Reorder channels in conda-build #1182
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Looks good to me. Squash and then merge. |
oleksandr-pavlyk
approved these changes
Sep 9, 2022
antonwolfy
added a commit
that referenced
this pull request
Sep 21, 2022
* dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter
The reason this crashed with CPU device and gave incorrect results on
Windows was deeper.
1. Adapter call allocated USM-shared buffer and copies data into it
2. Kernel is submitted to work on USM-shared pointer
3. dpnp_take_c returns kernel submission even
4. Adapter class goes out of scope and frees USM allocation
without making sure that the kernel that works on it has completed
its execution
5. If kernel execution was in progress we got a crash on CPU, or
incorrect result on GPU
If kernel execution was complete it worked as expected.
This change fixes the problem because it removes creation of unprotected
USM-shared temporary.
* Change to DPNPC_adapter to set/use events upon which deallocation must depend
The deallocation routine simply calls sycl::event::wait on the stored vector.
* Used DPNPC_ptr_adapter::depends_on
Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak.
* Get rid of "Improper Null Termination" issue
Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan.
* implemented PR feedback
* Reworked solution with a pointer on void
* Update dpnp/backend/kernels/dpnp_krnl_random.cpp
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/backend/kernels/dpnp_krnl_random.cpp
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Skip for two more tests till waiting fix (#1171)
* Skip for two more tests till waiting fix
tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan
tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan
Need to skip them because CI does not work due to this.
* The same tests skip for gpu
* dpnp_take failed on Windows due to memory corruption (#1172)
* dpnp_take failed on Windows due to memory corruption
* Add more tests
* Integer indexes types with different types of input data
* Add trailing empty line to .gitignore
* Add workflow for Win
Fix typo
Relax a strict pinning for numpy & cmake
Update run command for conda build on Win
Fix declaring DPLROOT env
Fix DPLROOT source
Fix DPLROOT for Win
Add missing double quotes
Try conda-incubator for Linux
Setup conda-incubator for Linux
Update caching
Exclude python 3.8
Strickly pin on 3.8.13
Change channel order
Fix artifcat uploading
Replace to single quotes
Add missing backslash
Corect backslash
* Attempt to fix workflow
* attempt to fix upload steps of the workflow on Linux
* Another attempt to fix upload step of conda-package workflow
* Set default shell in upload actions (#1180)
* Use pin_compatible for run-time dependency generation on numpy, restrict numpy version bracket for host section
* Reorder channels in conda-build (#1182)
* Reorder channels in conda-build
* Remove conda-build script for Linux
* Add tests running as a part of github actions (#1184)
* [Build] setuptools 63.4.1 breaks build for Windows (#1185)
* [SAT-5366] setuptools 63.4.1 breaks build for Windows
* Add TODO note as suggested in review comment
Co-authored-by: Oleksandr Pavlyk <[email protected]>
Co-authored-by: Alexander Rybkin <[email protected]>
antonwolfy
added a commit
that referenced
this pull request
Sep 29, 2022
* dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter
The reason this crashed with CPU device and gave incorrect results on
Windows was deeper.
1. Adapter call allocated USM-shared buffer and copies data into it
2. Kernel is submitted to work on USM-shared pointer
3. dpnp_take_c returns kernel submission even
4. Adapter class goes out of scope and frees USM allocation
without making sure that the kernel that works on it has completed
its execution
5. If kernel execution was in progress we got a crash on CPU, or
incorrect result on GPU
If kernel execution was complete it worked as expected.
This change fixes the problem because it removes creation of unprotected
USM-shared temporary.
* Change to DPNPC_adapter to set/use events upon which deallocation must depend
The deallocation routine simply calls sycl::event::wait on the stored vector.
* Used DPNPC_ptr_adapter::depends_on
Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak.
* Get rid of "Improper Null Termination" issue
Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan.
* implemented PR feedback
* Reworked solution with a pointer on void
* Update dpnp/backend/kernels/dpnp_krnl_random.cpp
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/backend/kernels/dpnp_krnl_random.cpp
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Skip for two more tests till waiting fix (#1171)
* Skip for two more tests till waiting fix
tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan
tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan
Need to skip them because CI does not work due to this.
* The same tests skip for gpu
* dpnp_take failed on Windows due to memory corruption (#1172)
* dpnp_take failed on Windows due to memory corruption
* Add more tests
* Integer indexes types with different types of input data
* Add trailing empty line to .gitignore
* Add workflow for Win
Fix typo
Relax a strict pinning for numpy & cmake
Update run command for conda build on Win
Fix declaring DPLROOT env
Fix DPLROOT source
Fix DPLROOT for Win
Add missing double quotes
Try conda-incubator for Linux
Setup conda-incubator for Linux
Update caching
Exclude python 3.8
Strickly pin on 3.8.13
Change channel order
Fix artifcat uploading
Replace to single quotes
Add missing backslash
Corect backslash
* Attempt to fix workflow
* attempt to fix upload steps of the workflow on Linux
* Another attempt to fix upload step of conda-package workflow
* Set default shell in upload actions (#1180)
* Use pin_compatible for run-time dependency generation on numpy, restrict numpy version bracket for host section
* Reorder channels in conda-build (#1182)
* Reorder channels in conda-build
* Remove conda-build script for Linux
* Add tests running as a part of github actions (#1184)
* [Build] setuptools 63.4.1 breaks build for Windows (#1185)
* [SAT-5366] setuptools 63.4.1 breaks build for Windows
* Add TODO note as suggested in review comment
* Add extra information on how to build dpdnp from source.
