<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Rupesh Raghavan</title><description>A working log of things I build: llm, embedded systems, local AI, robotics and the Linux servers that run them.</description><link>https://rupeshtr78.github.io/blog/</link><item><title>How to read language-model evaluation scores</title><link>https://rupeshtr78.github.io/blog/notes/llm-evaluation-metrics/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/llm-evaluation-metrics/</guid><description>A plain-English guide to held-out loss, nats, bits per byte, perplexity, and the PG-19 evaluation numbers from ghedi-c.</description><pubDate>Sat, 14 Nov 2026 00:00:00 GMT</pubDate></item><item><title>Gpu Unlock My RTX 3090 was doing tensor math at half speed</title><link>https://rupeshtr78.github.io/blog/notes/gpu-unlock-half-tensor-rate/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/gpu-unlock-half-tensor-rate/</guid><description>Gpu unlock results, NVIDIA caps FP32-accumulate tensor math on GeForce cards. I removed the cap on a 3090, measured it, and got 3 percent instead of the 60 the box promises.</description><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate></item><item><title>Real prose, and what is left to build</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-15-real-prose-and-what-is-left/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-15-real-prose-and-what-is-left/</guid><description>Part 15: measuring the same model on text people actually read, then the honest map of what this series did not reach.</description><pubDate>Thu, 12 Nov 2026 00:00:00 GMT</pubDate></item><item><title>Teaching it to call tools, and the form it fills wrongly</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-14-teaching-it-to-call-tools/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-14-teaching-it-to-call-tools/</guid><description>Part 14: a chatml-tools template, a JSON logit mask with a forced opener and a bounded agent loop, plus the held-out table that says the mechanism works and the model does not.</description><pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate></item><item><title>Long context without retraining, and the null result</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-13-long-context-without-retraining/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-13-long-context-without-retraining/</guid><description>Part 13: serving a 512-token model at 4,096 positions with four angle-scaling modes, and finding that every published trick made it worse.</description><pubDate>Fri, 09 Oct 2026 00:00:00 GMT</pubDate></item><item><title>Feeding it two gigabytes</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-12-feeding-it-two-gigabytes/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-12-feeding-it-two-gigabytes/</guid><description>Part 12: making the runtime load the full 2.08 GiB TinyStories corpus, and learning that data scale beats model scale at this size.</description><pubDate>Wed, 16 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Seeing inside the GPU while it runs</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-11-seeing-inside-the-gpu/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-11-seeing-inside-the-gpu/</guid><description>Part 11: live device metrics, Prometheus and Grafana, NVTX ranges and an in-process CUPTI subscriber, all without changing the default build.</description><pubDate>Fri, 14 Aug 2026 00:00:00 GMT</pubDate></item><item><title>Fusing to parity, and hitting one kernel</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-10-fusing-to-parity/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-10-fusing-to-parity/</guid><description>Part 10: a device-resident workspace and three fused operations took inference from 4.0 ms to 2.52 ms per token, exactly level with the CPU.</description><pubDate>Wed, 08 Jul 2026 00:00:00 GMT</pubDate></item><item><title>Rotary embeddings: an honest no-win</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-09-rotary-the-honest-no-win/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-09-rotary-the-honest-no-win/</guid><description>Part 9: replacing the learned position table with rotary encoding, proving it with a mathematical oracle, and shipping it despite no loss improvement at the small size.</description><pubDate>Wed, 10 Jun 2026 00:00:00 GMT</pubDate></item><item><title>Subwords beat characters, and the tokenizer is a quality lever</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-08-subwords-beat-characters/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-08-subwords-beat-characters/</guid><description>Part 8: byte-level BPE, then GPT-2&apos;s real vocabulary and SentencePiece, and the runs that showed compression moves quality more than architecture does.