-
Notifications
You must be signed in to change notification settings - Fork 113
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.
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
Add reduce then scan algorithm for transform scan family #1762
Commits on Aug 14, 2024
-
Checkpoint for reduce then scan integration
Test is currently compiling. None of the real device code has been integrated, but the preliminary host code and general structure has been defined. Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for d07ada2 - Browse repository at this point
Copy the full SHA d07ada2View commit details -
Introduce a parallel_backend_sycl_reduce_then_scan.h file to contain …
…implementation Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 6244266 - Browse repository at this point
Copy the full SHA 6244266View commit details -
Port of kernels from two-pass scan KT branch
Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for ccdb3b0 - Browse repository at this point
Copy the full SHA ccdb3b0View commit details -
Move the single-element last element storage for exclusive_scan after…
… the init computation Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for b465d84 - Browse repository at this point
Copy the full SHA b465d84View commit details -
Configuration menu - View commit details
-
Copy full SHA for 47360a0 - Browse repository at this point
Copy the full SHA 47360a0View commit details -
Lower single work-group upper limit to 2048 elements (empirically found)
Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 3bf0602 - Browse repository at this point
Copy the full SHA 3bf0602View commit details -
[PROTOTYPE] Generalized two pass algorithm and copy_if (#1700)
This PR changes the two pass algorithm to be more generalized for use with other scan-like algorithms like copy_if. This PR adds copy_if as an example --------- Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com> Signed-off-by: Matthew Michel <matthew.michel@intel.com> Co-authored-by: Adam Fidel <110841220+adamfidel@users.noreply.github.com> Co-authored-by: Matthew Michel <106704043+mmichel11@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 46c1a50 - Browse repository at this point
Copy the full SHA 46c1a50View commit details -
bug fix for global race on block carry-out
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 38c1b19 - Browse repository at this point
Copy the full SHA 38c1b19View commit details -
bugfix for elements to process in partial subgroup scan
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 72d42c2 - Browse repository at this point
Copy the full SHA 72d42c2View commit details -
[PROTOTYPE] Add unused temporary storage to single work-group scan to…
… fix use-after free error (#1712)
Configuration menu - View commit details
-
Copy full SHA for ecce124 - Browse repository at this point
Copy the full SHA ecce124View commit details -
Add temporary work-group size cap for FPGA_EMU testing
Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 39ebdbe - Browse repository at this point
Copy the full SHA 39ebdbeView commit details -
[PROTOTYPE] Resolve conversion issues between internal tuple and std:…
…:tuple in zip_iterator.pass (#1714) * Fix for zip_iterator.pass in copy_if assignment Signed-off-by: Matthew Michel <matthew.michel@intel.com> * Add similar fix to __simple_write_to_idx Signed-off-by: Matthew Michel <matthew.michel@intel.com> --------- Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for e4e30e1 - Browse repository at this point
Copy the full SHA e4e30e1View commit details -
Configuration menu - View commit details
-
Copy full SHA for 3732c12 - Browse repository at this point
Copy the full SHA 3732c12View commit details -
bugfix for overruning input for small non multiples of subgroup size
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 1745e0c - Browse repository at this point
Copy the full SHA 1745e0cView commit details -
Check if a subgroup is active before fetching its carry and grab the …
…last active subgroup otherwise Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 0921941 - Browse repository at this point
Copy the full SHA 0921941View commit details -
Comment out std::complex tests in scan_by_segment tests
Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 8effa03 - Browse repository at this point
Copy the full SHA 8effa03View commit details -
Configuration menu - View commit details
-
Copy full SHA for c22231a - Browse repository at this point
Copy the full SHA c22231aView commit details -
fixing device copyable for helpers
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 598f569 - Browse repository at this point
Copy the full SHA 598f569View commit details -
Remove commented code that remained after rebase
Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 96b4fd2 - Browse repository at this point
Copy the full SHA 96b4fd2View commit details -
[PROTOTYPE] Add fallback to legacy scan implementation for CPU device…
…s and devices that lack size 32 sub-groups (#1749) Signed-off-by: Matthew Michel <matthew.michel@intel.com> Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com> Co-authored-by: Adam Fidel <adam.fidel@intel.com> Co-authored-by: Dan Hoeflinger <109972525+danhoeflinger@users.noreply.github.com> Co-authored-by: Adam Fidel <110841220+adamfidel@users.noreply.github.com> Co-authored-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 8f759a3 - Browse repository at this point
Copy the full SHA 8f759a3View commit details -
[PROTOTYPE] partition, unique families and ranges API (#1708)
Enabling partition and unique family of scan-like algorithms includes ranges API Making legacy scan user `__result_and_scratch_storage` to match future type for return to compile Refactoring of `__pattern` / `__parallel` structure for scan-like algorithms for consistency --------- Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 6da54e7 - Browse repository at this point