* Fix import dpnp on Win & python 3.10 (#1189)
* Fix import dpnp on Win & python 3.10
* PATH is used by python till 3.10
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Apply review comments
Co-authored-by: Oleksandr Pavlyk <[email protected]>
Co-authored-by: Oleksandr Pavlyk <[email protected]>
Co-authored-by: Alexander Rybkin <[email protected]>
Co-authored-by: Diptorup Deb <[email protected]>
xaleryb
added a commit
that referenced
this pull request
Oct 10, 2022
* dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter
The reason this crashed with CPU device and gave incorrect results on
Windows was deeper.
1. Adapter call allocated USM-shared buffer and copies data into it
2. Kernel is submitted to work on USM-shared pointer
3. dpnp_take_c returns kernel submission even
4. Adapter class goes out of scope and frees USM allocation
without making sure that the kernel that works on it has completed
its execution
5. If kernel execution was in progress we got a crash on CPU, or
incorrect result on GPU
If kernel execution was complete it worked as expected.
This change fixes the problem because it removes creation of unprotected
USM-shared temporary.
* Change to DPNPC_adapter to set/use events upon which deallocation must depend
The deallocation routine simply calls sycl::event::wait on the stored vector.
* Used DPNPC_ptr_adapter::depends_on
Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak.
* Get rid of "Improper Null Termination" issue
Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan.
* implemented PR feedback
* Reworked solution with a pointer on void
* Update dpnp/backend/kernels/dpnp_krnl_random.cpp
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/backend/kernels/dpnp_krnl_random.cpp
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Skip for two more tests till waiting fix (#1171)
* Skip for two more tests till waiting fix
tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan
tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan
Need to skip them because CI does not work due to this.
* The same tests skip for gpu
* dpnp_take failed on Windows due to memory corruption (#1172)
* dpnp_take failed on Windows due to memory corruption
* Add more tests
* Integer indexes types with different types of input data
* Add trailing empty line to .gitignore
* Add workflow for Win
Fix typo
Relax a strict pinning for numpy & cmake
Update run command for conda build on Win
Fix declaring DPLROOT env
Fix DPLROOT source
Fix DPLROOT for Win
Add missing double quotes
Try conda-incubator for Linux
Setup conda-incubator for Linux
Update caching
Exclude python 3.8
Strickly pin on 3.8.13
Change channel order
Fix artifcat uploading
Replace to single quotes
Add missing backslash
Corect backslash
* Attempt to fix workflow
* attempt to fix upload steps of the workflow on Linux
* Another attempt to fix upload step of conda-package workflow
* Set default shell in upload actions (#1180)
* Use pin_compatible for run-time dependency generation on numpy, restrict numpy version bracket for host section
* Reorder channels in conda-build (#1182)
* Reorder channels in conda-build
* Remove conda-build script for Linux
* Add tests running as a part of github actions (#1184)
* [Build] setuptools 63.4.1 breaks build for Windows (#1185)
* [SAT-5366] setuptools 63.4.1 breaks build for Windows
* Add TODO note as suggested in review comment
* Add extra information on how to build dpdnp from source.
* Fix import dpnp on Win & python 3.10 (#1189)
* Fix import dpnp on Win & python 3.10
* PATH is used by python till 3.10
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Update dpnp/__init__.py
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Apply review comments
Co-authored-by: Oleksandr Pavlyk <[email protected]>
* Missing path towards TBB DLLs on Windows (#1195)
* Fix missing support of compute follows data in dpnp.erf (#1194)
* Fix support of compute follows data in dpnp.erf
* Include tests for dpnp.erf in validation with github action
* Fix github action for building docs (#1191)
* Fix github action for building docs
* Get rid of pinning for cmake
* Setting version to 0.10.2 (#1196)
Co-authored-by: Oleksandr Pavlyk <[email protected]>
Co-authored-by: Alexander Rybkin <[email protected]>
Co-authored-by: Diptorup Deb <[email protected]>
antonwolfy
added a commit
to antonwolfy/dpnp
that referenced
this pull request
Nov 2, 2022
commit e97a9a9 Author: Anton <[email protected]> Date: Wed Nov 2 19:04:44 2022 +0100 Roll back pinning on NumPy for gold/2021 branch (IntelPython#1217) The branch `gold/2021` is intended for internal CI needs, where it is required to explicitly take NumPy 1.19 during the build in upcoming release. The change mustn't aligned with `master` branch and it requires to be controlled every next push/merge on `gold/2021` branch. Alternative solution might be to update run dependencies with `max_pin` on `x.x` instead of `x`, like: ``` {{ pin_compatible('numpy', min_pin='x.x', max_pin='x.x') }} ``` but it can produce another unwanted consequences which has to be properly analyzed firstly. - [X] Have you provided a meaningful PR description? - [ ] Have you added a test, reproducer or referred to issue with a reproducer? - [ ] Have you tested your changes locally for CPU and GPU devices? - [X] Have you made sure that new changes do not introduce compiler warnings? - [ ] If this PR is a work in progress, are you filing the PR as a draft? commit 2f4e7d1 Merge: c9fab4c 78c4b8b Author: Anton Volkov <[email protected]> Date: Thu Oct 20 17:19:20 2022 +0200 Merge branch 'master' into gold/2021 commit 78c4b8b Author: Anton <[email protected]> Date: Thu Oct 20 16:23:30 2022 +0200 Get rid of env var which handles dep on dpctl.tensor (IntelPython#1205) * Get rid of env var stating dep on dpctl.tensor * Apply review comments commit c9fab4c Author: Evseniia Komarova <[email protected]> Date: Thu Oct 20 14:00:46 2022 +0200 update numpy restriction by adding numpy 1.22 (IntelPython#1208) commit 384a4ba Author: Evseniia Komarova <[email protected]> Date: Wed Oct 19 16:29:48 2022 +0200 update numpy restriction by adding numpy 1.22 (IntelPython#1207) * update numpy restriction by adding numpy 1.22 * Update conda-recipe/meta.yaml Also exclude alpha/beta builds of 1.23 sources Co-authored-by: Oleksandr Pavlyk <[email protected]> commit 35e28a4 Author: Lukicheva Polina <[email protected]> Date: Tue Oct 18 01:48:12 2022 +0300 Enable Compute Follows Data in Cython in fft and linalg (IntelPython#1132) commit a718039 Author: Anton <[email protected]> Date: Mon Oct 17 21:26:26 2022 +0200 Align normalize_queue_device call with CFD (IntelPython#1200) * Align normalize_queue_device call with CFD in dpnp * Update dpnp/dpnp_iface.