The Embedded Rustacean’s Post

The Return of the Frame Pointers https://buff.ly/3IH5Lyz (via @brendangregg)

New #rust release is out, feature highlights in 1.78 : - Diagnostic attributes - Asserting unsafe preconditions - Deterministic realignment https://lnkd.in/devFRaTi

Using floating point operations in embedded applications is evil. And there is a poison for that: https://lnkd.in/g_y_wmg8

We just drop a new TornadoVM version, 1.0.8 with improvements and many fixes. This version expands the profiler of the LevelZero/SPIRV backend with power metrics, new API calls for log and debug execution plans, fixes for running on OSx 14.6 and more! Check it out! 🔗 https://lnkd.in/dPbFrabr #java #gpu #performance #fpga #graalvm #ai

Batch Inference with Qwen2 Vision LLM (Sparrow) I'm explaining several hints how to optimize Qwen2 Visual LLM performance for batch processing. Complete video: https://lnkd.in/dcaezjq9 Code: https://lnkd.in/eXF5MpRP Sparrow UI: https://lnkd.in/dPk-ux6K

Clever trick to use in your projects: #pragma GCC poison new delete There we go, we have the NSA requirements for memory management in C++ completed 🤣

How to handle missing data using the is.na() function in the R programming language: https://lnkd.in/ehkr3B5J #programming #database #businessanalyst

Modern C++ : are "smart pointers" actually smart? What's the deal with smart pointers? What problems do they solve? Are they actually smart? In this video, I'll talk about what problems a smart pointer solves, the types of smart pointers in C++, and the pro's and con's of the way they solve certain problems.

Typically, creating and using LLM's for specialized use cases need power hungry GPU hardware. So, this is interesting and significant progress. Bitnet.cpp, The inference framework for 1-bit LLM's which provides the ability to run 100B LLM's locally without GPU support. https://lnkd.in/gFY_5EAX LLM's on edge devices like Raspberry Pi might just be possible! Looks like this is worth exploring for different use cases.

How to Get Location Information from an IP Address Using IPinfo https://lnkd.in/eRt2GTBq