How POWER, SPARC, and AIX Dominated HPC Before Linux and x86 Took Over

·1 views

From the 1990s into the 2000s, high-performance computing relied heavily on RISC-based processors like IBM's POWER and Fujitsu's SPARC64, which were engineered specifically for floating-point performance, memory bandwidth, and large-scale numerical calculations. Commercial UNIX systems such as AIX were tightly integrated with the underlying hardware, bundling compilers, numerical libraries, parallel execution environments, and file systems into a cohesive package. IBM's SP systems, for instance, shipped with AIX alongside tools like ESSL, MPI, LoadLeveler, and GPFS, making the software inseparable from the platform. As commodity x86 hardware matured and Linux gained capability, the balance shifted due to lower costs and greater procurement flexibility rather than any sudden technical leap. The transition is best understood not as a move away from specialized hardware for its own sake, but as a change in which option represented the most rational choice for HPC workloads at a given time.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Why a Developer Deleted Clean AI-Generated Code to Build Something Truly Original

A software developer writing on DEV Community argues that AI's advantage over any individual is limited precisely to the boundaries of that person's own ignorance — no further. While acknowledging AI's superior recall of syntax, algorithms, and documented solutions, the developer contends that it can only remix what already exists and cannot originate a vision for something entirely new. The author deleted a cleaner, AI-style function and replaced it with a slower, idiosyncratic one because that imperfect version contained a behaviour no standard solution would produce — one central to the product's purpose. The piece frames AI's knowledge gaps not as a threat but as a personal to-do list, arguing that every subject learned deeply removes one more area where the tool holds an edge. Ultimately, the developer concludes that authorship and taste are inseparable from the product itself, and outsourcing key decisions — even to a better output — hollows out the work's identity.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

CF7 to Google Sheets Drops Submissions Due to Shared API Quota Across All Users

WordPress site owners reported that roughly one in three Contact Form 7 submissions never reached their Google Sheets spreadsheet, even though email notifications and local database entries were recorded correctly. The GSheetConnector support team confirmed the cause: the free version of the plugin routes all API calls through the developer's single Google Cloud project, meaning thousands of sites share one 300-requests-per-minute quota. When that shared quota is exhausted, Google returns a 429 error and the submission is silently discarded without any alert to the site owner. The problem worsens as the plugin gains more installations, since more sites compete for the same limited quota bucket. The recommended fix is for each site owner to register their own Google Cloud project and connect the plugin to personal API credentials, effectively isolating their quota from all other users.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Developer Builds PDF Chat API in One Day Using FastAPI, Gemini, and Qdrant

A developer built a fully functional PDF Chat API in a single day, allowing users to upload PDF documents and ask questions about their content using natural language. The system uses a Retrieval Augmented Generation (RAG) architecture, combining Google Gemini for embeddings and chat, Qdrant as a vector database, and LangChain for pipeline orchestration. During ingestion, PDF text is split into chunks, converted into 3072-dimensional vectors via Gemini, and stored in Qdrant; at query time, the most relevant chunks are retrieved and passed to Gemini to generate an answer. The project includes a REST API with authentication, a plain HTML/CSS web interface, and is deployable on Qdrant's free cloud tier. The developer plans to add multi-user support, Docker deployment, and compatibility with other document formats such as Word and Excel.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

AI Agents Can't Be Sued: Why Liability Trails Back to Human Operators

When an AI agent autonomously negotiates and signs a contract, it cannot be held legally liable because software has no legal personhood. Legal systems universally agree that liability travels back along the chain of delegation until it reaches a human or entity that can be sued or fined. The challenge becomes more complex in multi-hop deployments where a principal authorizes an agent, which in turn delegates to sub-agents and third-party tools. Determining liability requires identifying, at each step, who held authority, who had sufficient information, and who could have intervened to prevent a bad outcome. Operators must maintain verifiable, timestamped authorization trails — not just policy statements — to demonstrate that governance controls were actually enforced at the moment an agent acted.

0 comments Read more at DEV Community