Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

CCSD_T2_7 GPU support #1029

Closed
wants to merge 113 commits into from
Closed

CCSD_T2_7 GPU support #1029

wants to merge 113 commits into from

Commits on Oct 5, 2023

  1. this works

    @jeffhammond
    jeffhammond committed Oct 5, 2023
    Configuration menu
    Copy the full SHA
    43ebd7f View commit details
    Browse the repository at this point in the history

  2. add DGEMM version too

    @jeffhammond
    jeffhammond committed Oct 5, 2023
    Configuration menu
    Copy the full SHA
    2fa9876 View commit details
    Browse the repository at this point in the history

  3. straight DGEMM works

    @jeffhammond
    jeffhammond committed Oct 5, 2023
    Configuration menu
    Copy the full SHA
    8062b6d View commit details
    Browse the repository at this point in the history

Commits on Oct 6, 2023

  1. remove the loops - DGEMM will always be better

    @jeffhammond
    jeffhammond committed Oct 6, 2023
    Configuration menu
    Copy the full SHA
    e80bac5 View commit details
    Browse the repository at this point in the history

  2. removing loops

    @jeffhammond
    jeffhammond committed Oct 6, 2023
    Configuration menu
    Copy the full SHA
    fc17d91 View commit details
    Browse the repository at this point in the history

  3. cleanup

    @jeffhammond
    jeffhammond committed Oct 6, 2023
    Configuration menu
    Copy the full SHA
    b3ce4c1 View commit details
    Browse the repository at this point in the history

  4. do the pure DGEMM T2_8 in ICSD/NTS too

    @jeffhammond
    jeffhammond committed Oct 6, 2023
    Configuration menu
    Copy the full SHA
    567fd44 View commit details
    Browse the repository at this point in the history

  5. move makefile include to the top so we can use its vars

    @jeffhammond
    jeffhammond committed Oct 6, 2023
    Configuration menu
    Copy the full SHA
    0a1b133 View commit details
    Browse the repository at this point in the history

Commits on Oct 10, 2023

  1. still debugging

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    6de9fbe View commit details
    Browse the repository at this point in the history

  2. so far, so good

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    e78fb8e View commit details
    Browse the repository at this point in the history

  3. so far, so good

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    9f12e90 View commit details
    Browse the repository at this point in the history

  4. okay, it works correctly now

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    409ba37 View commit details
    Browse the repository at this point in the history

  5. okay, it works correctly now

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    360e2e7 View commit details
    Browse the repository at this point in the history

  6. now time for double buffering

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    355a2c8 View commit details
    Browse the repository at this point in the history

  7. clean up

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    74ab406 View commit details
    Browse the repository at this point in the history

  8. arrays are column major. wow.

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    34242be View commit details
    Browse the repository at this point in the history

  9. n stream version using n=1

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    9335011 View commit details
    Browse the repository at this point in the history

  10. n stream version using n=1

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    2b18f0c View commit details
    Browse the repository at this point in the history

  11. 2 phase version is correct

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    12f47c9 View commit details
    Browse the repository at this point in the history

  12. comment syntax

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    7cf47de View commit details
    Browse the repository at this point in the history

  13. move T2_7 into separate file

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    aec8780 View commit details
    Browse the repository at this point in the history

  14. move the parent 7 term

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    d34e8a0 View commit details
    Browse the repository at this point in the history

  15. make it pretty

    @jeffhammond
    jeffhammond committed Oct 10, 2023
    Configuration menu
    Copy the full SHA
    62b5b91 View commit details
    Browse the repository at this point in the history

Commits on Oct 11, 2023

  1. more formatting

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    9ebe5ff View commit details
    Browse the repository at this point in the history

  2. more formatting

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    40d25c8 View commit details
    Browse the repository at this point in the history

  3. more formatting; code copy for _x version

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    a88ecc5 View commit details
    Browse the repository at this point in the history

  4. alloc hoist

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    bbb9b8b View commit details
    Browse the repository at this point in the history

  5. alloc reduction

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    35f5619 View commit details
    Browse the repository at this point in the history

