cuda

• CUDA Reference
  • Documentation
  • Other
  • Courses
  • Tools
  • Related Projects
• CUDA Forums
  • Retina MBP
• CUDA Toolkit
  • Detect CUDA version
  • Detect CUDA capability
• CUDA Random numbers
• CUDA Atomic
  • atomic funcs in chroma
  • Old GPU capability compatibility code
• CUDA Capability
• CUDA Emulation
  • GPU Ocelot
  • Portland Group PGI x86 CUDA Compiler
• CUDA TOOLKIT
  • GPU Hardware Check
  • Versions
  • 4.1
    • N belle7 (SL 5.1 Boron)
• rCUDA
• CUDA in Cloud
  • Virtualization
• CUDA OSX Driver
  • January 2015
  • March 12, 2014
    • 5.5.47
  • Hardware
  • PyCUDA Debugging
  • System Preferences
  • Simultaneous Integrated and Discrete GPU Operation ?
    • Integrated
• CUDA GDB
  • Docs
  • Dr Dobbs
  • Courses
  • (Manual) Configurations for GPU Debugging
    • Single GPU
    • Multi-GPU Debugging with the Desktop Manager Running
      • On Linux
      • On Mac OS X
  • Extras
  • Debug Setup
  • g4daechroma debug
  • kernel pileup ?
  • even 5 is enough
  • actually 2 will do it
  • even single photon (the 2nd photon in mock007) will do it
  • skipping the photons with axis aligned directions succeeds to complete
  • push the work up to 628 hardly touches the time
  • gdb compilation fail
  • Debug Example
  • info cuda threads
  • kernel debug
  • PyCUDA Version
  • pudb : Console based python debugger
• CUDA Profiling
  • nvprof
  • cuda clock
  • gpu burn
  • how many registers
  • cuda profile
  • cuda_profile_parse.py
  • method
  • occupancy
  • cputime
  • gputime
  • config
  • metrics
  • events
• CUDA Tests
• CUDA Error Handling including timeouts
  • Handling / Reset ?
  • Compute and Graphics
  • CUDA Driver API Errors
  • deviceQuery
• CUDA envvar
  • CUDA_CACHE_DISABLE
  • CUDA_DEVICE_WAITS_ON_EXCEPTION
• OSX GPU Panic
• CUDA Memory Types
  • Constant
• CUDA Math
  • matrix passing
  • role your own matrix4
  • math libs with CUDA compatibility
    • glm : OpenGL Mathematics
    • CudaVec single header
• CUDA Context Cleanup
  • Running with OSX in normal GUI mode
  • Pragmatic Solution
  • OSX VRAM
  • Running without GUI using “>console” login
  • Stuck Python Process
  • Search
• CUDA Persistent Threads
• CUDA GPU Comparisons
  • Multi GPU scaling
  • NVIDIA GeForce GT 750M vs NVIDIA Tesla K20m
  • NVIDIA GeForce GT 750M
  • NVIDIA Tesla K20m
• texture
  • CUDA Texture Memory
    • Reference
    • Texture Objects