</description><pubDate>Mon, 18 May 2026 00:00:00 GMT</pubDate></item><item><title>The first jump: ESP32-S3 startup and memory, simply</title><link>https://rupeshtr78.github.io/blog/notes/esp32s3-baremetal-memory-map/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/esp32s3-baremetal-memory-map/</guid><description>A short guide to the ESP32-S3 memory used by a bare-metal program: where its code, data, and stack go, and why two addresses can name the same RAM.</description><pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate></item><item><title>ESP32-S3 start.S and linker.ld, line by line</title><link>https://rupeshtr78.github.io/blog/notes/esp32s3-startup-linker-walkthrough/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/esp32s3-startup-linker-walkthrough/</guid><description>A guided reading of the startup assembly and linker script used by the ESP32-S3 bare-metal example, with each line tied back to the memory map.</description><pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate></item><item><title>Batching the window: 547 ms to 2 ms</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-07-batching-the-window/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-07-batching-the-window/</guid><description>Part 7: turning a training window into matrix problems across all positions, cutting a step by about 270x with an identical loss curve.</description><pubDate>Sun, 26 Apr 2026 00:00:00 GMT</pubDate></item><item><title>Training on the GPU: 762,936 launches per step</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-06-training-on-the-gpu/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-06-training-on-the-gpu/</guid><description>Part 6: moving backward propagation and Adam onto the device, matching the CPU to 4e-05, and finding the step was launch-bound rather than compute-bound.</description><pubDate>Sat, 04 Apr 2026 00:00:00 GMT</pubDate></item><item><title>The GPU is 4.5x slower, and the profile says why</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-05-the-gpu-is-slower/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-05-the-gpu-is-slower/</guid><description>Part 5: putting the chat loop on CUDA, measuring it honestly, and finding that 84% of the API time was spent shuttling activations across the bus.</description><pubDate>Thu, 12 Mar 2026 00:00:00 GMT</pubDate></item><item><title>Eight threads on an 82-SM GPU: my first kernels</title><link>https://rupeshtr78.github.io/blog/notes/gpu-kernels-first-lessons/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/gpu-kernels-first-lessons/</guid><description>Six small CUDA kernels, from hello world to a tiled matrix multiply, and the Nsight report that showed how little of the RTX 3090 I was actually using.</description><pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate></item><item><title>Tracing the game, scraping the model</title><link>https://rupeshtr78.github.io/blog/notes/thuruppu-04-otel-traces-vllm-metrics/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/thuruppu-04-otel-traces-vllm-metrics/</guid><description>Part 4: how Thuruppu emits OpenTelemetry traces, how the Collector sends them to Jaeger, and how Prometheus reads vLLM metrics.</description><pubDate>Sat, 28 Feb 2026 00:00:00 GMT</pubDate></item><item><title>First CUDA kernels: correct before fast</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-04-first-cuda-kernels/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-04-first-cuda-kernels/</guid><description>Part 4: bringing up one GPU inference operation at a time, matching the CPU to 1e-4, and discovering the first GPU build was 4.9x slower.</description><pubDate>Sun, 22 Feb 2026 00:00:00 GMT</pubDate></item><item><title>Two serving stacks, one card game</title><link>https://rupeshtr78.github.io/blog/notes/thuruppu-03-vllm-sglang-serving/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/thuruppu-03-vllm-sglang-serving/</guid><description>Part 3: the practical differences between running Thuruppu&apos;s local model endpoint with SGLang or vLLM.</description><pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate></item><item><title>The repeated prompt at the card table</title><link>https://rupeshtr78.github.io/blog/notes/thuruppu-02-kv-cache-prefix-caching/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/thuruppu-02-kv-cache-prefix-caching/</guid><description>Part 2: what KV cache and prefix caching can reuse in Thuruppu&apos;s repeated agent prompts, and what the repository does not measure.