That should help you with your financial decision makingDog stock. Time to trade these puppies. I’m waiting for mum to get home with more weed so I can pull another bong.
Correct.Isn't this similar to what we're trying to achieve to slash power consumption @Diogenese?
SNN's eliminating the need for matrix multiplication without affecting performance?
Researchers claim new approach to LLMs could slash AI power consumption
Work was conducted as concerns about power demands of AI have risen
Graeme Burton
27 June 2024• 2 min read![]()
SHARE
![]()
ChatGPT uses 10 times the electricity of a Google search – but cutting out ‘matrix multiplication’ could cut this number without affecting performance
New research suggests that eliminating the ‘matrix multiplication' stage of large-language models (LLMs) used in AI could slash power consumption without affecting performance.
The research was published at the same time that increasing concerns are being raised over the power demands that AI will make over the course of the decade.
Matrix multiplication – abbreviated to MatMul in the research paper – performs large numbers of multiplication operations in parallel, becoming the dominant operation in the neural networks that drive AI primarily due to GPUs being optimised for such operations.
But by leveraging Nvidia CUDA [Compute Unified Device Architecture] technology instead, along with optimised linear algebra libraries, this process can be efficiently parallelised and accelerated – cutting power consumption without penalising performance.
In the process, the researchers claim to have "demonstrated the feasibility and effectiveness of the first scalable MatMul-free language model", in a paper entitled Scalable MatMul-free Language Modelling.
The researchers created a custom 2.7 billion parameter model without using matrix multiplication, and found that performance was on a par with state-of-the-art deep learning models (called ‘transformers', a fundamental element of natural language processing). The performance gap between their model and conventional approaches narrows as the MatMul-free scales, they claim.
They add: "We also provide a GPU-efficient implementation of this model, which reduces memory usage by up to 61% over an unoptimised baseline during training. By utilising an optimised kernel during inference, our model's memory consumption can be reduced by more than 10 times compared to unoptimised models."
Less hardware-heavy AI models could also enable more pervasive AI, freeing the technology from its dependence on the data centre and the cloud. In addition, both OpenAI and Meta are set to unveil new models that, they claim, will be capable of reasoning and planning.
However, they cautioned, "one limitation of our work is that the MatMul-free language model has not been tested on extremely large-scale models (for example, 100+ billion parameters) due to computational constraints". They called for well-resourced institutions and organisations to build LLMs utilising lightweight models, "prioritising the development and deployment of matrix multiplication-free architectures".
Moreover, the researchers' work has not yet been subject to peer review. Indeed, power consumption on its own matters less than energy usage per unit of ‘output' – information that is not provided in the research paper.
![]()
Researchers claim new approach to LLMs could slash AI power consumption
ChatGPT uses 10 times the electricity of a Google search – but cutting out ‘matrix multiplication’ could cut this number without affecting performancewww.computing.co.uk
Scalable MatMul-free Language Modeling
Matrix multiplication (MatMul) typically dominates the overall computational cost of large language models (LLMs). This cost only grows as LLMs scale to larger embedding dimensions and context lengths. In this work, we show that MatMul operations can be completely eliminated from LLMs while...arxiv.org
Her Dad must be so proud.Fantastic PR and advice on how to increase the Brainchip share price
Someone can send this to Sean & the team
Have a great weekend everyone!
View attachment 65615
Off course he is - she owns shares in BRCHFHer Dad must be so proud.![]()
Hi Bravo,Isn't this similar to what we're trying to achieve to slash power consumption @Diogenese?
SNN's eliminating the need for matrix multiplication without affecting performance?
Researchers claim new approach to LLMs could slash AI power consumption
Work was conducted as concerns about power demands of AI have risen
Graeme Burton
27 June 2024• 2 min read![]()
SHARE
![]()
ChatGPT uses 10 times the electricity of a Google search – but cutting out ‘matrix multiplication’ could cut this number without affecting performance
New research suggests that eliminating the ‘matrix multiplication' stage of large-language models (LLMs) used in AI could slash power consumption without affecting performance.
The research was published at the same time that increasing concerns are being raised over the power demands that AI will make over the course of the decade.