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/dpnp_iface.py Co-authored-by: Oleksandr Pavlyk <[email protected]> Co-authored-by: Oleksandr Pavlyk <[email protected]> commit cfb62a6 Author: Anton <[email protected]> Date: Mon Oct 17 15:46:06 2022 +0200 Test runs to update to node 16 (IntelPython#1198) commit 9767f69 Author: Anton <[email protected]> Date: Mon Oct 10 18:46:54 2022 +0200 Merge changes for OOC release (IntelPython#1197) * dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter The reason this crashed with CPU device and gave incorrect results on Windows was deeper. 1. Adapter call allocated USM-shared buffer and copies data into it 2. Kernel is submitted to work on USM-shared pointer 3. dpnp_take_c returns kernel submission even 4. Adapter class goes out of scope and frees USM allocation without making sure that the kernel that works on it has completed its execution 5. If kernel execution was in progress we got a crash on CPU, or incorrect result on GPU If kernel execution was complete it worked as expected. This change fixes the problem because it removes creation of unprotected USM-shared temporary. * Change to DPNPC_adapter to set/use events upon which deallocation must depend The deallocation routine simply calls sycl::event::wait on the stored vector. * Used DPNPC_ptr_adapter::depends_on Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak. * Get rid of "Improper Null Termination" issue Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan. * implemented PR feedback * Reworked solution with a pointer on void * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Skip for two more tests till waiting fix (IntelPython#1171) * Skip for two more tests till waiting fix tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan Need to skip them because CI does not work due to this. * The same tests skip for gpu * dpnp_take failed on Windows due to memory corruption (IntelPython#1172) * dpnp_take failed on Windows due to memory corruption * Add more tests * Integer indexes types with different types of input data * Add trailing empty line to .gitignore * Add workflow for Win Fix typo Relax a strict pinning for numpy & cmake Update run command for conda build on Win Fix declaring DPLROOT env Fix DPLROOT source Fix DPLROOT for Win Add missing double quotes Try conda-incubator for Linux Setup conda-incubator for Linux Update caching Exclude python 3.8 Strickly pin on 3.8.13 Change channel order Fix artifcat uploading Replace to single quotes Add missing backslash Corect backslash * Attempt to fix workflow * attempt to fix upload steps of the workflow on Linux * Another attempt to fix upload step of conda-package workflow * Set default shell in upload actions (IntelPython#1180) * Use pin_compatible for run-time dependency generation on numpy, restrict numpy version bracket for host section * Reorder channels in conda-build (IntelPython#1182) * Reorder channels in conda-build * Remove conda-build script for Linux * Add tests running as a part of github actions (IntelPython#1184) * [Build] setuptools 63.4.1 breaks build for Windows (IntelPython#1185) * [SAT-5366] setuptools 63.4.1 breaks build for Windows * Add TODO note as suggested in review comment * Add extra information on how to build dpdnp from source. * Fix import dpnp on Win & python 3.10 (IntelPython#1189) * Fix import dpnp on Win & python 3.10 * PATH is used by python till 3.10 * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Apply review comments Co-authored-by: Oleksandr Pavlyk <[email protected]> * Missing path towards TBB DLLs on Windows (IntelPython#1195) * Fix missing support of compute follows data in dpnp.erf (IntelPython#1194) * Fix support of compute follows data in dpnp.erf * Include tests for dpnp.erf in validation with github action * Fix github action for building docs (IntelPython#1191) * Fix github action for building docs * Get rid of pinning for cmake * Setting version to 0.10.2 (IntelPython#1196) Co-authored-by: Oleksandr Pavlyk <[email protected]> Co-authored-by: Alexander Rybkin <[email protected]> Co-authored-by: Diptorup Deb <[email protected]> commit 42ec3a2 Merge: 95d47fd 5471e19 Author: Anton <[email protected]> Date: Mon Oct 10 17:18:49 2022 +0200 Merge branch 'gold/2021' into master commit 95d47fd Author: Anton <[email protected]> Date: Mon Oct 10 16:56:57 2022 +0200 Setting version to 0.10.2 (IntelPython#1196) commit 9830f99 Author: Anton <[email protected]> Date: Fri Oct 7 16:57:20 2022 +0200 Fix github action for building docs (IntelPython#1191) * Fix github action for building docs * Get rid of pinning for cmake commit 63ff524 Author: Anton <[email protected]> Date: Fri Oct 7 14:42:11 2022 +0200 Fix missing support of compute follows data in dpnp.erf (IntelPython#1194) * Fix support of compute follows data in dpnp.erf * Include tests for dpnp.erf in validation with github action commit da21159 Author: Anton <[email protected]> Date: Thu Oct 6 17:39:03 2022 +0200 Missing path towards TBB DLLs on Windows (IntelPython#1195) commit 5471e19 Author: Anton <[email protected]> Date: Thu Sep 29 10:40:40 2022 +0200 Merge changes for nightly tests in CI (IntelPython#1190) * dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter The reason this crashed with CPU device and gave incorrect results on Windows was deeper. 