</description><pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate></item><item><title>A device boundary, before any device</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-03-a-device-boundary/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-03-a-device-boundary/</guid><description>Part 3: moving every buffer and operation behind an interface while the CPU still did all the math, so CUDA could land later without touching the model.</description><pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate></item><item><title>Three agents at one card table</title><link>https://rupeshtr78.github.io/blog/notes/thuruppu-01-agents-at-the-card-table/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/thuruppu-01-agents-at-the-card-table/</guid><description>Part 1: how I model the Twenty-eight players as seat-specific agents, and why the LLM does not make every card-play decision.</description><pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate></item><item><title>The first vertical slice: a terminal chat</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-02-first-vertical-slice-chat/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-02-first-vertical-slice-chat/</guid><description>Part 2: building the whole chat loop on the CPU before any GPU work, so the destination was defined before the road was built.</description><pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate></item><item><title>Ray on Kind: GPU plumbing is the hard part</title><link>https://rupeshtr78.github.io/blog/notes/ray-kind-gpu-cluster/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ray-kind-gpu-cluster/</guid><description>A local GPU-enabled Ray cluster crosses several boundaries: host devices, Kind nodes, Kubernetes scheduling, KubeRay, and model-cache storage.</description><pubDate>Thu, 15 Jan 2026 00:00:00 GMT</pubDate></item><item><title>A GPT in C, end to end</title><link>https://rupeshtr78.github.io/blog/notes/ghedi-01-a-gpt-in-c-end-to-end/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/ghedi-01-a-gpt-in-c-end-to-end/</guid><description>Part 1: a dependency-free character-level GPT in C that trains on a text file and generates names — made reproducible before it was made fast.</description><pubDate>Mon, 05 Jan 2026 00:00:00 GMT</pubDate></item><item><title>Watching the server without drowning in it</title><link>https://rupeshtr78.github.io/blog/notes/home-server-13-monitoring-the-server/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-13-monitoring-the-server/</guid><description>Part 13: Prometheus, node and GPU exporters, Glance and cAdvisor, and choosing the one GPU signal that stayed useful during a freeze.</description><pubDate>Wed, 03 Dec 2025 00:00:00 GMT</pubDate></item><item><title>DNS failed while Wi-Fi was still connected</title><link>https://rupeshtr78.github.io/blog/notes/home-server-12-dns-fails-silently/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-12-dns-fails-silently/</guid><description>Part 12: systemd-resolved fell back from EDNS0, Pi-hole refused a noisy client&apos;s queries, and the router kept resolving without filtering.</description><pubDate>Mon, 17 Nov 2025 00:00:00 GMT</pubDate></item><item><title>Pi-hole became the DNS backbone</title><link>https://rupeshtr78.github.io/blog/notes/home-server-11-pihole-dns-backbone/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-11-pihole-dns-backbone/</guid><description>Part 11: a Raspberry Pi at 10.0.0.42 filters DNS for the workstation, until one noisy client hit the per-client query limit.</description><pubDate>Sun, 02 Nov 2025 00:00:00 GMT</pubDate></item><item><title>Why the Tesla P40 stayed out of this build</title><link>https://rupeshtr78.github.io/blog/notes/home-server-10-why-the-p40-stayed-out/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-10-why-the-p40-stayed-out/</guid><description>Part 10: 24 GiB is not the whole story; Pascal lacks tensor cores, modern inference stacks draw a compute-capability line, and the driver branches conflict.</description><pubDate>Fri, 17 Oct 2025 00:00:00 GMT</pubDate></item><item><title>Docker and GPUs do not mix by accident</title><link>https://rupeshtr78.github.io/blog/notes/home-server-09-docker-and-gpus/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-09-docker-and-gpus/</guid><description>Part 9: a metrics container bound /dev/nvidia-uvm directly, held a module reference, and made the suspend guard&apos;s job harder.