Copy the full SHA 6da54e7View commit details -
fix windows issue regression __out
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 13cecbf - Browse repository at this point
Copy the full SHA 13cecbfView commit details -
fix for missing assigner in copy if pattern
standardize setup for __parallel_scan_copy Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 2daefab - Browse repository at this point
Copy the full SHA 2daefabView commit details -
fix unique same mangled name problem
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 4a83e1b - Browse repository at this point
Copy the full SHA 4a83e1bView commit details -
[PROTOTYPE] Cleanup reduce-then-scan code (#1760)
* Cleanup reduce-then-scan code Signed-off-by: Matthew Michel <matthew.michel@intel.com> * Remove old comment Signed-off-by: Matthew Michel <matthew.michel@intel.com> --------- Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 299b28b - Browse repository at this point
Copy the full SHA 299b28bView commit details -
restoring removed whitespace line
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 8266882 - Browse repository at this point
Copy the full SHA 8266882View commit details -
removing unnecessay storage type from kernel name
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 453d4ca - Browse repository at this point
Copy the full SHA 453d4caView commit details -
remove unique pattern family from reduce_then_scan
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 78e33ac - Browse repository at this point
Copy the full SHA 78e33acView commit details -
remove partition pattern family from reduce_then_scan
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 8267513 - Browse repository at this point
Copy the full SHA 8267513View commit details -
remove copy_if pattern family from reduce_then_scan
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for d37746e - Browse repository at this point
Copy the full SHA d37746eView commit details -
remove unnecessary barrier + cleanup unnecessary lazy value
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 404c4ef - Browse repository at this point
Copy the full SHA 404c4efView commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 060f649 - Browse repository at this point
Copy the full SHA 060f649View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 0beebd1 - Browse repository at this point
Copy the full SHA 0beebd1View commit details -
Configuration menu - View commit details
-
Copy full SHA for 90e6e62 - Browse repository at this point
Copy the full SHA 90e6e62View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for ef5d377 - Browse repository at this point
Copy the full SHA ef5d377View commit details -
reverting formatting only changes
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for bca0002 - Browse repository at this point
Copy the full SHA bca0002View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 68c75e5 - Browse repository at this point
Copy the full SHA 68c75e5View commit details -
Configuration menu - View commit details
-
Copy full SHA for dddb050 - Browse repository at this point
Copy the full SHA dddb050View commit details -
removing formatting only changes
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for dc2de26 - Browse repository at this point
Copy the full SHA dc2de26View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 165b1a5 - Browse repository at this point
Copy the full SHA 165b1a5View commit details -
removing unused forwarding references
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for b9f0f4e - Browse repository at this point
Copy the full SHA b9f0f4eView commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for bd144a4 - Browse repository at this point
Copy the full SHA bd144a4View commit details -
adding comment and different threshold for different implementations
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for d809051 - Browse repository at this point
Copy the full SHA d809051View commit details -
checking is_gpu rather than !is_cpu
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 1647722 - Browse repository at this point
Copy the full SHA 1647722View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 0271b40 - Browse repository at this point
Copy the full SHA 0271b40View commit details -
removing bad formatting only changes (::std::)
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 6cfc979 - Browse repository at this point
Copy the full SHA 6cfc979View commit details
Commits on Aug 15, 2024
-
fixing result_and_scratch_storage creation
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for cc03af1 - Browse repository at this point
Copy the full SHA cc03af1View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 98de25d - Browse repository at this point
Copy the full SHA 98de25dView commit details -
fixing single pass scan KT from change to single-wg check
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 59933c1 - Browse repository at this point
Copy the full SHA 59933c1View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 94e6e97 - Browse repository at this point
Copy the full SHA 94e6e97View commit details -
refactor subgroup scan to reduce redundant code
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for ddaad55 - Browse repository at this point
Copy the full SHA ddaad55View commit details -
refactoring full block / full thread logic to remove redundancy
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 1fc0f59 - Browse repository at this point
Copy the full SHA 1fc0f59View commit details -
passing storage container by ref
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for a5753d0 - Browse repository at this point
Copy the full SHA a5753d0View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 761ec51 - Browse repository at this point
Copy the full SHA 761ec51View commit details -