1. Adapter call allocated USM-shared buffer and copies data into it 2. Kernel is submitted to work on USM-shared pointer 3. dpnp_take_c returns kernel submission even 4. Adapter class goes out of scope and frees USM allocation without making sure that the kernel that works on it has completed its execution 5. If kernel execution was in progress we got a crash on CPU, or incorrect result on GPU If kernel execution was complete it worked as expected. This change fixes the problem because it removes creation of unprotected USM-shared temporary. * Change to DPNPC_adapter to set/use events upon which deallocation must depend The deallocation routine simply calls sycl::event::wait on the stored vector. * Used DPNPC_ptr_adapter::depends_on Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak. * Get rid of "Improper Null Termination" issue Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan. * implemented PR feedback * Reworked solution with a pointer on void * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Skip for two more tests till waiting fix (IntelPython#1171) * Skip for two more tests till waiting fix tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan Need to skip them because CI does not work due to this. * The same tests skip for gpu * dpnp_take failed on Windows due to memory corruption (IntelPython#1172) * dpnp_take failed on Windows due to memory corruption * Add more tests * Integer indexes types with different types of input data * Add trailing empty line to .gitignore * Add workflow for Win Fix typo Relax a strict pinning for numpy & cmake Update run command for conda build on Win Fix declaring DPLROOT env Fix DPLROOT source Fix DPLROOT for Win Add missing double quotes Try conda-incubator for Linux Setup conda-incubator for Linux Update caching Exclude python 3.8 Strickly pin on 3.8.13 Change channel order Fix artifcat uploading Replace to single quotes Add missing backslash Corect backslash * Attempt to fix workflow * attempt to fix upload steps of the workflow on Linux * Another attempt to fix upload step of conda-package workflow * Set default shell in upload actions (IntelPython#1180) * Use pin_compatible for run-time dependency generation on numpy, restrict numpy version bracket for host section * Reorder channels in conda-build (IntelPython#1182) * Reorder channels in conda-build * Remove conda-build script for Linux * Add tests running as a part of github actions (IntelPython#1184) * [Build] setuptools 63.4.1 breaks build for Windows (IntelPython#1185) * [SAT-5366] setuptools 63.4.1 breaks build for Windows * Add TODO note as suggested in review comment * Add extra information on how to build dpdnp from source. * Fix import dpnp on Win & python 3.10 (IntelPython#1189) * Fix import dpnp on Win & python 3.10 * PATH is used by python till 3.10 * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Apply review comments Co-authored-by: Oleksandr Pavlyk <[email protected]> Co-authored-by: Oleksandr Pavlyk <[email protected]> Co-authored-by: Alexander Rybkin <[email protected]> Co-authored-by: Diptorup Deb <[email protected]> commit 1d19119 Merge: efa1a19 6377629 Author: Anton <[email protected]> Date: Thu Sep 29 09:21:57 2022 +0200 Merge branch 'gold/2021' into master commit efa1a19 Author: Anton <[email protected]> Date: Thu Sep 29 09:13:10 2022 +0200 Fix import dpnp on Win & python 3.10 (IntelPython#1189) * Fix import dpnp on Win & python 3.10 * PATH is used by python till 3.10 * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/__init__.py Co-authored-by: Oleksandr Pavlyk <[email protected]> * Apply review comments Co-authored-by: Oleksandr Pavlyk <[email protected]> commit 8b44b4a Merge: 614c829 e3e04c7 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Sep 21 16:38:09 2022 -0500 Merge pull request IntelPython#1188 from IntelPython/update/readme Add extra information on how to build dpnp from source. commit e3e04c7 Author: Diptorup Deb <[email protected]> Date: Wed Sep 21 16:32:40 2022 -0500 Add extra information on how to build dpdnp from source. commit 6377629 Author: Anton <[email protected]> Date: Wed Sep 21 14:30:20 2022 +0200 Merge changes for nightly tests in CI (IntelPython#1186) * dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter The reason this crashed with CPU device and gave incorrect results on Windows was deeper. 1. Adapter call allocated USM-shared buffer and copies data into it 2. Kernel is submitted to work on USM-shared pointer 3. dpnp_take_c returns kernel submission even 4. Adapter class goes out of scope and frees USM allocation without making sure that the kernel that works on it has completed its execution 5. If kernel execution was in progress we got a crash on CPU, or incorrect result on GPU If kernel execution was complete it worked as expected. This change fixes the problem because it removes creation of unprotected USM-shared temporary. * Change to DPNPC_adapter to set/use events upon which deallocation must depend The deallocation routine simply calls sycl::event::wait on the stored vector. * Used DPNPC_ptr_adapter::depends_on Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak. * Get rid of "Improper Null Termination" issue Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan. * implemented PR feedback * Reworked solution with a pointer on void * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Skip for two more tests till waiting fix (IntelPython#1171) * Skip for two more tests till waiting fix tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan Need