</description><pubDate>Wed, 01 Oct 2025 00:00:00 GMT</pubDate></item><item><title>Adding a third GPU moved the target</title><link>https://rupeshtr78.github.io/blog/notes/home-server-08-third-gpu-changed-indices/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-08-third-gpu-changed-indices/</guid><description>Part 8: the Quadro P400 changed PCI addresses and indices, and one stale GGML_VK_VISIBLE_DEVICES setting quietly moved transcription onto the wrong card.</description><pubDate>Wed, 17 Sep 2025 00:00:00 GMT</pubDate></item><item><title>The suspend that never returned</title><link>https://rupeshtr78.github.io/blog/notes/home-server-07-suspend-never-returned/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-07-suspend-never-returned/</guid><description>Part 7: an NVIDIA lock held through suspend, a compositor stuck in the display path, and the cold-suspend mitigation that finally woke cleanly.</description><pubDate>Tue, 02 Sep 2025 00:00:00 GMT</pubDate></item><item><title>When nvidia-smi stops answering</title><link>https://rupeshtr78.github.io/blog/notes/home-server-06-nvidia-smi-hangs/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-06-nvidia-smi-hangs/</guid><description>Part 6: the 12-second timeout, the clean-looking kernel log, and two GPU exporters spinning at 100 percent after resume.</description><pubDate>Mon, 18 Aug 2025 00:00:00 GMT</pubDate></item><item><title>Docker became the default home for services</title><link>https://rupeshtr78.github.io/blog/notes/home-server-05-docker-default-home/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-05-docker-default-home/</guid><description>Part 5: forty Compose folders, a smaller running set, and why a port map and a clear boundary matter more than another dashboard.</description><pubDate>Fri, 01 Aug 2025 00:00:00 GMT</pubDate></item><item><title>Ubuntu 24.04 on server hardware</title><link>https://rupeshtr78.github.io/blog/notes/home-server-04-ubuntu-kernels-drivers/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-04-ubuntu-kernels-drivers/</guid><description>Part 4: a kernel update is also an NVIDIA module update, the three retained kernels, and why I checked DKMS before rebooting.</description><pubDate>Thu, 17 Jul 2025 00:00:00 GMT</pubDate></item><item><title>The 12 GiB card, the 24 GiB card, and the index</title><link>https://rupeshtr78.github.io/blog/notes/home-server-03-split-and-index/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-03-split-and-index/</guid><description>Part 3: the measured LM Studio split, model size versus context memory, and why an app&apos;s device index can pick the wrong card.</description><pubDate>Wed, 02 Jul 2025 00:00:00 GMT</pubDate></item><item><title>Two GPUs, two jobs</title><link>https://rupeshtr78.github.io/blog/notes/home-server-02-two-gpus-two-jobs/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-02-two-gpus-two-jobs/</guid><description>The 3090 for LLM work and the 3060 for display and gaming: the 24 GiB and 12 GiB split, the role split, and why the 3090 stays headless.</description><pubDate>Wed, 18 Jun 2025 00:00:00 GMT</pubDate></item><item><title>ThinkStation P920 Server</title><link>https://rupeshtr78.github.io/blog/notes/home-server-01-the-machine/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/home-server-01-the-machine/</guid><description>Part 1: a dual Xeon workstation that became the home server, the disk split, and the three choices that shaped everything that follows.</description><pubDate>Wed, 04 Jun 2025 00:00:00 GMT</pubDate></item><item><title>DHT11: a protocol built out of delays</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-12-dht11/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-12-dht11/</guid><description>Part 12: reading a temperature sensor with nothing but pin toggling and microsecond timing, and the comment in the source that has it backwards.</description><pubDate>Mon, 19 May 2025 00:00:00 GMT</pubDate></item><item><title>I2C: two wires and a roll call</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-11-i2c/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-11-i2c/</guid><description>Part 11: a shared bus with addresses and acknowledgements, a repeated start, and a driver that reads an MPU6050 over it.