARM based processors named 'Grace' (named after Grace Hopper) and these will first appear in servers and offer over 10x increase in performance compared to an AMD or Intel X86 processor because of massive amounts of memory bandwidth and IO.

Nvidia's "Ampere Next" GPU is said to launch in 2022, and Ampere Next Next in 2024. Most likely, the GPU release in 2022 is going to be LOVELACE, with possibly 'hopper' being the release in 2024.

Nvidia's RTX 40 series cards are almost certain therefore to launch in 2022, and I've already discussed the GPU performance targets in an exclusive

Also, Nvidia's ARM based processors I had leaked, but when I was told about them it was for gaming - which possibly means that this is a different architecture entirely, or a very different configuration (assuming my info was accurate).

AMAZON AFFILIATE LINKS

GreenManGaming Affiliate link -

SOCIAL MEDIA

0:00 Nvidia GTC 2021
12:10 Metal Gear Solid rumours

[Chazzajoe]

Lovelace gonna be here sooner than Ampere has stock :P

[loki]

LOL JenSun Huang with that long hair is starting to morph into Lisa Su..

[thisguy]

Thanks for the lowdown, guys. Now I can go to the pub with a sound mind.

[RepsUp]

No new NVIDIA GPUs in 2021... thank god!

[JohnDoe-ssnn]

LOL We're almost at the half-way point of 2021 and I've yet still see an RTX 3000 series card that's not being scalped. 2022 doesn't look like it's going to shape up to be any better.

[hivemind]

I figured out if I dont subscribe and watch all your videos I actually get better notification for you somehow.

[rager]

I've heard that they are also going to not use the Lovelace name for GPUs, instead naming the ARM CPU after Grace as Linda.

[mixk]

Fuckin hell, here's to hoping 2022 will be a better year.
I still have a tiny piece of hope that maybe this year will finally improve.

[ASTROTZUR]

what's up with the corrupted frames? in this video it's at 14:42 and it also happens in some of your other videos

[rRefuseToSelfCensor]

Every YouTuber streaming todays gtc like 3080 ti reveal?

[jackmiller]

Regarding "Ampere Next" release in 2022, I have a reason to believe it references the Lovelance successor to the Ampere based A100 datacenter GPUs announced and released in 1H 2020, more specifically May. Consumer Geforce Ampere GPUs did not release until 2H 2020, so I am thinking that Lovelance will follow a similar pattern.

[allenpost]

"Power tends to corrupt and absolute power corrupts absolutely" Lord Acton

[grbm]

More competition = better for us. Hopefully Nvidia and AMD can force Intel to innovate for once.

[wheelchairskater]

I think it is possible Nintendo does use a more advanced sock than usual. I think Nvidia has something to prove in the console space and could give Nintendo a break on chips. Especially after being locked out by MS and Sony with them using AMD twice already with more to come if history is a indicator. I could see an Turing based soc for the Switch Pro/2.

[TeflonBilly]

FYI Hopper is named after the early Computer Scientist Rear Admiral Grace Hopper, it seems strange that they'd use that name if it's not unrelated to the other project.

[TomRFL]

Any bets on continued GPU shortages well into 22 and beyond?

[ChePennyDK]

I think, now that we have Windows for Arm, we will see a big/little combo as in smartphones, just with the big cores being X86 and the small ones Arm, so we can achieve low power consumption at idle or low loads.
Actually I suggested this to AMD 7-8 years ago...

[famousfighter]

Damn I wanted to be second (I think it would be cool to have a nvidia cpu and amd cpu or nvidia cpu and intel gpu)

[wazzamolloy]

RISC-V might prove interesting over the next 5 years and challenge Arm.

[ADB-zfzr]

Hi Paul, how do I communicate with you directly (in private).?