Commit Graph

11 Commits

Author SHA1 Message Date
Ben Ashbaugh
8d4a870059 fix correctly rounded behavior for math bruteforce tests (#2397)
fixes #2387 

Corrects the "correctly rounded" behavior for the math bruteforce tests.
Specifically:

* Only applies the `-cl-fp32-correctly-rounded-divide-sqrt` build option
for the `divide_cr` and `sqrt_cr` tests. The other tests do not receive
this build option. This means that there is a difference in the behavior
of the `divide` and `divide_cr` tests and the `sqrt` and `sqrt_cr`
tests, and the "correctly rounded" build option is not applied to the
fp16 or fp64 tests.
* Removes the build option to toggle testing the correctly rounded
divide and square root tests since it no longer needed. Instead, the
test names can be used to choose whether to test the correctly rounded
functions or the non-correctly rounded functions.

Additionally:

* Relaxes the fp16 sqrt accuracy requirements to 1 ULP. This is needed
to pass this test on some of our devices. This part is still under
discussion, so I will keep this PR as a draft until it is settled.
2025-07-15 09:01:19 -07:00
Sven van Haastregt
9116bb7acb [NFC] math_brute_force: move TestInfoBase to common.h (#2059)
The various forms of `TestInfoBase` have many members in common, so
avoid duplicating the struct definition and move it to `common.h`.

Provide a description and initializer for every struct member, and drop
initializations done with `memset`.

Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
2024-09-03 18:30:13 +01:00
Sven van Haastregt
b6941b6c61 Add fp16 testing to conversions and bruteforce (#1975)
Merge the `fp16-staging` branch into `main`, adding fp16 (`half`)
testing to the conversions and math bruteforce tests.

---------

Signed-off-by: Ahmed Hesham <ahmed.hesham@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
Signed-off-by: Guo, Yilong <yilong.guo@intel.com>
Signed-off-by: John Kesapides <john.kesapides@arm.com>
Co-authored-by: Marcin Hajder <marcin.hajder@gmail.com>
Co-authored-by: Ewan Crawford <ewan@codeplay.com>
Co-authored-by: Wawiorko, Grzegorz <grzegorz.wawiorko@intel.com>
Co-authored-by: Sreelakshmi Haridas Maruthur <sharidas@quicinc.com>
Co-authored-by: Harald van Dijk <harald@gigawatt.nl>
Co-authored-by: Ben Ashbaugh <ben.ashbaugh@intel.com>
Co-authored-by: Haonan Yang <haonan.yang@intel.com>
Co-authored-by: Ahmed Hesham <117350656+ahesham-arm@users.noreply.github.com>
Co-authored-by: niranjanjoshi121 <43807392+niranjanjoshi121@users.noreply.github.com>
Co-authored-by: Wenwan Xing <wenwan.xing@intel.com>
Co-authored-by: Yilong Guo <yilong.guo@intel.com>
Co-authored-by: Romaric Jodin <89833130+rjodinchr@users.noreply.github.com>
Co-authored-by: joshqti <127994991+joshqti@users.noreply.github.com>
Co-authored-by: Pekka Jääskeläinen <pekka.jaaskelainen@tuni.fi>
Co-authored-by: imilenkovic00 <155085410+imilenkovic00@users.noreply.github.com>
Co-authored-by: John Kesapides <46718829+JohnKesapidesARM@users.noreply.github.com>
Co-authored-by: Aharon Abramson <aharon.abramson@mobileye.com>
2024-06-18 09:43:11 -07:00
Sven van Haastregt
6ecf824122 math_brute_force: Use clKernelWrapper in threaded tests (#1562)
Simplify code by relying on RAII to free resources.  Reduce code
duplication.

This commit only affects tests that use `BuildKernelInfo`, which are
the multi-threaded tests.  Another patch will deal with the
single-threaded tests, i.e., those using `BuildKernelInfo2`.

Original patch by Marco Antognini.

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
2022-11-27 13:25:32 -08:00
Sven van Haastregt
63274f97b7 math_brute_force: Factor out GetUnaryKernel and GetBinaryKernel (#1525)
Use common functions to create the kernel source code for testing
1-argument and 2-argument math builtins.  This reduces code duplication.

Use appropriate patterns to initialise variables to their full bit
widths.  For example, `0xdead` was previously used to initialise 32-bit
integers, while now a larger number spanning all bytes is used.

Co-authored-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
Co-authored-by: Marco Antognini <marco.antognini@arm.com>
2022-11-01 13:08:36 -07:00
Sven van Haastregt
d9a938b698 Factor out GetTernaryKernel (#1511)
Use a common function to create the kernel source code for testing
3-argument math builtins.  This reduces code duplication.  1-argument
and 2-argument math kernel construction will be factored out in future
work.

Change the kernels to use preprocessor defines for argument types and
undef values, to make the CTS code easier to read.

Co-authored-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
Co-authored-by: Marco Antognini <marco.antognini@arm.com>
2022-10-04 09:28:29 -07:00
Sven van Haastregt
f94c135755 Use clMemWrapper in math_brute_force (#1476)
Simplify code by avoiding manual resource management.

Original patch by Marco Antognini.

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
2022-08-31 22:43:43 -07:00
Sven van Haastregt
c12bff46c6 Tidy up BuildKernelInfo (#1461)
Remove the `offset` field from both structures, because it was always
set to the global `gMinVectorSizeIndex`.

Improve documentation and rename some variables:
 - `i` becomes `vectorSize`;
 - `kernel_count` becomes `threadCount`.

Original patch by Marco Antognini.

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
2022-08-04 15:03:52 +01:00
Sven van Haastregt
b06ccc6cd9 Share BuildKernelInfo struct definition (#1453)
Move the main `BuildKernelInfo` definition into `common.h` to reduce
code duplication.

Some tests (e.g. `i_unary_double.cpp`) use a different struct; rename
those structs to `BuildKernelInfo2` for now to avoid ambiguity.

Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
2022-07-29 18:22:18 +01:00
Sven van Haastregt
0a5a8f90c9 Use clProgramWrapper in math_brute_force (#1451)
Simplify code by avoiding manual resource management.

This allows removing clReleaseProgram from `MakeKernels` to reduce
behavioral differences between `MakeKernels` and `MakeKernel`.

Original patch by Marco Antognini.

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
Signed-off-by: Sven van Haastregt <sven.vanhaastregt@arm.com>
2022-07-28 13:33:16 +01:00
Marco Antognini
ed839ebf10 Avoid manual memory management (#1260)
* Avoid manual memory management

Prefer std::vector over malloc and free. This will allow removing goto
statements by leveraging RAII.

Use appropriate type (bool) to store overflow predicates and allocate
std::vector<bool> of appropriate sizes: before this change the
allocation was unnecessary bigger than required.

No longer attempt to catch "out of host memory" issues, given that in
such situation it is generally not possible to cleanly report an error.
Rely on std::bad_alloc exception to report such issues.

Introduce a new header for common code in the math_brute_force
component. It is currently complementary to utility.h and is expected to
hold cleaned up content extracted from future refactoring operations.

List all headers as source in CMake for better compatibility with IDEs.

Signed-off-by: Marco Antognini <marco.antognini@arm.com>

* Remove manual or unnecessary memset

In order to use non-POD types as fields of TestInfo, memset must be
replaced with a compatible zero-initialisation.

Remove an unnecessary memset in MakeKernels.

Signed-off-by: Marco Antognini <marco.antognini@arm.com>
2021-05-24 16:34:54 +01:00