Adhyayan

Tuesday, December 16, 2025

Computer achitecture

 Advanced Computer architecture and systems

scuttleblurb nvda1 CPU to GPU


Running MCP server in cloud


Model flops utilization

HW for deep learning

Data gated conventional MAC shown in the video.

DEEP COMPRESSION: COMPRESSING DEEP NEURAL NETWORKS WITH PRUNING, TRAINED QUANTIZATION AND HUFFMAN CODING

deeplearningbook MIT

Posted by adhyayan at 2:46 AM
Email ThisBlogThis!Share to XShare to FacebookShare to Pinterest

No comments:

Post a Comment

Newer Post Older Post Home
Subscribe to: Post Comments (Atom)

Search This Blog

About Me

adhyayan
View my complete profile

Popular Posts

  • (no title)
    lladro  figurine
  • (no title)
    This book  Bill Gates sends ex-con to law school after supreme court win win_a_year_s_supply_what_does_that_mean
  • Advantages of Google books
    1. You can use them as inspiration for your own writing 2. You can make up stories for kids using it as a starting board.
  • Warning sign
    In 'hard target' movie, Jean Claude Van Damme bites off the rattle of the rattlesnake saying, " I will leave a little surprise ...
  • (no title)
    Title IX
  • Essays from 'Discovery book'
    'The Will'  by Anne C Blackford is a very short essay with a single scene of a lady dictating her will to a boy who has words COCA ...
  • Good example of a bad lecture
    Not only is the content obvious but trying to state 'what not to do' to take up literature as vocation has been written in a very un...
  • (no title)
    Nutone
  • Links
    On a birthday cake, I saw Growing old is mandatory, Growing up is optional -  Chili Davis . When I came across  “Reaching the summit is o...
  • A poem by its other name
    This poem with title 'race' had me guessing the end with some injustice. Race is such a high power word that it masks its homonymity...