</description><pubDate>Sat, 10 May 2025 00:00:00 GMT</pubDate></item><item><title>SPI in practice: a 25LC512 EEPROM driver</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-10-eeprom/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-10-eeprom/</guid><description>Part 10: writing a real SPI device driver, with write-enable, status polling and the page boundary that silently wraps your data.</description><pubDate>Thu, 01 May 2025 00:00:00 GMT</pubDate></item><item><title>What a logic analyser shows on an SPI bus</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-09-logic-analyzer/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-09-logic-analyzer/</guid><description>Part 9: reading a real capture of two chips talking, from the chip-select frame to the byte the slave sends one transaction late.</description><pubDate>Sat, 19 Apr 2025 00:00:00 GMT</pubDate></item><item><title>SPI: two chips passing notes</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-08-spi/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-08-spi/</guid><description>Part 8: a four-wire bus with a shared clock, two ATmega328Ps exchanging bytes, and the reason a send is always also a receive.</description><pubDate>Thu, 10 Apr 2025 00:00:00 GMT</pubDate></item><item><title>Servos: when pulse width is a position</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-07-servo/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-07-servo/</guid><description>Part 7: a 50 Hz signal where the width of each pulse is the angle, and the power wiring that decides whether it moves or jitters.</description><pubDate>Tue, 01 Apr 2025 00:00:00 GMT</pubDate></item><item><title>PWM: switching too fast to see</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-06-pwm/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-06-pwm/</guid><description>Part 6: using Timer1 to make a pin flicker faster than the eye can follow, and letting a potentiometer set the duty cycle.</description><pubDate>Wed, 19 Mar 2025 00:00:00 GMT</pubDate></item><item><title>Timers: accurate time without delay loops</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-05-timers/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-05-timers/</guid><description>Part 5: the chip&apos;s built-in counters, CTC versus overflow, and a watchdog that resets the chip when the program hangs.</description><pubDate>Mon, 10 Mar 2025 00:00:00 GMT</pubDate></item><item><title>Interrupts: two ways to let hardware call your code</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-04-interrupts/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-04-interrupts/</guid><description>Part 4: a button on INT0 and the same button on a pin-change interrupt, the rules an ISR has to follow, and the one line people forget.</description><pubDate>Sat, 01 Mar 2025 00:00:00 GMT</pubDate></item><item><title>ADC: turning a voltage into a number</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-03-adc/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-03-adc/</guid><description>Part 3: reading a potentiometer with the ATmega328P&apos;s 10-bit converter, and the prescaler setting that decides how accurate the answer is.</description><pubDate>Wed, 19 Feb 2025 00:00:00 GMT</pubDate></item><item><title>UART: giving the chip a voice on one wire</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-02-uart/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-02-uart/</guid><description>Part 2: a serial channel out of the chip, the four settings both ends must agree on, and what a baud mismatch looks like on an analyser.</description><pubDate>Mon, 10 Feb 2025 00:00:00 GMT</pubDate></item><item><title>Bare metal on the ATmega328P: pins, fuses and a first blink</title><link>https://rupeshtr78.github.io/blog/notes/atmega328p-01-pins-fuses-blink/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/atmega328p-01-pins-fuses-blink/</guid><description>Part 1: making a pin an output, and the fuse setting that decides whether the chip runs at 16 MHz, at 1 MHz, or not at all.</description><pubDate>Sat, 01 Feb 2025 00:00:00 GMT</pubDate></item><item><title>Wake on Lan Proxy server for triggering remote wakes</title><link>https://rupeshtr78.github.io/blog/notes/wol-proxy-remote-wake/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/wol-proxy-remote-wake/</guid><description>A small Rust proxy on a Raspberry Pi that wakes the home music server from a phone, and the magic packet, the signed cookie and the JSON field that silently did nothing.