Nvidia discusses their lossless delta color compression is on its fourth generation in Pascal, and that it transparently saves bandwidth behind the scenes for rendering workloads. AMD has a similar technology, discussed in detail here: https://gpuopen.com/dcc-overview/

Do the same limitations apply to Nvidia's compression? They don't seem to discuss it in context of developer guidance.

  • $\begingroup$ As a designer of a GPU lossless image compression system, most of these seem to make sense, with the possible exception of using 32-bit float over 16-bit. Surely the LSBs of floating point data become increasingly less coherent and thus harder to compress. $\endgroup$
    – Simon F
    Jan 29 '18 at 9:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.