News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products

[regen1]

Few Jensen takeaways:
> mentioned something like a year or so of collaboration going on for this.
> Claims no change in ARM roadmaps.
Wants to reach whatever general computing platform that has market reach. Mentions Intel unquestionably having the largest enterprise, industrial, cloud, consumer footprint. Praises FOVEROS.
> With x86 platform they were using external PCIe retimers(scaling up to NVLink 8) whereas with ARM they were doing NVL72. Collab brings parity and thus more rack-scale on x86.

The scope of collab is still not fully clear, more clarity with time.

 

[Josh128]

So I listened to the call. My takeaways (in no particular order).

• AI! AI! AI!
• They focused more on server CPUs than client GPUs.
• They said Intel is selling server CPUs to Nvidia who will sell them to customers.
• The deal has supposedly been in the works for nearly a year
• If I heard him right, Lip-Bu Tan mentioned unified memory for the client SOCs which would imply it's a true iGPU.
• Huang said this deal has no impact on Nvidia's ARM roadmap.
• No announcement of Nvidia products being fabbed by Intel Foundry
• No mention of what happens to Intel's graphics IP or datacenter AI accelerators.
• When asked why the cash infusion into Intel, Huang said "It's a good investment".

Fantastic screen name and avatar by the way. 10/10

 

[fastandfurious6]

NVIntel Halo incoming

but how soon? all nvidia has is VERY FAT gpus, can they shrink them efficiently for APU/SOC mix with hot&hungry intel x86?

Tegra? lol!

 

[adroc_thurston]

can they shrink them efficiently for APU/SOC mix

yeah.

 

[fastandfurious6]

Yeah NV went from "ARM is all you need, we're gonna literally buy the goddamn company" to this.
whatever that is (total complete surrender to x86?).

instinct says NV saw early Medusa Halo results and decided "damn we need x86 ASAP"

 

[adroc_thurston]

instinct says NV saw early Medusa Halo results and decided "damn we need x86 ASAP"

Kind of.
WoA is just dead on arrival.

 

[jpiniero]

All the high margin opportunities are on graphics/AI, not CPU.

AI is a dead opportunity really. You should pretty much expect even AMD to not get much. The Cloud Providers will roll their own, and that will be the main competitor to nVidia.

 

[eek2121]

Kind of.
WoA is just dead on arrival.

As it has always been. For many years folks on this forum and elsewhere have been swearing ARM will overtake x86. I suspect many more years will pass at minimum. Many folks want cheap/efficient/fast/compatible. Not going to happen for quite a while. x86 will continue to offer a good balance for any serious workload.

I didn’t come here to talk about that, however. I came here to talk about this:

https://videocardz.com/newz/intel-nvidia-gpu-collaboration-is-complementary-gpu-product-offerings-to-continue

This is corporate speak for: “it’s dead, jim.”

 

[poke01]

WoA is dead cause of Microsoft, Qualcomm and various other things. There STILL not a single CPU on that platform that is worthy.

 

[poke01]

Hell yeah it does (yes, cinebench, yadda-yadda).
View attachment 130529

this chart is what Intel suits won't show. They talked about Lunar Lake improved battery life but that was just in video playback which is performed by the very good media engines in Lunar. The actual CPU core IP is still dog crap.

 

[511]

this chart is what Intel suits won't show. They talked about Lunar Lake improved battery life but that was just in video playback which is performed by the very good media engines in Lunar. The actual CPU core IP is still dog crap.

The LP-E cores are perfectly fine though they handle everything on battery that is why LNL has so good Battery Life.
Wish test has been done on Linux instead of windows.

 

[Win2012R2]

AMD better speed things up now

 

[adroc_thurston]

that is why LNL has so good Battery Life.

It's MoP and PMIC spam.

 

[511]

I have seen Some AMD investor complaining that Intel getting so much Money is unfair 🤣🤣 and MLIDs tweet is pure gold.

 

[511]

It's MoP and PMIC spam.

LP-E Cores do play a role if they were bad like P cores in PnP at such lower power. LNL would have sucked.

 

[marees]

I have seen Some AMD investor complaining that Intel getting so much Money is unfair 🤣🤣 and MLIDs tweet is pure gold.

Intel design is using its sick fab as an excuse to get govt & Nvidia money

 

[adroc_thurston]

LP-E Cores do play a role if they were bad like P cores in PnP at such lower power. LNL would have sucked.

Obviously, but platform power floor is achieved thru primarily platform engineering means.

 

[511]

Intel design is using its sick fab as an excuse to get govt & Nvidia money

Intel is an IDM and Design has been funding fabs for some time before it was fab were carrying Design group.