</description><pubDate>Wed, 15 Jan 2025 00:00:00 GMT</pubDate></item><item><title>Seven logic gates on a breadboard, and the one wired differently</title><link>https://rupeshtr78.github.io/blog/notes/seven-logic-gates/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/seven-logic-gates/</guid><description>Working through the 74HC family one chip at a time, from NOT to NAND, and the NOR chip whose pinout is not like the others.</description><pubDate>Thu, 01 Aug 2024 00:00:00 GMT</pubDate></item><item><title>Kernel modules: no main(), and a thread bug that hung the load</title><link>https://rupeshtr78.github.io/blog/notes/kernel-modules-first-lessons/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/notes/kernel-modules-first-lessons/</guid><description>Writing a first Linux kernel module, adding parameters and kthreads, and the mistakes that taught me how the kernel expects a driver to behave.</description><pubDate>Fri, 01 Dec 2023 00:00:00 GMT</pubDate></item><item><title>AWS Big Data IOT With Spark Streaming</title><link>https://rupeshtr78.github.io/blog/archive/aws-bigdata-iot/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/aws-bigdata-iot/</guid><description>Business case :Use AWS Big data services to process streaming data from IOT device.</description><pubDate>Sat, 30 Jan 2021 00:00:00 GMT</pubDate></item><item><title>Airflow Pipeline Kafka Cassandra</title><link>https://rupeshtr78.github.io/blog/archive/airflow-pipeline/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/airflow-pipeline/</guid><description>Business case : Create Airflow Pipeline using Kafka for Realtime stock price streaming, Cassandra for Realtime stock data warehousing Real time stock price can be streamed from alphavantage.com(subscription required)....</description><pubDate>Wed, 01 Jul 2020 00:00:00 GMT</pubDate></item><item><title>MQTT Spark Streaming With Watermark Part 2</title><link>https://rupeshtr78.github.io/blog/archive/mqtt-spark-streaming/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/mqtt-spark-streaming/</guid><description>Earlier we created a IOT Device / MQTT Client using Raspberry Pi.</description><pubDate>Sat, 02 May 2020 00:00:00 GMT</pubDate></item><item><title>MQTT Publish Subscribe IOT Part 1</title><link>https://rupeshtr78.github.io/blog/archive/mqtt-iot-pub-sub/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/mqtt-iot-pub-sub/</guid><description>We will create a IOT Device / MQTT Client using Raspberry Pi.</description><pubDate>Fri, 01 May 2020 00:00:00 GMT</pubDate></item><item><title>Custom MQTT Spark Receiver</title><link>https://rupeshtr78.github.io/blog/archive/spark-custom-receiver/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/spark-custom-receiver/</guid><description>Spark Streaming’s Receivers accept data in parallel and buffer it in the memory of Spark’s workers nodes.</description><pubDate>Wed, 15 Apr 2020 00:00:00 GMT</pubDate></item><item><title>Spark SQL Hive</title><link>https://rupeshtr78.github.io/blog/archive/spark-hdfs-hive/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/spark-hdfs-hive/</guid><description>In this post we will explore Spark SQL reading processing and writing data stored in Apache Hive.</description><pubDate>Tue, 14 Apr 2020 00:00:00 GMT</pubDate></item><item><title>Spark Structured Streaming Integrate Kafka Streams to Cassandra Sink</title><link>https://rupeshtr78.github.io/blog/archive/spark-structured-kafka-cassandra/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/spark-structured-kafka-cassandra/</guid><description>Kafka Streams integrated with Cassandra Sink using Spark Structured Streaming.</description><pubDate>Tue, 14 Apr 2020 00:00:00 GMT</pubDate></item><item><title>Kafka Streams</title><link>https://rupeshtr78.github.io/blog/archive/kafka-streams/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/kafka-streams/</guid><description>kafka Stream Word Count.</description><pubDate>Sat, 04 Apr 2020 00:00:00 GMT</pubDate></item><item><title>Inject Spark Stream to ElastiSearch Sink With Kibana</title><link>https://rupeshtr78.github.io/blog/archive/elastisearch-spark/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/elastisearch-spark/</guid><description>Elastisearch is a distributed document search and analytics engine, real-time search.</description><pubDate>Mon, 16 Mar 2020 00:00:00 GMT</pubDate></item><item><title>AWS EMR Spark</title><link>https://rupeshtr78.github.io/blog/archive/spark-aws-emr/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/spark-aws-emr/</guid><description>Steps for Deploying Spark App on Amazon EMR Make sure there are no paths to your local filesystem used in your script!