  6. alloc reduction

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    307156f View commit details
    Browse the repository at this point in the history

  7. remove pointless DGEMM from ccsd_e

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    707c738 View commit details
    Browse the repository at this point in the history

  8. eliminate pointless call to sort

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    2df0127 View commit details
    Browse the repository at this point in the history

  9. eliminate pointless call to sort

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    b7e7054 View commit details
    Browse the repository at this point in the history

  10. eliminate dgemm in favor of loop

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    b4d79cd View commit details
    Browse the repository at this point in the history

  11. my test

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    315b62b View commit details
    Browse the repository at this point in the history

  12. fix non-F90 case

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    b437000 View commit details
    Browse the repository at this point in the history

  13. fix my dumb error in dot product

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    1ba9af4 View commit details
    Browse the repository at this point in the history

  14. Merge branch 'nwchemgit:master' into ccsd_t2_dgemm_cublas_7

    @jeffhammond
    jeffhammond authored Oct 11, 2023
    Configuration menu
    Copy the full SHA
    47aeddd View commit details
    Browse the repository at this point in the history

  15. eliminated the transpose from e+=T2*V2 finally

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    51447a3 View commit details
    Browse the repository at this point in the history

  16. remove pointless MA stack allocation for scalar

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    45fe7cb View commit details
    Browse the repository at this point in the history

  17. cleanup

    @jeffhammond
    jeffhammond committed Oct 11, 2023
    Configuration menu
    Copy the full SHA
    64325e9 View commit details
    Browse the repository at this point in the history

Commits on Oct 12, 2023

  1. works

    @jeffhammond
    jeffhammond committed Oct 12, 2023
    Configuration menu
    Copy the full SHA
    af9ead8 View commit details
    Browse the repository at this point in the history

  2. this works

    @jeffhammond
    jeffhammond committed Oct 12, 2023
    Configuration menu
    Copy the full SHA
    e9a6539 View commit details
    Browse the repository at this point in the history

  3. this works

    @jeffhammond
    jeffhammond committed Oct 12, 2023
    Configuration menu
    Copy the full SHA
    271169b View commit details
    Browse the repository at this point in the history

  4. this works

    @jeffhammond
    jeffhammond committed Oct 12, 2023
    Configuration menu
    Copy the full SHA
    713d5f0 View commit details
    Browse the repository at this point in the history

  5. this works

    @jeffhammond
    jeffhammond committed Oct 12, 2023
    Configuration menu
    Copy the full SHA
    2089b9a View commit details
    Browse the repository at this point in the history

  6. this works with stdpar=gpu

    @jeffhammond
    jeffhammond committed Oct 12, 2023
    Configuration menu
    Copy the full SHA
    7d8377a View commit details
    Browse the repository at this point in the history

Commits on Oct 17, 2023

  1. input

    @jeffhammond
    jeffhammond committed Oct 17, 2023
    Configuration menu
    Copy the full SHA
    b90bbff View commit details
    Browse the repository at this point in the history

Commits on Nov 6, 2023

  1. add CCSD term timers

    @jeffhammond
    jeffhammond committed Nov 6, 2023
    Configuration menu
    Copy the full SHA
    b773443 View commit details
    Browse the repository at this point in the history

  2. another T2_7 term runs on GPU

    @jeffhammond
    jeffhammond committed Nov 6, 2023
    Configuration menu
    Copy the full SHA
    fcd2a3d View commit details
    Browse the repository at this point in the history

Commits on Nov 7, 2023

  1. test

    @jeffhammond
    jeffhammond committed Nov 7, 2023
    Configuration menu
    Copy the full SHA
    f62cba3 View commit details
    Browse the repository at this point in the history

Commits on Nov 24, 2023

  1. fix misplaced preprocessing and add _y version

    @jeffhammond
    jeffhammond committed Nov 24, 2023
    Configuration menu
    Copy the full SHA
    6d51b8f View commit details
    Browse the repository at this point in the history