Blog Archive

  • ►  2026 (35)
    • ►  January (35)
  • ▼  2025 (155)
    • ▼  December (73)
      • Counteractive RL vs. SEED RL and TIS: How Algorith...
      • Why TIS Matters in RL: Boosting Performance Despit...
      • TTT - Test time training
      • System Verilog as concurrent reactive system
      • What is projection?
      • Hardware of Primal Attention
      • Understanding Design and Parallel Programming Patt...
      • Cornell Courses
      • Parallel Processing: From Early Hardware to AI‑Dri...
      • High performer questions
      • Doctor's Appointment.
      • Deepseek V3 Deep dive
      • From Python to FPGA: Modern Toolchains for Machine...
      • quantbeckman Golden Python rules
      • Does Groq LPU not use systolic because it uses SRA...
      • A Weekly Routine for Long-Term Judgment
      • Deep dive of ACM paper - A domain specific archite...
      • QwenLong-L1.5 and the Return of Streaming Models
      • Inference Is the New Training
      • Transformers Broke the Hardware Model
      • The Memory Wall: Why Modern ML Is Bandwidth-Bound
      • Why GPUs Forced Deep Learning to Become Linear Alg...
      • How hardware ↔ ML co-evolved (the big picture)
      • TPU vs GPU difference at microarchitecture level
      • Where TPUs outperform GPUs
      • Why MoE breaks the systolic dream
      • From Algorithms to Accelerators: A Narrative of De...
      • Hardware for Deep Learning
      • QwenLong-L1.5
      • model inference optimization
      • Why Hardware People Should Still Read ML Textbooks
      • How a Hardware-Minded Reader Should Use d2l.ai
      • How to Learn about the hardware behind AI
      • Pretty pictures
      • Scale
      • What is a “kernel”?
      • Is matrix multiplication a Data flow?
      • Nvidia + Groq
      • How Attention Runs on Real Hardware
      • CMU Intro to Deep Learning
      • Deep Learning
      • TPU
      • Hardware for Deep Learning
      • In-Datacenter Performance Analysis of a Tensor Pro...
      • Modern AI Hardware: A Conceptual Primer for GPUs, ...
      • Chapter 7 Precision, Sparsity, and Energy: Why Spe...
      • Chapter 6 Scaling Many Chips: When Communication D...
      • Chapter 5 Scaling One Chip: Extracting Performance...
      • Chapter 4 TPUs: Dataflow Machines
      • Chapter 3 GPUs: Latency-Hiding Machines
      • Compute Is Cheap, Data Is Expensive
      • Chapter 2 Why Everything in Deep Learning Becomes ...
      • Chapter 1 The Roofline Model: The One Graph That E...
      • correctness checking of kernel with floats
      • Compute Is Cheap, Data Is Expensive - explore with...
      • Chapter 0 Why Deep Learning Needs Special hardware
      • Before the beginning
      • TPU
      • Drones
      • Matmul
      • Computer Systems architecture
      • Computer achitecture
      • How to think about GPUs
      • ML Resources
      • Good Staff Engineer
      • Hardware acceleration
      • Culinary interest for kids
      • Real Intent connectivity
      • DSL - Domain specific language from chatgpt
      • Scalable Collaborative zk-SNARK and Its Applicatio...
      • Structure for college essay from chatgpt
      • Example college essay from chatgpt
      • optimizing GPU kernels for high performance
    • ►  November (2)
    • ►  October (1)
    • ►  September (3)
    • ►  August (18)
    • ►  July (28)
    • ►  June (20)
    • ►  May (7)
    • ►  April (2)
    • ►  February (1)
  • ►  2024 (17)
    • ►  June (5)
    • ►  May (10)
    • ►  February (2)
  • ►  2023 (7)
    • ►  December (1)
    • ►  September (3)
    • ►  July (2)
    • ►  January (1)
  • ►  2022 (59)
    • ►  December (4)
    • ►  November (1)
    • ►  October (2)
    • ►  September (17)
    • ►  August (2)
    • ►  July (2)
    • ►  June (9)
    • ►  May (4)
    • ►  April (1)
    • ►  March (2)
    • ►  February (8)
    • ►  January (7)
  • ►  2021 (35)
    • ►  December (3)
    • ►  November (5)
    • ►  October (10)
    • ►  September (3)
    • ►  August (1)
    • ►  July (7)
    • ►  June (3)
    • ►  May (1)
    • ►  January (2)
  • ►  2020 (16)
    • ►  November (2)
    • ►  October (4)
    • ►  July (1)
    • ►  June (9)
  • ►  2019 (2)
    • ►  August (1)
    • ►  March (1)
  • ►  2017 (6)
    • ►  June (5)
    • ►  January (1)
  • ►  2016 (2)
    • ►  December (2)
  • ►  2013 (65)
    • ►  September (1)
    • ►  August (11)
    • ►  July (1)
    • ►  June (3)
    • ►  May (9)
    • ►  April (25)
    • ►  March (6)
    • ►  February (7)
    • ►  January (2)
  • ►  2012 (154)
    • ►  November (7)
    • ►  October (12)
    • ►  September (3)
    • ►  August (10)
    • ►  July (25)
    • ►  June (20)
    • ►  May (27)
    • ►  April (24)
    • ►  March (19)
    • ►  February (6)
    • ►  January (1)
  • ►  2011 (51)
    • ►  November (2)
    • ►  October (1)
    • ►  August (1)
    • ►  May (8)
    • ►  April (6)
    • ►  March (5)
    • ►  February (21)
    • ►  January (7)
  • ►  2010 (96)
    • ►  December (2)
    • ►  October (2)
    • ►  September (9)
    • ►  August (3)
    • ►  July (13)
    • ►  June (14)
    • ►  May (3)
    • ►  April (5)
    • ►  March (24)
    • ►  February (18)
    • ►  January (3)
  • ►  2009 (26)
    • ►  December (14)
    • ►  November (5)
    • ►  October (2)
    • ►  June (1)
    • ►  May (1)
    • ►  March (1)
    • ►  February (2)
  • ►  2008 (50)
    • ►  November (6)
    • ►  October (2)
    • ►  September (5)
    • ►  July (5)
    • ►  June (4)
    • ►  May (11)
    • ►  April (4)
    • ►  March (1)
    • ►  February (8)
    • ►  January (4)
  • ►  2007 (86)
    • ►  December (2)
    • ►  November (3)
    • ►  October (11)
    • ►  September (21)
    • ►  August (19)
    • ►  July (8)
    • ►  June (2)
    • ►  May (4)
    • ►  April (1)
    • ►  March (2)
    • ►  February (5)
    • ►  January (8)
  • ►  2006 (41)
    • ►  December (10)
    • ►  November (5)
    • ►  October (2)
    • ►  September (4)
    • ►  August (1)
    • ►  July (4)
    • ►  June (12)
    • ►  May (1)
    • ►  April (1)
    • ►  January (1)
  • ►  2005 (16)
    • ►  November (1)
    • ►  July (2)
    • ►  May (4)
    • ►  April (9)
Picture Window theme. Powered by Blogger.