summaryrefslogtreecommitdiff
path: root/api
Commit message (Collapse)AuthorAgeFilesLines
* Add rsDebug for halfPirama Arumuga Nainar2016-04-041-0/+9
| | | | | | | | http://b/27556050 Update API generator specification and generated rs_debug.rsh file. Change-Id: Icc731e16c142720d5cca5523a904d800f67f6253
* Enable tests for native_{expm1,tan,tanpi}Pirama Arumuga Nainar2016-03-311-5/+5
| | | | | | Bug: http://b/27512382 Change-Id: Ic19e1de9cd80e87453e28ad4f709ac7060f532c2
* For custom tests, initialize double counterpart of Float16 variablePirama Arumuga Nainar2016-03-312-0/+14
| | | | | | | | | | Bug: http://b/27512382 For functions with 'custom' test label, initialize an 'outDouble' variable which has the 'short' output value 'out' converted to double. Also, print this converted value in case of error. Change-Id: Iff4123f2103bb9ccad2028bd170f95ce3ebb8515
* Enable tests for Float16 clampPirama Arumuga Nainar2016-03-311-2/+0
| | | | | Bug: http://b/27512382 Change-Id: Ibb8c69d6bed07c1551669144e8cf8d710287d02f
* Merge "Enable tests for some float16 native_* functions" into nyc-devPirama Arumuga Nainar2016-03-311-22/+27
|\
| * Enable tests for some float16 native_* functionsPirama Arumuga Nainar2016-03-311-22/+27
| | | | | | | | | | | | | | | | | | Bug: http://b/27512382 Enable tests for native_* functions whose specification allows an absolute error value. Change-Id: I06e3598216dd86399997ed9c503bf68e736f7ee3
* | Merge "Fix typo in Conversion Functions doc" into nyc-devBrian Carlstrom2016-03-301-1/+1
|\ \
| * | Fix typo in Conversion Functions docBrian Carlstrom2016-03-291-1/+1
| | | | | | | | | | | | | | | Bug: 27529815 Change-Id: Ib024501bddf35e7897e490338b407c289222c672
* | | Merge "Update rs_object_types.spec about USAGE_GRAPHICS_TEXTURE." into nyc-devMiao Wang2016-03-301-1/+1
|\ \ \ | |_|/ |/| |
| * | Update rs_object_types.spec about USAGE_GRAPHICS_TEXTURE.Miao Wang2016-03-291-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | Bug: 25227549 - Remove "Dreprecated" and add description to it, since it is still being used. Change-Id: Ifaac2756bd1d4858893443d4c11641bc5f5767a1
* | | Add convert_halfN(halfN) variants to spec and rs_convert.rshPirama Arumuga Nainar2016-03-291-1/+1
| | | | | | | | | | | | | | | Bug: http://b/27529996 Change-Id: I894818cdd7ebef1172e087be0deab851abebb983
* | | Enable tests for Float16 native_distancePirama Arumuga Nainar2016-03-281-1/+1
| |/ |/| | | | | | | Bug: http://b/27547875 Change-Id: I439debafaa70edadc62cce3af07368239d3c5799
* | Merge changes I0527dd5c,Ib8e76b7e into nyc-devPirama Arumuga Nainar2016-03-281-0/+11
|\ \ | | | | | | | | | | | | | | | * changes: Reorder tests for Float16 vector functions Add native_distance for half.
| * | Add native_distance for half.Pirama Arumuga Nainar2016-03-281-0/+11
| |/ | | | | | | | | | | | | | | | | | | | | Bug: http://b/27547875 Add native_distance to the .spec files and update the auto-generated rs_vector_math.rsh. The definition already exists in driver/runtime/rs_cl.c Change-Id: Ib8e76b7e784c826d060e462859cb33559d66fe32
* | Enable tests for Float16 native_* functionsPirama Arumuga Nainar2016-03-251-46/+11
| | | | | | | | | | Bug: http://b/27512382 Change-Id: Ia7e9f7e1216d879d2c4afdb1fb0838e12ddb75c5
* | Enable tests for float16 math functionsPirama Arumuga Nainar2016-03-241-79/+8
| | | | | | | | | | Bug: http://b/27512382 Change-Id: I63ba284ec033653ce6445c4615ddf7771e390650
* | Enable tests for FP16-related convert_* functions.Pirama Arumuga Nainar2016-03-241-2/+0
|/ | | | | | http://b/27512382 Change-Id: Ib38a07c5a24816bc5857162c184e6b1e18c7edef
* Enable tests for vector float16 math functionsPirama Arumuga Nainar2016-03-251-6/+6
| | | | | Bug: http://b/27512382 Change-Id: I2c45fed9d78183a22f7e277034d41160970e0fe1
* Initialize Double field for Float16 parameterPirama Arumuga Nainar2016-03-231-0/+7
| | | | | | | | | Bug: http://b/27512382 Add missing initialization of the Double field corresponding to Float16 parameter to vector parameters when vector size is one. Change-Id: Ida81ebc6151019df164dede5f769d79d60c32afa
* Prevent truncation in maxDoubleForIntegerPirama Arumuga Nainar2016-03-221-2/+2