  2. stdpar sort

    @jeffhammond
    jeffhammond committed Nov 24, 2023
    Configuration menu
    Copy the full SHA
    a733194 View commit details
    Browse the repository at this point in the history

  3. GPU stdpar sort

    @jeffhammond
    jeffhammond committed Nov 24, 2023
    Configuration menu
    Copy the full SHA
    a883c9d View commit details
    Browse the repository at this point in the history

  4. massive cleanup of GPU T2 7

    @jeffhammond
    jeffhammond committed Nov 24, 2023
    Configuration menu
    Copy the full SHA
    8f1ddd0 View commit details
    Browse the repository at this point in the history

  5. cleanup T2 7 further

    @jeffhammond
    jeffhammond committed Nov 24, 2023
    Configuration menu
    Copy the full SHA
    7657146 View commit details
    Browse the repository at this point in the history

Commits on Jan 16, 2024

  1. performance

    @jeffhammond
    jeffhammond committed Jan 16, 2024
    Configuration menu
    Copy the full SHA
    b98edb2 View commit details
    Browse the repository at this point in the history

  2. preprocessor guard on CUBLAS module

    @jeffhammond
    jeffhammond committed Jan 16, 2024
    Configuration menu
    Copy the full SHA
    cefa173 View commit details
    Browse the repository at this point in the history

Commits on Oct 15, 2024

  1. move makefile include to the top so we can use its vars

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    775f9d1 View commit details
    Browse the repository at this point in the history

  2. still debugging

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    07d444c View commit details
    Browse the repository at this point in the history

  3. so far, so good

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    b6e8a9d View commit details
    Browse the repository at this point in the history

  4. so far, so good

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    ca86b66 View commit details
    Browse the repository at this point in the history

  5. okay, it works correctly now

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    8950e69 View commit details
    Browse the repository at this point in the history

  6. okay, it works correctly now

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    09f9729 View commit details
    Browse the repository at this point in the history

  7. now time for double buffering

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    f4ef36e View commit details
    Browse the repository at this point in the history

  8. clean up

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    db91ef7 View commit details
    Browse the repository at this point in the history

  9. arrays are column major. wow.

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    3b767d5 View commit details
    Browse the repository at this point in the history

  10. n stream version using n=1

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    be9d24d View commit details
    Browse the repository at this point in the history

  11. n stream version using n=1

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    79499fe View commit details
    Browse the repository at this point in the history

  12. 2 phase version is correct

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    241b9b1 View commit details
    Browse the repository at this point in the history

  13. comment syntax

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    4cd5ddc View commit details
    Browse the repository at this point in the history

  14. move T2_7 into separate file

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    7a8222c View commit details
    Browse the repository at this point in the history

  15. move T2_7 into separate file

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    25e7304 View commit details
    Browse the repository at this point in the history

  16. move the parent 7 term

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    24a594b View commit details
    Browse the repository at this point in the history

  17. make it pretty

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    f76c000 View commit details
    Browse the repository at this point in the history

  18. more formatting

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    84ead7d View commit details
    Browse the repository at this point in the history

  19. more formatting

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    61b26da View commit details
    Browse the repository at this point in the history

  20. more formatting; code copy for _x version

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    8d7dfb6 View commit details
    Browse the repository at this point in the history

  21. alloc hoist

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    d06d802 View commit details
    Browse the repository at this point in the history

  22. alloc reduction

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    7b3e483 View commit details
    Browse the repository at this point in the history

  23. alloc reduction

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    40f82f7 View commit details
    Browse the repository at this point in the history

  24. eliminate pointless call to sort

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    9a8bd1e View commit details
    Browse the repository at this point in the history

  25. eliminate pointless call to sort

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    098db98 View commit details
    Browse the repository at this point in the history

  26. my test

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    b365451 View commit details
    Browse the repository at this point in the history

  27. eliminated the transpose from e+=T2*V2 finally

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    5d5caaa View commit details
    Browse the repository at this point in the history

  28. cleanup

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    8c2f216 View commit details
    Browse the repository at this point in the history

  29. works

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    c597fba View commit details
    Browse the repository at this point in the history

