Intel is so cute. Thinking they can beat Nvidia in DEEP/machine memes.

File: Intel-Xeon-Lake-Crest-Deep-Learning-Features.jpg (574KB, 1600x832px) Image search: [Google]

574KB, 1600x832px

>>60529131

>>60529131
Not like they have anything to lose by trying.

>>60529180
Besides face and development money/time better

>>60529131
Intel has a lot things better nVidia, but Nvidia software is way better.

>>60529131
More deep memes? I wonder when AMD will jump on tensor memetrain.

>>60529950
With Vega.

>>60529976
No tensor memes, just pure compute.

>>60530084
4x4x4 matrix multiply is just one of several deep/machine learning workloads.

>>60530109
Yes but everyone but AMD already jumped on tensor memetrain.

>>60530133
Only Nvidia did, Intel's solution(in the OP) isn't even out yet.

>>60530260
Volta is also paperlaunced. The only working tensor meme is google one.

File: 1492224325484.jpg (78KB, 637x613px) Image search: [Google]

78KB, 637x613px

>>60529131
>proprietary inter-chip links

>>60529131
>10x of state-of-art
LOL

>>60529131
Ok. Next time they ask you what to do.