to skip them because CI does not work due to this. * The same tests skip for gpu * dpnp_take failed on Windows due to memory corruption (IntelPython#1172) * dpnp_take failed on Windows due to memory corruption * Add more tests * Integer indexes types with different types of input data * Add trailing empty line to .gitignore * Add workflow for Win Fix typo Relax a strict pinning for numpy & cmake Update run command for conda build on Win Fix declaring DPLROOT env Fix DPLROOT source Fix DPLROOT for Win Add missing double quotes Try conda-incubator for Linux Setup conda-incubator for Linux Update caching Exclude python 3.8 Strickly pin on 3.8.13 Change channel order Fix artifcat uploading Replace to single quotes Add missing backslash Corect backslash * Attempt to fix workflow * attempt to fix upload steps of the workflow on Linux * Another attempt to fix upload step of conda-package workflow * Set default shell in upload actions (IntelPython#1180) * Use pin_compatible for run-time dependency generation on numpy, restrict numpy version bracket for host section * Reorder channels in conda-build (IntelPython#1182) * Reorder channels in conda-build * Remove conda-build script for Linux * Add tests running as a part of github actions (IntelPython#1184) * [Build] setuptools 63.4.1 breaks build for Windows (IntelPython#1185) * [SAT-5366] setuptools 63.4.1 breaks build for Windows * Add TODO note as suggested in review comment Co-authored-by: Oleksandr Pavlyk <[email protected]> Co-authored-by: Alexander Rybkin <[email protected]> commit 614c829 Author: Anton <[email protected]> Date: Wed Sep 21 11:23:02 2022 +0200 [Build] setuptools 63.4.1 breaks build for Windows (IntelPython#1185) * [SAT-5366] setuptools 63.4.1 breaks build for Windows * Add TODO note as suggested in review comment commit 3be4b2e Author: Anton <[email protected]> Date: Tue Sep 20 20:56:43 2022 +0200 Add tests running as a part of github actions (IntelPython#1184) commit bd1a414 Author: Anton <[email protected]> Date: Sat Sep 10 15:51:29 2022 +0200 Reorder channels in conda-build (IntelPython#1182) * Reorder channels in conda-build * Remove conda-build script for Linux commit 5f87d3a Merge: ac79e9f 64478ae Author: Oleksandr Pavlyk <[email protected]> Date: Fri Sep 9 08:17:36 2022 -0500 Merge pull request IntelPython#1178 from IntelPython/use-variant Build for numpy 1.21 commit 64478ae Author: Oleksandr Pavlyk <[email protected]> Date: Fri Sep 9 07:14:39 2022 -0500 Use pin_compatible for run-time dependency generation on numpy, restrict numpy version bracket for host section commit ac79e9f Author: Anton <[email protected]> Date: Fri Sep 9 13:52:56 2022 +0200 Set default shell in upload actions (IntelPython#1180) commit 2ea4aac Author: Oleksandr Pavlyk <[email protected]> Date: Thu Sep 8 19:52:57 2022 -0500 Another attempt to fix upload step of conda-package workflow commit 26d190c Merge: 31144c6 ee65713 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Sep 8 18:22:50 2022 -0500 Merge pull request IntelPython#1177 from IntelPython/finess-conda-package attempt to fix upload steps of the workflow on Linux commit ee65713 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Sep 8 16:52:28 2022 -0500 attempt to fix upload steps of the workflow on Linux commit 31144c6 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Sep 8 15:57:06 2022 -0500 Attempt to fix workflow commit ac0f6f6 Merge: 59ef493 c91f912 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Sep 8 15:19:42 2022 -0500 Merge pull request IntelPython#1176 from antonwolfy/build_and_upload_for_win Add git workflow factions or Windows commit c91f912 Author: Anton Volkov <[email protected]> Date: Thu Sep 8 07:15:32 2022 -0500 Add workflow for Win Fix typo Relax a strict pinning for numpy & cmake Update run command for conda build on Win Fix declaring DPLROOT env Fix DPLROOT source Fix DPLROOT for Win Add missing double quotes Try conda-incubator for Linux Setup conda-incubator for Linux Update caching Exclude python 3.8 Strickly pin on 3.8.13 Change channel order Fix artifcat uploading Replace to single quotes Add missing backslash Corect backslash commit 1ca7191 Author: Anton <[email protected]> Date: Thu Sep 1 10:39:42 2022 +0200 Merge changes for nightly tests in CI (IntelPython#1173) * dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter The reason this crashed with CPU device and gave incorrect results on Windows was deeper. 1. Adapter call allocated USM-shared buffer and copies data into it 2. Kernel is submitted to work on USM-shared pointer 3. dpnp_take_c returns kernel submission even 4. Adapter class goes out of scope and frees USM allocation without making sure that the kernel that works on it has completed its execution 5. If kernel execution was in progress we got a crash on CPU, or incorrect result on GPU If kernel execution was complete it worked as expected. This change fixes the problem because it removes creation of unprotected USM-shared temporary. * Change to DPNPC_adapter to set/use events upon which deallocation must depend The deallocation routine simply calls sycl::event::wait on the stored vector. * Used DPNPC_ptr_adapter::depends_on Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak. * Get rid of "Improper Null Termination" issue Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan. * implemented PR feedback * Reworked solution with a pointer on void * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Skip for two more tests till waiting fix (IntelPython#1171) * Skip for two more tests till waiting fix tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan Need to skip them because CI does not work due to this. * The same tests skip for gpu * dpnp_take failed on Windows due to memory corruption (IntelPython#1172) * dpnp_take failed on Windows due to memory corruption * Add more tests * Integer indexes