  30. this works

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    6bc2012 View commit details
    Browse the repository at this point in the history

  31. this works

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    fc0a851 View commit details
    Browse the repository at this point in the history

  32. this works

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    c0b4a97 View commit details
    Browse the repository at this point in the history

  33. this works

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    b1b5eb5 View commit details
    Browse the repository at this point in the history

  34. this works with stdpar=gpu

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    58752ad View commit details
    Browse the repository at this point in the history

  35. input

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    fe7c87a View commit details
    Browse the repository at this point in the history

  36. add CCSD term timers

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    eed2c64 View commit details
    Browse the repository at this point in the history

  37. another T2_7 term runs on GPU

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    9954c5d View commit details
    Browse the repository at this point in the history

  38. test

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    c05af60 View commit details
    Browse the repository at this point in the history

  39. fix misplaced preprocessing and add _y version

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    29df905 View commit details
    Browse the repository at this point in the history

  40. stdpar sort

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    745c033 View commit details
    Browse the repository at this point in the history

  41. GPU stdpar sort

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    4676886 View commit details
    Browse the repository at this point in the history

  42. massive cleanup of GPU T2 7

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    ba2631d View commit details
    Browse the repository at this point in the history

  43. cleanup T2 7 further

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    534d78a View commit details
    Browse the repository at this point in the history

  44. performance

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    03b44bf View commit details
    Browse the repository at this point in the history

  45. preprocessor guard on CUBLAS module

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    ff20a9d View commit details
    Browse the repository at this point in the history

  46. reset generic input file

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    39070b5 View commit details
    Browse the repository at this point in the history

  47. Merge branch 'ccsd_t2_dgemm_cublas_7' of https://github.com/jeffhammo… 

    …nd/nwchem into ccsd_t2_dgemm_cublas_7
    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    ba3242e View commit details
    Browse the repository at this point in the history

  48. oops

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    2ddc447 View commit details
    Browse the repository at this point in the history

  49. add 2D sorts for GPU

    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    aee190d View commit details
    Browse the repository at this point in the history

  50. GPU version of ccsd_t2_7_2_x 

    this is an O(N^5) term, so it probably doesn't matter...

Signed-off-by: Jeff Hammond <[email protected]>
    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    7c1459f View commit details
    Browse the repository at this point in the history

  51. remove commented-out code 

    Signed-off-by: Jeff Hammond <[email protected]>
    @jeffhammond
    jeffhammond committed Oct 15, 2024
    Configuration menu
    Copy the full SHA
    c562fc8 View commit details
    Browse the repository at this point in the history

Commits on Oct 16, 2024

  1. add alg that does O(N^5) on CPU but other on GPU, which is slower

    @jeffhammond
    jeffhammond committed Oct 16, 2024
    Configuration menu
    Copy the full SHA
    f3b5539 View commit details
    Browse the repository at this point in the history

  2. reset

    @jeffhammond
    jeffhammond committed Oct 16, 2024
    Configuration menu
    Copy the full SHA
    2904ac6 View commit details
    Browse the repository at this point in the history

Commits on Oct 23, 2024

  1. fix and improve CCSD_T2_8 

    sync on oldphase was wrong.  sync on the phase associated with the buffers about to be overwritten.
remove oldphase variable since no longer needed.
fuse the two accumulates at the end since accumulate is more expensive than memcpy.

Signed-off-by: Jeff Hammond <[email protected]>
    @jeffhammond
    jeffhammond committed Oct 23, 2024
    Configuration menu
    Copy the full SHA
    239ce8b View commit details
    Browse the repository at this point in the history

Commits on Oct 25, 2024

  1. add 2 phase ccsd_t2_7_3 and broken cublas of the same

    @jeffhammond
    jeffhammond committed Oct 25, 2024
    Configuration menu
    Copy the full SHA
    9d4040a View commit details
    Browse the repository at this point in the history

  2. almost there, i hope

    @jeffhammond
    jeffhammond committed Oct 25, 2024
    Configuration menu
    Copy the full SHA
    450b52e View commit details
    Browse the repository at this point in the history