#matrixmultiplication

A Novel Compiler Transformation for Fast Sparse Matrix Multiplication in GPUs

#CUDA #Compilers #Sparse #MatrixMultiplication

https://hgpu.org/?p=29951

“Beating NumPy’s Matrix Multiplication In 150 Lines Of C Code”, Aman Salykov (https://salykova.github.io/matmul-cpu).

Via HN: https://news.ycombinator.com/item?id=40870345

On Lobsters: https://lobste.rs/s/6cktqx/beating_numpy_s_matrix_multiplication

#C #MatrixMultiplication #Math #Performance #BLAS #LinearAlgebra #MatMul #Speed #NumPy #Optimization

Researchers upend AI status quo by eliminating matrix multiplication in LLMs - Enlarge / Illustration of a brain inside of a light bulb. (credit: Gett... - https://arstechnica.com/?p=2033314 #matrixmultiplication #machinelearning #googlegemini #ucsantacruz #matrixmath #chatgpt #ternary #biz⁢ #matmul #gpu #ai

How much can we gain from Tensor Kernel Fusion on GPUs?

#CUDA #MatrixMultiplication #LinearAlgebra #Performance

https://hgpu.org/?p=29255

Fast and Practical Strassen’s Matrix Multiplication using FPGAs

#OpenCL #FPGA #MatrixMultiplication #BLAS #LinearAlgebra #GEMM #Package

https://hgpu.org/?p=29241

Evaluation of computational and energy performance in matrix multiplication algorithms on CPU and GPU using MKL, cuBLAS and SYCL

#CUDA #SYCL #MKL #CUBLAS #MatrixMultiplication #LinearAlgebra #Performance #Package

https://hgpu.org/?p=29229

New Breakthrough Brings #MatrixMultiplication Closer to Ideal https://www.quantamagazine.org/new-breakthrough-brings-matrix-multiplication-closer-to-ideal-20240307/

By eliminating a hidden inefficiency, computer scientists Ran Duan, Renfei Zhou and Hongxun Wu have come up with a new way to multiply large matrices that’s faster than ever

#matrixmultiplication

https://www.quantamagazine.org/new-breakthrough-brings-matrix-multiplication-closer-to-ideal-20240307/

Matrix multiplication breakthrough could lead to faster, more efficient AI models - Enlarge / When you do math on a computer, you fly through a numerical t... - https://arstechnica.com/?p=2008905 #massachusettsinstituteoftechnology #matrixmultiplication #aiandtheenvironment #tsinghuauniversity #machinelearning #williamkuszmaul #quantamagazine #volkerstrassen #aiefficiency #lasermethod #openaisora #chatgpt #chatgtp #biz⁢ #zhou #ai

#DailyBloggingChallenge (137/200)

When looking at short term memory, a portion of it is the active working memory.

One way to look at the active working memory is like matrix multiplication.

Let's say one gets a 5 digit number. It is fairly simple to return that number in the same order as given, which would equate to the identity matrix.

If one had to invert the order of number, then the diagonal of the matrix would be flipped compared to the identity matrix, so equating to a negative determinant.

Now imagine if one had to do more complex calculations like ordering the months of the year alphabetically. Creating a matrix that transposes the months vector is initially not that simple, though once instantiated, it becomes quite simple to repeat.

Depending on how many row alterations are needed, starting from the identity matrix could be a way to quantify the complexity of the task.

#ShortTermMemory #ActiveWorkingMemory #memory #matrix #MatrixMultiplication #maths

#AI Reveals New Possibilities in #MatrixMultiplication.
Last month, a team at the artificial intelligence company #DeepMind showed how to tackle the problem from a new direction, reporting in a paper in Nature that they’d successfully trained a neural network to discover new fast algorithms for matrix multiplication. It was as if the #AI had found an unknown strategy for solving a monstrously complex Rubik’s Cube.

“It’s a very neat result,” said #JoshAlman, a computer scientist at #ColumbiaUniversity. But he and other matrix multiplication specialists also emphasized that such #AI assistance will complement rather than replace existing methods — at least in the near term. “It’s like a proof of concept for something that could become a breakthrough,” Alman said. The result will simply help researchers on their quest.

When ever I see a cool new result in #ComputationalMath, I like to see if I can replicate it. So, last month when that Nature article came out about #MatrixMultiplication formulas from #AlphaTensor I set out see if I could get their formulas and verify them symbolically.
I was able to do that and of course they were right. But I was excited to see Kauers and Moosbauer publish a response a couple days later. So, here's their results replicated in a Maple Jupyter notebook https://github.com/johnpmay/MapleSnippets/blob/main/KMtoFFM.ipynb

DeepMind breaks 50-year math record using AI; new record falls a week later

https://arstechnica.com/?p=1887641

#matrixmultiplication #machinelearning #AlpheTensor #AlphaZero #deepmind #alphago #Biz&IT #AI