Merge branch 'better_gpu_docs' into serialization_docs

Author: david-cortes-intel

doc/sources/array_api.rst

@@ -32,7 +32,7 @@ on GPU without moving the data from host to device.
     be :external+sklearn:doc:`enabled in scikit-learn <modules/array_api>`, which requires either changing
     global settings or using a ``config_context``, plus installing additional dependencies such as ``array-api-compat``.
 
-When passing array API inputs whose data is on a SyCL-enabled device (e.g. an Intel GPU), as
+When passing array API inputs whose data is on a SYCL-enabled device (e.g. an Intel GPU), as
 supported for example by `PyTorch <https://docs.pytorch.org/docs/stable/notes/get_start_xpu.html>`__
 and |dpnp|, if array API support is enabled and the requested operation (e.g. call to ``.fit()`` / ``.predict()``
 on the estimator class being used) is :ref:`supported on device/GPU <sklearn_algorithms_gpu>`, computations
@@ -50,10 +50,10 @@ through options ``allow_sklearn_after_onedal`` (default is ``True``) and ``allow
 
 If array API is enabled for |sklearn| and the estimator being used has array API support on |sklearn| (which can be
 verified by attribute ``array_api_support`` from :obj:`sklearn.utils.get_tags`), then array API inputs whose data
-is allocated neither on CPU nor on a SyCL device will be forwarded directly to the unpatched methods from |sklearn|,
+is allocated neither on CPU nor on a SYCL device will be forwarded directly to the unpatched methods from |sklearn|,
 without using the accelerated versions from this library, regardless of option ``allow_sklearn_after_onedal``.
 
-While other array API inputs (e.g. torch arrays with data allocated on a non-SyCL device) might be supported
+While other array API inputs (e.g. torch arrays with data allocated on a non-SYCL device) might be supported
 by the |sklearnex| in cases where the same class from |sklearn| doesn't support array API, note that the data will
 be transferred to host if it isn't already, and the computations will happen on CPU.
 

doc/sources/input-types.rst

@@ -50,6 +50,6 @@ enabled the input is unsupported).
   - For SciPy CSR matrix / array, index arrays are always copied. Note that sparse matrices in formats other than CSR
     will be converted to CSR, which implies more than just data copying.
   - Heterogeneous NumPy array
-  - If SyCL queue is provided for device without ``float64`` support but data are ``float64``, data are copied with reduced precision.
+  - If SYCL queue is provided for device without ``float64`` support but data are ``float64``, data are copied with reduced precision.
   - If :ref:`Array API <array_api>` is not enabled then data from GPU devices are always copied to the host device and then result table 
     (for applicable methods) is copied to the source device.

doc/sources/oneapi-gpu.rst

@@ -22,11 +22,11 @@ GPU support
 Overview
 --------
 
-|sklearnex| can execute computations on different devices (CPUs and GPUs, including integrated GPUs from laptops and desktops) supported by the SyCL framework.
+|sklearnex| can execute computations on different devices (CPUs and GPUs, including integrated GPUs from laptops and desktops) supported by the SYCL framework.
 
 The device used for computations can be easily controlled through the ``target_offload`` option in config contexts, which moves data to GPU if it's not already there - see :ref:`config_contexts` and rest of this page for more details).
 
-For finer-grained controlled (e.g. operating on arrays that are already in a given device's memory), it can also interact with on-device :ref:`array API classes <array_api>` like |dpnp_array|, and with SyCL-related objects from package |dpctl| such as :obj:`dpctl.SyclQueue`.
+For finer-grained controlled (e.g. operating on arrays that are already in a given device's memory), it can also interact with on-device :ref:`array API classes <array_api>` like |dpnp_array|, and with SYCL-related objects from package |dpctl| such as :obj:`dpctl.SyclQueue`.
 
 .. Note:: Note that not every operation from every estimator is supported on GPU - see the :ref:`GPU support table <sklearn_algorithms_gpu>` for more information. See also :ref:`verbose` to verify where computations are performed.
 
@@ -94,7 +94,7 @@ Target offload option
 
 Just like |sklearn|, the |sklearnex| can use configuration contexts and global options to modify how it interacts with different inputs - see :ref:`config_contexts` for details.
 
-In particular, the |sklearnex| allows an option ``target_offload`` which can be passed a SyCL device name like ``"gpu"`` indicating where the operations should be performed, moving the data to that device in the process if it's not already there; or a :obj:`dpctl.SyclQueue` object from an already-existing queue on a device.
+In particular, the |sklearnex| allows an option ``target_offload`` which can be passed a SYCL device name like ``"gpu"`` indicating where the operations should be performed, moving the data to that device in the process if it's not already there; or a :obj:`dpctl.SyclQueue` object from an already-existing queue on a device.
 
 .. hint:: If repeated operations are going to be performed on the same data (e.g. cross-validators, resamplers, missing data imputers, etc.), it's recommended to use the array API option instead - see the next section for details.
 
@@ -114,7 +114,7 @@ Example:
                model.fit(X, y)
                pred = model.predict(X)
 
-    .. tab:: Passing a SyCL queue
+    .. tab:: Passing a SYCL queue
        .. code-block:: python
 
            import dpctl
@@ -139,7 +139,7 @@ Example:
 GPU arrays through array API
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
-As another option, computations can also be performed on data that is already on a SyCL device without moving it there if it belongs to an array API-compatible class, such as |dpnp_array| or `torch.tensor <https://docs.pytorch.org/docs/stable/tensors.html>`__.
+As another option, computations can also be performed on data that is already on a SYCL device without moving it there if it belongs to an array API-compatible class, such as |dpnp_array| or `torch.tensor <https://docs.pytorch.org/docs/stable/tensors.html>`__.
 
 This is particularly useful when multiple operations are performed on the same data (e.g. cross validators, stacked ensembles, etc.), or when the data is meant to interact with other libraries besides the |sklearnex|. Be aware that it requires enabling array API support in |sklearn|, which comes with additional dependencies.
 
@@ -230,7 +230,7 @@ Example:
     model.fit(X, y)
 
 
-Note that, if array API had been enabled, the snippet above would use the data as-is on the device where it resides, but without array API, it implies data movements using the SyCL queue contained by those objects.
+Note that, if array API had been enabled, the snippet above would use the data as-is on the device where it resides, but without array API, it implies data movements using the SYCL queue contained by those objects.
 
 .. note::
     All the input data for an algorithm must reside on the same device.