types with different types of input data * Add trailing empty line to .gitignore Co-authored-by: Oleksandr Pavlyk <[email protected]> Co-authored-by: Alexander Rybkin <[email protected]> commit 59ef493 Merge: 9b14f0c 4102660 Author: Anton Volkov <[email protected]> Date: Tue Aug 30 12:17:57 2022 -0500 Merge branch 'gold/2021' commit 9b14f0c Author: Anton <[email protected]> Date: Tue Aug 30 14:35:28 2022 +0200 dpnp_take failed on Windows due to memory corruption (IntelPython#1172) * dpnp_take failed on Windows due to memory corruption * Add more tests * Integer indexes types with different types of input data * Add trailing empty line to .gitignore commit 4102660 Author: Alexander Rybkin <[email protected]> Date: Wed Aug 24 16:23:44 2022 +0200 Merge from master to gold to pick-up fixes (IntelPython#1170) * dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter The reason this crashed with CPU device and gave incorrect results on Windows was deeper. 1. Adapter call allocated USM-shared buffer and copies data into it 2. Kernel is submitted to work on USM-shared pointer 3. dpnp_take_c returns kernel submission even 4. Adapter class goes out of scope and frees USM allocation without making sure that the kernel that works on it has completed its execution 5. If kernel execution was in progress we got a crash on CPU, or incorrect result on GPU If kernel execution was complete it worked as expected. This change fixes the problem because it removes creation of unprotected USM-shared temporary. * Change to DPNPC_adapter to set/use events upon which deallocation must depend The deallocation routine simply calls sycl::event::wait on the stored vector. * Used DPNPC_ptr_adapter::depends_on Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak. * Get rid of "Improper Null Termination" issue Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan. * implemented PR feedback * Reworked solution with a pointer on void * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> * Skip for two more tests till waiting fix (IntelPython#1171) * Skip for two more tests till waiting fix tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan Need to skip them because CI does not work due to this. * The same tests skip for gpu Co-authored-by: Oleksandr Pavlyk <[email protected]> Co-authored-by: Anton <[email protected]> Co-authored-by: Anton Volkov <[email protected]> commit 2fed06a Author: Alexander Rybkin <[email protected]> Date: Wed Aug 24 15:14:31 2022 +0200 Skip for two more tests till waiting fix (IntelPython#1171) * Skip for two more tests till waiting fix tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_all_nan tests/third_party/cupy/statistics_tests/test_order.py::TestOrder::test_ptp_nan Need to skip them because CI does not work due to this. * The same tests skip for gpu commit ce6d880 Merge: cfc9b40 629d0e0 Author: Anton <[email protected]> Date: Fri Aug 19 14:03:34 2022 +0200 Merge pull request IntelPython#1168 from IntelPython/antonwolfy/checkmarx_fix Get rid of "Improper Null Termination" issue commit 629d0e0 Author: Anton <[email protected]> Date: Thu Aug 18 22:07:07 2022 +0200 Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> commit b29d957 Author: Anton <[email protected]> Date: Thu Aug 18 22:06:54 2022 +0200 Update dpnp/backend/kernels/dpnp_krnl_random.cpp Co-authored-by: Oleksandr Pavlyk <[email protected]> commit a8da387 Merge: 818dc82 cfc9b40 Author: Anton <[email protected]> Date: Thu Aug 18 22:03:33 2022 +0200 Merge branch 'master' into antonwolfy/checkmarx_fix commit 818dc82 Author: Anton Volkov <[email protected]> Date: Thu Aug 18 14:48:14 2022 -0500 Reworked solution with a pointer on void commit cfc9b40 Merge: 51369c3 d839ea1 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 18 11:38:38 2022 -0500 Merge pull request IntelPython#1167 from IntelPython/fix-take Fix take commit d839ea1 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 17 13:32:08 2022 -0500 implemented PR feedback commit 16a7632 Author: Anton <[email protected]> Date: Wed Aug 17 13:04:52 2022 +0200 Get rid of "Improper Null Termination" issue Add a null-terminated symbol at the end of char array to avoid "Improper Null Termination" issue reported by Checkmarx scan. commit 37386bb Author: Oleksandr Pavlyk <[email protected]> Date: Sun Aug 14 07:08:37 2022 -0500 Used DPNPC_ptr_adapter::depends_on Also applied DPCTLEvent_Delete in legacy interfaces to avoid memory leak. commit 7220d87 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 11 10:31:12 2022 -0500 Change to DPNPC_adapter to set/use events upon which deallocation must depend The deallocation routine simply calls sycl::event::wait on the stored vector. commit 885f91a Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 11 10:27:02 2022 -0500 dpnp_take_c uses SYCL kernel, no need to use no_sycl parameter in adapter The reason this crashed with CPU device and gave incorrect results on Windows was deeper. 1. Adapter call allocated USM-shared buffer and copies data into it 2. Kernel is submitted to work on USM-shared pointer 3. dpnp_take_c returns kernel submission even 4. Adapter class goes out of scope and frees USM allocation without making sure that the kernel that works on it has completed its execution 5. If kernel execution was in progress we got a crash on CPU, or incorrect result on GPU If kernel execution was complete it worked as expected. This change fixes the problem because it removes creation of unprotected USM-shared temporary. commit 7aabc45 Merge: 2332c2d 51369c3 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 10 09:11:06 2022 -0500 Merge remote-tracking branch 'origin/master' into gold/2021 commit 51369c3 Merge: 4f9ed25 8ddaeb7 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 10 09:10:05 2022 -0500 Merge pull request IntelPython#1166 from IntelPython/debug_take_for_main Corrected logic