</description><pubDate>Sun, 15 Mar 2020 00:00:00 GMT</pubDate></item><item><title>Cassandra</title><link>https://rupeshtr78.github.io/blog/archive/cassandra/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/cassandra/</guid><description>Setup Multi-node Cassandra cluster on a single machine We will create three node Cassandra cluster on single local machine.</description><pubDate>Wed, 11 Mar 2020 00:00:00 GMT</pubDate></item><item><title>Hue Oozie Hdfs</title><link>https://rupeshtr78.github.io/blog/archive/hue-oozie-hive/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/hue-oozie-hive/</guid><description>Hue provides UI for main hadoop features.</description><pubDate>Mon, 09 Mar 2020 00:00:00 GMT</pubDate></item><item><title>Flink Streaming</title><link>https://rupeshtr78.github.io/blog/archive/flink-streaming/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/flink-streaming/</guid><description>Flink is a stream processing engine.</description><pubDate>Fri, 21 Feb 2020 00:00:00 GMT</pubDate></item><item><title>Presto Hive Cassandra</title><link>https://rupeshtr78.github.io/blog/archive/presto-hive-cassandra/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/presto-hive-cassandra/</guid><description>Presto can connect to many different “big data” databases and data stores at once, and query across them using SQL syntax.</description><pubDate>Wed, 12 Feb 2020 00:00:00 GMT</pubDate></item><item><title>Spark PostGres Cassandra MongoDb</title><link>https://rupeshtr78.github.io/blog/archive/spark-postgres-mongodb-cassandra/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/spark-postgres-mongodb-cassandra/</guid><description>Integrate spark with Postgres, Cassandra,MongoDb Explore how to write spark streaming data into databases.</description><pubDate>Mon, 03 Feb 2020 00:00:00 GMT</pubDate></item><item><title>Twitter Custom Spark Receiver</title><link>https://rupeshtr78.github.io/blog/archive/twitter-custom-receiver/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/twitter-custom-receiver/</guid><description>Create your twitter feed filtering the handles you follow and also filtering out retweets by those handles using custom Spark receiver.</description><pubDate>Sat, 11 Jan 2020 00:00:00 GMT</pubDate></item><item><title>Spark Partitioning</title><link>https://rupeshtr78.github.io/blog/archive/spark-partitioning/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/spark-partitioning/</guid><description>Spark isn’t totally magic – you need to think about how your data is partitioned Operations that are expensive, and Spark won’t distribute data on its own.</description><pubDate>Wed, 01 Jan 2020 00:00:00 GMT</pubDate></item><item><title>Kafka Cluster</title><link>https://rupeshtr78.github.io/blog/archive/kafka-zookeeper/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/kafka-zookeeper/</guid><description>- Kafka is simply a collection of topics split into one or more partitions.</description><pubDate>Tue, 03 Dec 2019 00:00:00 GMT</pubDate></item><item><title>Understanding Consensus algorithms in Distributed Systems</title><link>https://rupeshtr78.github.io/blog/archive/raft-leader-election/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/raft-leader-election/</guid><description>Consensus Algorithms Consensus algorithms allow a collection of machines to work as a coherent group that can survive the failures of some of its members.</description><pubDate>Mon, 11 Nov 2019 00:00:00 GMT</pubDate></item><item><title>Kafka Zookeeper Cluster Leader Election</title><link>https://rupeshtr78.github.io/blog/archive/kafka-docker-zookeeper/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/kafka-docker-zookeeper/</guid><description>This article will configure kafka zookeeper cluster and explore the leader election process.</description><pubDate>Wed, 02 Oct 2019 00:00:00 GMT</pubDate></item><item><title>Kubernetes Docker</title><link>https://rupeshtr78.github.io/blog/archive/kubernetes-docker-hyperledger/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/kubernetes-docker-hyperledger/</guid><description>The objective of this post is to implement a basic hyperledger Block Chain network on Kubernetes platform.</description><pubDate>Tue, 01 Oct 2019 00:00:00 GMT</pubDate></item><item><title>Technical Blogging</title><link>https://rupeshtr78.github.io/blog/archive/technical-blogging/</link><guid isPermaLink="true">https://rupeshtr78.github.io/blog/archive/technical-blogging/</guid><description>How I started.</description><pubDate>Tue, 22 Dec 2015 00:00:00 GMT</pubDate></item></channel></rss>