INSYDIUM’s Post

Modelithics is pleased to announce a strategic partnership with Signal Edge Solutions aimed at advancing #signalintegrity and #powerintegrity modeling. This partnership will result in new, robust model libraries that combine Modelithics’ established catalog of high frequency models, many of which are applicable to high-speed SI/PI applications, with customized SI/PI models from Signal Edge Solutions. Read more! https://lnkd.in/e3g__5h #modelithics #modelingexperts #engineeringsolutions #modeling #modelingandsimulation

Great news for those with long-term simulation experience, especially with MIMO: Recently, many performance optimizations have been made to the codebases that will be released as the 5G-LENA v3.1 and as the ns-3.42. These performance optimizations are important for the new full MIMO model introduced in nr-v3.0, allowing for greater scalability of complex simulations. In the following blog of 5G-LENA we explain the code analysis and recent performance optimizations. More to come, stay tunned! https://lnkd.in/d6qWUV5d

SOOTHSAYER 1.8 is out now. It's our second feature release of the year and now that we've retired the big ol' VM the download has shrunk from 12GB to 200MB, you can share your GPU and get more control over patching and data. New Features: 🏠 3D Multipath 🛰 Satellite coverage 🚁 General Purpose Model ✈ Flight path route analysis 🐜 Dozens of bugfixes - thanks for the reports. A feature summary and change log is on the blog here: https://lnkd.in/enf4xe_W Lastly, to counter the (half true) rumour that it takes days to install, we've broken protocol and put the 9 minute install video on youtube 😉 https://lnkd.in/e4Gh_tGA

Interesting time to see the rapid advancement in models that help address some of the fundamental challenges to fine-tuning LLMs.

💡 New Special Issue Online! Title: Advances in #NumericalModeling Based on #GraphTheory Deadline: 30 April 2025 Submit via this link: https://buff.ly/4eRixZX #numericalmethods #simulation #optimization #MDPIOpenAccess #ComSciMath_Mdpi

Nearly completed with H1 2024, and the third update to DEWESoftX 2024 was just released. This update brings key enhancements to the sequencer, Polygon, Modal Geometry, GigE cameras, and LIN Bus. If you own DEWESoft hardware, you can now enjoy these updates to DEWESoftX professional software and your owned plug-ins immediately, at no additional cost. #DEWESoftUSA #DEWESoftX #SoftwareUpdate

𝐀𝐝𝐚𝐦𝐬𝐨𝐧’𝐬 𝐀𝐫𝐫𝐚𝐲𝐈𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐜𝐞 𝐕𝟏.𝟐 𝐚𝐝𝐝𝐬 𝐒𝐮𝐛 𝐀𝐫𝐫𝐚𝐲 𝐎𝐩𝐭𝐢𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧, 𝐕𝐆𝐭 𝐚𝐧𝐝 𝐃𝐞𝐥𝐭𝐚 𝐓𝐢𝐦𝐞 𝐒𝐢𝐦𝐮𝐥𝐚𝐭𝐢𝐨𝐧 Adamson has announced the release of ArrayIntelligence V1.2, the latest version of our innovative loudspeaker design and control software. Key highlights of ArrayIntelligence V1.2 include: VGt Now Included: Adamson's new flagship line array, VGt, is now available in AI V1.2, allowing users to include it in simulations and evaluate the optimization feature. Sub Array Optimization: A new tool that simplifies the design and optimization of subwoofer arrays for superior sound coverage and impact. Delta Time Simulation: A feature that allows for precise timing adjustments and simulations, enhancing the accuracy and coherence of sound systems. Plus More.... Follow the link: https://lnkd.in/g_3tkKKB Lee Stevens Travis Taylor Neil Colliss Scott Mathison Paul Thomas Felipe Balbino Jess Bickmore-Hutt #adamson #teamadamson #arrayintelligence #update #V2

Switching to a new borehole data management includes migrating your historical borehole data. Use #RSLog latest OCR feature and $ave hundreds of hours. Contact us to see how easy you can import your PDFs to your RSLog account

Are you overwhelmed by the extensive borehole data you've accumulated over the years? 📃😰 Introducing RSLog's innovative AI-powered Optical Character Recognition (OCR) feature that transforms your paper borehole logs into easily accessible digital records within your RSLog account. Download a free trial today and see how RSLog can streamline your workflow: https://bit.ly/3YqmuNV #Rocscience #OCR #EngineeringInnovation #BoreholeData

Hello ✋, I am pleased to share that both Llama-3-8B-Instruct and Llama-3-8B, which I helped to tune, have been quantized and are now available in 4-bit format. This means we can now host Llama-3-8B using only 10 GB of RAM, a reduction of approximately 30 GB from the standard model. Both models can be found in the following links: #meta #llama3 #quantization #4bit 1) https://lnkd.in/es7jDBHr (<- Llama-3-8B-Instruct-4-bit) 2) https://lnkd.in/espq3S3Q (<- Llama-3-8B-4-bit)