__group_start_idx -> __group_start_id
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 4d8c92d - Browse repository at this point
Copy the full SHA 4d8c92dView commit details -
minor variable naming and helpers
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 55db83e - Browse repository at this point
Copy the full SHA 55db83eView commit details -
improving comments, removing unused variable
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for f3768bf - Browse repository at this point
Copy the full SHA f3768bfView commit details -
__prefer_reduce_then_scan -> __is_gpu_with_sg_32
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for f1361d2 - Browse repository at this point
Copy the full SHA f1361d2View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for b67b987 - Browse repository at this point
Copy the full SHA b67b987View commit details -
fold initial value into __carry_offset
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for f3aec73 - Browse repository at this point
Copy the full SHA f3aec73View commit details -
running tally of __reduction_scan_id
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 15d09e2 - Browse repository at this point
Copy the full SHA 15d09e2View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 6bbe469 - Browse repository at this point
Copy the full SHA 6bbe469View commit details -
running tally of __load_reduction_id rather than recalculating
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for a7d00db - Browse repository at this point
Copy the full SHA a7d00dbView commit details -
running tally of __reduction_id rather than recalculating
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for f54e298 - Browse repository at this point
Copy the full SHA f54e298View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for d11dd6f - Browse repository at this point
Copy the full SHA d11dd6fView commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 1a29790 - Browse repository at this point
Copy the full SHA 1a29790View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for e936e83 - Browse repository at this point
Copy the full SHA e936e83View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 1b4f365 - Browse repository at this point
Copy the full SHA 1b4f365View commit details -
rename variables for consistency
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 0ca6f48 - Browse repository at this point
Copy the full SHA 0ca6f48View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for df6a223 - Browse repository at this point
Copy the full SHA df6a223View commit details
Commits on Aug 19, 2024
-
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 6e470e5 - Browse repository at this point
Copy the full SHA 6e470e5View commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 528e04a - Browse repository at this point
Copy the full SHA 528e04aView commit details -
Simplify conversions in __gen_transform_input
Signed-off-by: Matthew Michel <matthew.michel@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 8104f1f - Browse repository at this point
Copy the full SHA 8104f1fView commit details -
Configuration menu - View commit details
-
Copy full SHA for a5367d1 - Browse repository at this point
Copy the full SHA a5367d1View commit details
Commits on Aug 21, 2024
-
Configuration menu - View commit details
-
Copy full SHA for 6096e7a - Browse repository at this point
Copy the full SHA 6096e7aView commit details -
auto -> real types and formatting
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 60c8516 - Browse repository at this point
Copy the full SHA 60c8516View commit details -
fixing type of subgroup id returns
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 8121d67 - Browse repository at this point
Copy the full SHA 8121d67View commit details -
shrinking subgroup size id types
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 48724db - Browse repository at this point
Copy the full SHA 48724dbView commit details -
adjust type to depend on input range
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 3cc61db - Browse repository at this point
Copy the full SHA 3cc61dbView commit details
Commits on Aug 22, 2024
-
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for c2c7e35 - Browse repository at this point
Copy the full SHA c2c7e35View commit details -
shrinking types, switch branch to min, remove double deref
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for ff7b256 - Browse repository at this point
Copy the full SHA ff7b256View commit details -
Configuration menu - View commit details
-
Copy full SHA for 8a36d5a - Browse repository at this point
Copy the full SHA 8a36d5aView commit details -
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 9520f3c - Browse repository at this point
Copy the full SHA 9520f3cView commit details
Commits on Aug 23, 2024
-
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 5a928fd - Browse repository at this point
Copy the full SHA 5a928fdView commit details
Commits on Aug 27, 2024
-
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for e57573f - Browse repository at this point
Copy the full SHA e57573fView commit details
Commits on Aug 29, 2024
-
minor comments from review + formatting
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for 93189b0 - Browse repository at this point
Copy the full SHA 93189b0View commit details -
Co-authored-by: Matthew Michel <106704043+mmichel11@users.noreply.github.com>
Configuration menu - View commit details
-
Copy full SHA for 4e4568e - Browse repository at this point
Copy the full SHA 4e4568eView commit details -
rounding workgroup size down to mult of subgroup size
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Configuration menu - View commit details
-
Copy full SHA for af82182 - Browse repository at this point
Copy the full SHA af82182View commit details