for kernel function pointer lookup for dpnp_take_c commit 8ddaeb7 Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 11:22:21 2022 -0500 Corrected logic for kernel function pointer lookup for dpnp_take_c commit 2332c2d Merge: 24f9921 0be9100 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 10 09:07:30 2022 -0500 Merge pull request IntelPython#1165 from IntelPython/debug_take Corrected logic for kernel function pointer lookup for dpnp_take_c commit 0be9100 Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 11:22:21 2022 -0500 Corrected logic for kernel function pointer lookup for dpnp_take_c commit 4f9ed25 Merge: cc098b1 e3fff8c Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 22:27:15 2022 -0500 Merge pull request IntelPython#1164 from IntelPython/cmake-cleanup-master Cmake cleanup master commit e3fff8c Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 16:33:57 2022 -0500 Add -fno-approx-func to build_clib too commit b9ccfbb Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 15:25:03 2022 -0500 Cleaned up backend/CMakeLists.txt This should allow correct propagation of compiler flags, including on Windows commit 24f9921 Merge: cfe5d63 c0c0552 Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 18:12:42 2022 -0500 Merge pull request IntelPython#1163 from IntelPython/cmake-cleanup Cmake cleanup commit c0c0552 Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 16:33:57 2022 -0500 Add -fno-approx-func to build_clib too commit 5381a9c Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 9 15:25:03 2022 -0500 Cleaned up backend/CMakeLists.txt This should allow correct propagation of compiler flags, including on Windows commit cfe5d63 Merge: 6768325 dd939cb Author: Alexander Rybkin <[email protected]> Date: Fri Aug 5 16:17:17 2022 +0200 Merge pull request IntelPython#1161 from IntelPython/xaleryb/deps Dependencies update for 2022.2 commit dd939cb Author: Alexander Rybkin <[email protected]> Date: Fri Aug 5 16:04:44 2022 +0200 Dependencies update for 2022.2 commit 6768325 Merge: 17539b8 cc098b1 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 4 17:12:15 2022 -0500 Merge remote-tracking branch 'origin/master' into gold/2021 commit cc098b1 Merge: 51ee8c8 fa647a9 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 4 17:12:00 2022 -0500 Merge pull request IntelPython#1160 from IntelPython/dot-crash GEMM to use trans only if matrix if not C-contig commit fa647a9 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 4 14:33:58 2022 -0500 GEMM to use trans only if matrix if not C-contig Fixed logic for determinining transA and transB in call to gemm commit 17539b8 Merge: 08748a1 51ee8c8 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 4 06:51:31 2022 -0500 Merge remote-tracking branch 'origin/master' into gold/2021 commit 51ee8c8 Author: Oleksandr Pavlyk <[email protected]> Date: Thu Aug 4 06:51:19 2022 -0500 Fixed typo commit 08748a1 Merge: eec1214 2907869 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 3 23:06:04 2022 -0500 Merge remote-tracking branch 'origin/master' into gold/2021 commit 2907869 Merge: af5401a f3a13ce Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 3 23:05:05 2022 -0500 Merge pull request IntelPython#1159 from IntelPython/test-failures Do not test on host device in test_sycl.py for now commit f3a13ce Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 3 16:37:52 2022 -0500 Do not test on host device for now. Also make sure to skip rejected devices commit eec1214 Merge: 2953c8e af5401a Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 3 12:09:54 2022 -0500 Merge remote-tracking branch 'origin/master' into gold/2021 commit af5401a Merge: 2c4ee08 2f417b2 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 3 12:08:31 2022 -0500 Merge pull request IntelPython#1158 from IntelPython/transcendental-precision-control Transcendental precision control commit 2f417b2 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 3 08:05:32 2022 -0500 Use -fno-approx-func instead of -fno-fast-math commit f62d806 Author: Oleksandr Pavlyk <[email protected]> Date: Wed Aug 3 07:55:56 2022 -0500 spacing change commit 2953c8e Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 2 16:57:50 2022 -0500 Use -fno-fast-math to ensure that device code uses accurate transcendentals (IntelPython#1157) commit 2c4ee08 Author: Oleksandr Pavlyk <[email protected]> Date: Tue Aug 2 16:57:50 2022 -0500 Use -fno-fast-math to ensure that device code uses accurate transcendentals (IntelPython#1157) commit 0fc7017 Author: Oleksandr Pavlyk <[email protected]> Date: Fri Jul 29 09:28:12 2022 -0500 Setting version to 0.10.1 commit 49c4401 Author: Oleksandr Pavlyk <[email protected]> Date: Fri Jul 29 08:48:24 2022 -0500 Added two test exceptions commit 7e99250 Merge: 3411f28 90a297e Author: Oleksandr Pavlyk <[email protected]> Date: Fri May 6 16:36:03 2022 -0500 Merge pull request IntelPython#1156 from IntelPython/sycl_cython2_check_sasha Merges changes from subsumed IntelPython#1131, IntelPython#1137, IntelPython#1140 as well as changes to address compiler warnings, changes to make call_origin follow CFD for the allocation of the output array, and adding PR template to hopefully stop harmful team practice of not properly documenting PRs commit 90a297e Author: Oleksandr Pavlyk <[email protected]> Date: Fri May 6 11:42:08 2022 -0500 Ignore emacs backup copies commit ef735e2 Author: Oleksandr Pavlyk <[email protected]> Date: Fri May 6 11:41:38 2022 -0500 Added PR template. The madness with empty PR description must stop commit b0175f7 Author: Lukicheva Polina <[email protected]> Date: Wed Mar 2 04:17:08 2022 -0600 Enable Compute Follows Data in Cython in random commit 62d048c Author: Oleksandr Pavlyk <[email protected]> Date: Thu May 5 21:23:49 2022 -0500 Skipped tests that crash on GPU pending investigation commit 01d1c5d Author: Oleksandr Pavlyk <[email protected]> Date: Thu May 5 21:23:22 2022 -0500 addressed compiler warnings commit c8a5343 Author: Oleksandr Pavlyk <[email protected]> Date: Thu May 5 13:55:19 2022 -0500 Applied unique changes from IntelPython#1131 These were: ``` commit e57df0e (pr_1131) Author: Denis Smirnov <[email protected]> Date: Fri Apr 1 17:57:51 2022 +0300 Revert fallback to numpy commit 92c2fae Author: Denis Smirnov <[email protected]> Date: Wed Mar 30 12:32:58 2022 +0300 Normilize queue and device + change copying from numpy in call_origin commit 9fb287c Author: Denis Smirnov <[email protected]> Date: Wed Mar 23 11:36:59 2022 +0300 Change output array creation in dpnp_diag (part 2) commit a1597d0 Author: Denis Smirnov <[email protected]> Date: Wed Mar 23 09:52:50 2022 +0300 Change output array creation in dpnp_diag ``` commit 404b587 Author: Oleksandr Pavlyk <[email protected]> Date: Thu May 5 10:45:34 2022 -0500 Modified call_origin to deduce result allocation queue from input allocation queues using CFD commit c3a5cab Author: Oleksandr Pavlyk <[email protected]> Date: Thu May 5 10:44:34 2022 -0500 dpnp_array must expose sycl_queue, sycl_device, sycl_context, and device properties commit 9c6bd87 Author: Oleksandr Pavlyk <[email protected]> Date: Thu May 5 10:40:54 2022 -0500 expanded assert_sycl_queue_equal commit dea42f6 Merge: 174c0f1 3411f28 Author: Oleksandr Pavlyk <[email protected]> Date: Mon May 2 13:16:00 2022 -0500 Merge remote-tracking branch 'origin/master' into sycl_to_cython2_check commit 3411f28 Author: Anton <[email protected]> Date: Thu Mar 31 11:27:22 2022 +0300 Add '-fsycl-device-code-split=per_kernel' option while biulding and (IntelPython#1152) linking DPNP backend for Windows Co-authored-by: Alexander-Makaryev <[email protected]> commit 1b5173d Author: Alexander-Makaryev <[email protected]> Date: Wed Mar 30 22:03:28 2022 +0300 Using device or sycl_queue keyword, not both (IntelPython#1147) * Using device or sycl_queue keyword, not both commit 395f051 Author: KsanaKozlova <[email protected]> Date: Wed Mar 30 21:52:02 2022 +0300 support ifft with real input (IntelPython#1150) Co-authored-by: Alexander-Makaryev <[email protected]> commit f8b0ce8 Author: Anton <[email protected]> Date: Wed Mar 30 21:49:36 2022 +0300 Function "linalg.qr" raised an exception with mode='reduced' (IntelPython#1148) * DPNP function "linalg.qr" raised an exception when mode='reduced'. commit 0234590 Author: KsanaKozlova <[email protected]> Date: Thu Mar 24 13:22:29 2022 +0300 add ifft support (IntelPython#1142) * add ifft support commit 6874366 Author: Alexander-Makaryev <[email protected]> Date: Mon Mar 14 16:51:00 2022 +0300 to_device array method impl (IntelPython#1143) * to_device array method impl commit 67ac3c6 Author: KsanaKozlova <[email protected]> Date: Fri Mar 11 15:33:06 2022 +0300 support real input in fft with mkl (IntelPython#1120) * add mkl to handle real input in fft commit 174c0f1 Author: Denis Smirnov <[email protected]> Date: Sat Mar 5 12:13:23 2022 +0300 Unskip half of test_1in_1out commit 78d6e28 Author: Denis Smirnov <[email protected]> Date: Sat Mar 5 10:22:07 2022 +0300 Skip test_1in_1out commit 39f8b87 Author: Denis Smirnov <[email protected]> Date: Fri Mar 4 21:48:53 2022 +0300 Unskip test_1in_1out commit 8e63208 Author: Denis Smirnov <[email protected]> Date: Fri Mar 4 20:20:50 2022 +0300 Unskip test_2in_1out commit 47edea4 Author: Denis Smirnov <[email protected]> Date: Fri Mar 4 17:37:36 2022 +0300 Unskip test_1in_1out and test_2in_1out commit f4f466a Author: Denis Smirnov <[email protected]> Date: Fri Mar 4 16:01:18 2022 +0300 Unskip tests/test_sycl_queue.py::test_broadcasting commit 3f99c62 Author: Denis Smirnov <[email protected]> Date: Fri Mar 4 15:13:34 2022 +0300 Unskip tests/test_sycl_queue.py::test_out commit eb8dd52 Author: Denis Smirnov <[email protected]> Date: Fri Mar 4 13:50:34 2022 +0300 Unskip test for modf commit dd92cdb Author: Denis Smirnov <[email protected]> Date: Fri Mar 4 10:34:10 2022 +0300 Skip all tests from test_sycl_queue.py commit 7595c82 Author: Sergey Pokhodenko <[email protected]> Date: Thu Mar 3 19:33:21 2022 +0300 Build with conda packages instead of with oneAPI (IntelPython#1138) * Remove oneAPI installation commit 181490e Author: Denis Smirnov <[email protected]> Date: Thu Mar 3 11:18:32 2022 +0300 Skip level_zero tests commit 0ac3872 Author: Denis Smirnov <[email protected]> Date: Wed Mar 2 22:59:18 2022 +0300 Skip tests which test function conjugate commit 9a496d8 Author: Denis Smirnov <[email protected]> Date: Wed Mar 2 22:09:51 2022 +0300 Skip tests which test function ceil commit 6783c04 Author: Denis Smirnov <[email protected]> Date: Wed Mar 2 18:19:52 2022 +0300 Skip tests which test function abs commit dc50bff Merge: 8431a64 5d65a33 Author: Denis Smirnov <[email protected]> Date: Tue Mar 1 14:15:09 2022 +0300 Merge branch 'master' of https://github.com/IntelPython/dpnp into feature/sycl_to_cython2 commit 8431a64 Author: Denis Smirnov <[email protected]> Date: Thu Feb 24 04:39:28 2022 -0600 Fix conjugate commit 9f7d199 Author: Denis Smirnov <[email protected]> Date: Tue Feb 22 11:56:05 2022 -0600 Add copy_when_nondefault_queue=False to call of get_dpnp_descriptor commit 9a93e16 Author: Denis Smirnov <[email protected]> Date: Mon Feb 21 05:21:06 2022 -0600 Add tests for compute follows data for math funcs commit d7783af Author: Denis Smirnov <[email protected]> Date: Fri Feb 4 04:42:29 2022 -0600 Extand main cython funcs with queues and events (part 2) commit 7bd11b2 Author: Denis Smirnov <[email protected]> Date: Fri Feb 4 03:09:25 2022 -0600 Extand main cython funcs with queues and events
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Intel channel must have higher priority then default channel in conda-build command.
Align channel ordering and conda-build command for both Linux and Windows.