Matrix multiplication – abbreviated to MatMul in the research paper – performs large numbers of multiplication operations in parallel, becoming the dominant operation in the neural networks that drive AI primarily due to GPUs being optimised for such operations.
But by leveraging Nvidia CUDA [Compute Unified Device Architecture] technology instead, along with optimised linear algebra libraries, this process can be efficiently parallelised and accelerated – cutting power consumption without penalising performance.
In the process, the researchers claim to have "demonstrated the feasibility and effectiveness of the first scalable MatMul-free language model", in a paper entitled Scalable MatMul-free Language Modelling.
The researchers created a custom 2.7 billion parameter model without using matrix multiplication, and found that performance was on a par with state-of-the-art deep learning models (called ‘transformers', a fundamental element of natural language processing). The performance gap between their model and conventional approaches narrows as the MatMul-free scales, they claim.
They add: "We also provide a GPU-efficient implementation of this model, which reduces memory usage by up to 61% over an unoptimised baseline during training. By utilising an optimised kernel during inference, our model's memory consumption can be reduced by more than 10 times compared to unoptimised models."
Less hardware-heavy AI models could also enable more pervasive AI, freeing the technology from its dependence on the data centre and the cloud. In addition, both OpenAI and Meta are set to unveil new models that, they claim, will be capable of reasoning and planning.
However, they cautioned, "one limitation of our work is that the MatMul-free language model has not been tested on extremely large-scale models (for example, 100+ billion parameters) due to computational constraints". They called for well-resourced institutions and organisations to build LLMs utilising lightweight models, "prioritising the development and deployment of matrix multiplication-free architectures".
Moreover, the researchers' work has not yet been subject to peer review. Indeed, power consumption on its own matters less than energy usage per unit of ‘output' – information that is not provided in the research paper.
![]()
Researchers claim new approach to LLMs could slash AI power consumption
ChatGPT uses 10 times the electricity of a Google search – but cutting out ‘matrix multiplication’ could cut this number without affecting performancewww.computing.co.uk
Scalable MatMul-free Language Modeling
Matrix multiplication (MatMul) typically dominates the overall computational cost of large language models (LLMs). This cost only grows as LLMs scale to larger embedding dimensions and context lengths. In this work, we show that MatMul operations can be completely eliminated from LLMs while...arxiv.org
Looks like Pedro is enjoying his new toy so far![]()
LinkedIn Login, Sign in | LinkedIn
Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.www.linkedin.com
Pedro Machado, a Senior Lecturer in Computer Science at Nottingham Trent University must have been a little giddy and trembling with excitement while typing this LinkedIn post about his uni joining “the race of Neuromorphic Computing by being [one] of the pioneers accelerating SNNs on the Alkida’s Brainchip.”You gotta love his enthusiasm, though!
View attachment 64801
View attachment 64803
View attachment 64804
Akida will be useful for their ongoing project on “Position Aware Activity Recognition”:
View attachment 64805
Agree. 25 years of experience as a leading salesperson. I think shareholders can demand a better presentation than that.I think Kuchenbuch urgently needs to work on his presentation skills. There are courses for that.
Already his first slide and and confusing presentation made me shut it off.
We want the share price to go lower so we can buy back in. HeheheAgree. 25 years of experience as a leading salesperson. I think shareholders can demand a better presentation than that.
Bring Rob Telson back!
Still here, and ready to buy back in when it gets lower.
We want the share price to go lower so we can buy back in. Hehehe![]()
sarcasm....................You had so many chances to buy low and did not use them? And now you want the SP to fall again after it has recovered a bit?
How stupid can one be?
I am a stubbie short of a six pack.You had so many chances to buy low and did not use them? And now you want the SP to fall again after it has recovered a bit?
How stupid can one be?
Not down ramping at all.
I was annoyed because of empty promises made and lack of real commercial updates from the company.
Sold some weeks ago which has saved me a further 30-40% decline.
"Watching the financials" and ready to buy when things actually look positive on paper, but if there is SFA happening for the rest of the year, then I think this will go down to 10-15 cents. In that case I might buy back in depending on the "financials" that we were told to keep an eye on.
As always, Not advice. DYOR and of course "watch the financials".
So yeah, "watch the financials"