▲ 1054 ▼ Dell says the quiet part out loud: Consumers don't actually care about AI PCs — "AI probably confuses them more than it helps them" (www.windowscentral.com) submitted 10 months ago by throws_lemy@reddthat.com to c/technology@lemmy.world 127 comments fedilink hide all child comments
[–] Gsus4@mander.xyz 5 points 10 months ago (5 children) Can the NPU at least stand in as a GPU in case you need it? permalink fedilink source parent hideshow 5 child comments replies: [–] Sam_Bass@lemmy.world 5 points 10 months ago Nope. Don't need it permalink fedilink source parent [–] UsoSaito@feddit.uk 4 points 10 months ago (3 children) No as it doesn't compute graphical information and is solely for running computations for "AI stuff". permalink fedilink source parent hideshow 3 child comments replies: [–] Gsus4@mander.xyz 1 point 10 months ago* (last edited 10 months ago) (2 children) GPUs aren't just for graphics. They speed up vector operations, including those used in "AI stuff". I just never heard of NPUs before, so I imagine they may be hardwired for graph architecture of neural nets instead of linear algebra, maybe, so that's why they can't be used as GPUs. permalink fedilink source parent hideshow 2 child comments replies: [–] JATtho@lemmy.world 2 points 10 months ago* (last edited 10 months ago) (1 child) Initially, x86 CPUs didn't have a FPU. It cost extra, and was delivered as a separate chip. Later, GPU is just a overgrown SIMD FPU. NPU is a specialized GPU that operates on low-precision floating-point numbers, and mostly does matrix-multiply-and-add operations. There is zero neural processing going on here, which would mean the chip operates using bursts of encoded analog signals, within power consumption of about 20W, and would be able to adjust itself on the fly online, without having a few datacenters spending exceeding amount of energy to update the weights of the model. permalink fedilink source parent hideshow 1 child comment replies: [–] UsoSaito@feddit.uk 1 point 10 months ago NPUs do those calculations far more effectively than a GPU though is what I was meaning. permalink fedilink source parent
[–] UsoSaito@feddit.uk 4 points 10 months ago (3 children) No as it doesn't compute graphical information and is solely for running computations for "AI stuff". permalink fedilink source parent hideshow 3 child comments replies: [–] Gsus4@mander.xyz 1 point 10 months ago* (last edited 10 months ago) (2 children) GPUs aren't just for graphics. They speed up vector operations, including those used in "AI stuff". I just never heard of NPUs before, so I imagine they may be hardwired for graph architecture of neural nets instead of linear algebra, maybe, so that's why they can't be used as GPUs. permalink fedilink source parent hideshow 2 child comments replies: [–] JATtho@lemmy.world 2 points 10 months ago* (last edited 10 months ago) (1 child) Initially, x86 CPUs didn't have a FPU. It cost extra, and was delivered as a separate chip. Later, GPU is just a overgrown SIMD FPU. NPU is a specialized GPU that operates on low-precision floating-point numbers, and mostly does matrix-multiply-and-add operations. There is zero neural processing going on here, which would mean the chip operates using bursts of encoded analog signals, within power consumption of about 20W, and would be able to adjust itself on the fly online, without having a few datacenters spending exceeding amount of energy to update the weights of the model. permalink fedilink source parent hideshow 1 child comment replies: [–] UsoSaito@feddit.uk 1 point 10 months ago NPUs do those calculations far more effectively than a GPU though is what I was meaning. permalink fedilink source parent
[–] Gsus4@mander.xyz 1 point 10 months ago* (last edited 10 months ago) (2 children) GPUs aren't just for graphics. They speed up vector operations, including those used in "AI stuff". I just never heard of NPUs before, so I imagine they may be hardwired for graph architecture of neural nets instead of linear algebra, maybe, so that's why they can't be used as GPUs. permalink fedilink source parent hideshow 2 child comments replies: [–] JATtho@lemmy.world 2 points 10 months ago* (last edited 10 months ago) (1 child) Initially, x86 CPUs didn't have a FPU. It cost extra, and was delivered as a separate chip. Later, GPU is just a overgrown SIMD FPU. NPU is a specialized GPU that operates on low-precision floating-point numbers, and mostly does matrix-multiply-and-add operations. There is zero neural processing going on here, which would mean the chip operates using bursts of encoded analog signals, within power consumption of about 20W, and would be able to adjust itself on the fly online, without having a few datacenters spending exceeding amount of energy to update the weights of the model. permalink fedilink source parent hideshow 1 child comment replies: [–] UsoSaito@feddit.uk 1 point 10 months ago NPUs do those calculations far more effectively than a GPU though is what I was meaning. permalink fedilink source parent
[–] JATtho@lemmy.world 2 points 10 months ago* (last edited 10 months ago) (1 child) Initially, x86 CPUs didn't have a FPU. It cost extra, and was delivered as a separate chip. Later, GPU is just a overgrown SIMD FPU. NPU is a specialized GPU that operates on low-precision floating-point numbers, and mostly does matrix-multiply-and-add operations. There is zero neural processing going on here, which would mean the chip operates using bursts of encoded analog signals, within power consumption of about 20W, and would be able to adjust itself on the fly online, without having a few datacenters spending exceeding amount of energy to update the weights of the model. permalink fedilink source parent hideshow 1 child comment replies: [–] UsoSaito@feddit.uk 1 point 10 months ago NPUs do those calculations far more effectively than a GPU though is what I was meaning. permalink fedilink source parent
[–] UsoSaito@feddit.uk 1 point 10 months ago NPUs do those calculations far more effectively than a GPU though is what I was meaning. permalink fedilink source parent