|
38 | 38 | - [Other restrictions](#other-restrictions) |
39 | 39 |
|
40 | 40 | ## Other content: |
41 | | -* [ESIMD API/doxygen reference](https://intel.github.io/llvm-docs/doxygen/group__sycl__esimd.html) |
42 | 41 | * [Examples](./examples/README.md) |
43 | 42 | * [ESIMD LIT tests - working code examples](https://github.com/intel/llvm/blob/sycl/sycl/test-e2e/ESIMD/) |
44 | 43 |
|
@@ -162,7 +161,6 @@ The element type must either be a vectorizable type or the `sycl::half` type. |
162 | 161 | The set of vectorizable types is the |
163 | 162 | set of fundamental SYCL arithmetic types excluding `bool`. The length of the |
164 | 163 | vector is the second template parameter. |
165 | | -See the complete [API reference](https://intel.github.io/llvm-docs/doxygen/classcl_1_1____ESIMD__NS_1_1simd.html#details) for the `simd` class for more details. |
166 | 164 |
|
167 | 165 | ESIMD compiler back-end does the best it can to map each `simd` class object to a |
168 | 166 | contiguous block of registers in the general register file (GRF). |
@@ -351,8 +349,6 @@ reduction operations are supported: |
351 | 349 | - maximum |
352 | 350 | - minimum |
353 | 351 |
|
354 | | -See more details on the API documentation [page TODO](https://intel.github.io/llvm-docs/doxygen). |
355 | | - |
356 | 352 | ### Memory access APIs |
357 | 353 |
|
358 | 354 | Explicit SIMD memory access interface is quite different from the standard SYCL |
@@ -598,9 +594,6 @@ Many memory access APIs accept offsets as arguments, which are used to determine |
598 | 594 | actual memory location for the access. Offsets are always expressed in bytes |
599 | 595 | rather than element units. |
600 | 596 |
|
601 | | -See more details in the API documentation |
602 | | -[page TODO](https://intel.github.io/llvm-docs/doxygen). |
603 | | -
|
604 | 597 | ### Math operations |
605 | 598 |
|
606 | 599 | #### Extended math |
@@ -649,10 +642,6 @@ ESIMD supports the following non-standard math functions implemented in hardware |
649 | 642 | - Plane equation - `plane`. Solves a component-wise plane equation |
650 | 643 | `w = p*u + q*v + r` where `u`, `v`, `w` are vectors and `p`, `q`, `r` are scalars. |
651 | 644 |
|
652 | | -
|
653 | | -See more details in the API documentation |
654 | | -[page TODO](https://intel.github.io/llvm-docs/doxygen). |
655 | | -
|
656 | 645 | ### Dot Product Accumulate Systolic - `DPAS` API |
657 | 646 |
|
658 | 647 | DPAS is the matrix multiply-add-and-accumulate operation performed on limited size matrices/tiles. |
@@ -870,9 +859,6 @@ There are other useful miscellaneous APIs provided by ESIMD. |
870 | 859 | - Find least significant set bit - `fbl`. |
871 | 860 | - Find most significant set bit - `fbh`. |
872 | 861 |
|
873 | | -See more details in the API documentation |
874 | | -[page TODO](https://intel.github.io/llvm-docs/doxygen). |
875 | | - |
876 | 862 | <br> |
877 | 863 |
|
878 | 864 | ### Private Global Variables. |
|
0 commit comments