AMD Introduces GAIA - an Open-Source Project That Runs Local LLMs on Ryzen AI NPUs

[NoLoihi]

It is though. The item at hand is just the NPU. The GPU was not mentioned in that statement and thus is not the one being compared to.

In the heading right above the sentence I’ve quoted, it says “Comparing NPU and iGPU”.

wake me up when they integrate the npu into the big desktop processors.

Such as Intel has done?

 

[Cheeseball]

In the heading right above the sentence I’ve quoted, it says “Comparing NPU and iGPU”.

Good catch. The NPU is still more performant as it does 50 TOPS compared to the 890M’s ~30 TOPS.

 

[lexluthermiester]

Scripting language will never be fast and memory efficient.

Not true at all.

 

[trsttte]

Scripting language will never be fast and memory efficient.

Have fun with those decade old misconceptions then

 

[regs]

Not true at all.

Self-hypnosis, or as americans like to say, wishful thinking, wouldn't help to change the fact.

 

[lexluthermiester]

Self-hypnosis, or as americans like to say, wishful thinking, wouldn't help to change the fact.

Are going to meander about the forums shouting nonsense for long?