| | | | | | | | | | Bug: http://b/27512382 Use uint64_t instead of 'unsigned long' as the latter's size may not always be 64-bits. Doing so makes maxDoubleForInteger return the expected result. Change-Id: I1db0af0e11ff55a99994b6b81a9cb5d18d525e7c
* Merge "Add documentation on invoke-only APIs" into nyc-devYang Ni2016-03-222-0/+16
|\
| * Add documentation on invoke-only APIsYang Ni2016-03-222-0/+16
| | | | | | | | | | | | | | | | | | | | Bug: 22872577 Four Allocation related runtime API functions are not supposed to be called from inside a kernel. Edit the documentation to clarify this restriction. Change-Id: I6766a216627fe485d2429c5455c02d0a151c68f8
* | Enable tests for Float16 cross, cos, floor, rsqrtPirama Arumuga Nainar2016-03-072-4/+1
|/ | | | | | | | | | http://b/27512382 Enable select Float16 tests to validate changes to CoreMathVerifier and Target in cts/. Other tests will be enabled once the current changes get reviewed and submitted. Change-Id: I679353ac4bd5fb64b4e1fece03bc7380cd82ced3
* Extend test generator to generate tests for Float16Pirama Arumuga Nainar2016-03-073-5/+58
| | | | | | | | | | | | http://b/27512382 Extend the test generator to add a short and double field to the Arguments class for Float16 parameters. Input short values get converted to double while initializating the Arguments class and output short values get converted to double before invoking Target.Floaty.couldBe(). Change-Id: I9848cb4b2ba7b79c5b77d849a4c48e827d24f64c
* Merge "Do not test rsForEach and rsForEachWithOptions in the all-api test." ↵Pirama Arumuga Nainar2016-03-031-0/+3
|\ | | | | | | into nyc-dev
| * Do not test rsForEach and rsForEachWithOptions in the all-api test.Pirama Arumuga Nainar2016-02-251-0/+3
| | | | | | | | | | | | | | | | | | http://b/27358969 This patch skips both rsForEach and rsForEachWithOptions from the all-api test. Change-Id: I270be38a5a4f4e6cd25f42e5000d85b38ea0a4df
* | Stop generate all_api.rs for CTSMiao Wang2016-02-251-2/+0
|/ | | | | Bug: 27358969 Change-Id: I186547f79a0fec9ccfe6b9075ca779b33467c1b5
* Fix rs_quaternion.spec for API level 24.Stephen Hines2016-02-231-12/+12
| | | | | | | | | | A recent change made these functions non-inline (so that implementations could accelerate it differently). This change makes things more consistent, since we don't want a potential missing set of functions. (cherry picked from commit b1bc853f19a2db925edbcfdd393b3a7def42604e) Change-Id: I457145ed408ce46c5d013c1592760a1f9e15efd8
* Pass new parameters while generating Target objectPirama Arumuga Nainar2016-02-191-4/+41
| | | | | | | | | http://b/23536224 While generating tests, identify and pass FunctionType and ReturnType parameters to Target's constructor. Change-Id: I0e1be6279d21028175571d53a7def7fc197778eb
* Assign API level 24 to f16 math functionsPirama Arumuga Nainar2016-02-183-129/+129
| | | | | | http://b/23536224 Change-Id: Iae4cb0b333bc2c42cab63f01b2bf4e2c0c01a2f1
* Add length, normalize and their native variants for f16Pirama Arumuga Nainar2016-02-181-2/+40
| | | | | | | | | | | | http://b/23536224 https://android-review.googlesource.com/#/c/186326/ included implementations for length, native_length, normalize and native_normalize in driver/runtime. I had missed adding them to the API. This CL adds specs for these functions. Change-Id: Idaefb7ec2e45adb7b0e4b502404b1098f4c9d8c9 (cherry picked from commit 09311ef98909f1c759239cf0b1e258e93feca6e9)
* Add RS_TYPE_FLOAT_16 and allocation helpers for f16Pirama Arumuga Nainar2016-02-182-8/+8
| | | | | | | | | | http://b/24862914 - Add RS_TYPE_FLOAT_16 value to rs_data_type enum. - Add helper functions to create f16 allocations. Change-Id: I937a1de6f1d80ee50521edd106962e243886198e (cherry picked from commit f4fee99a12a915fca447370bcae85a40b23cb8d5)
* Move implementations of quaternion functions into driver/runtime.Verena Beckham2016-02-111-0/+121
| | | | | | | | | | | | | | Move the implementation of all quaternion functions from rs_math.rsh driver/runtime. The old implementation remains in place for older versions, and the move occurs only for an UNRELEASED API level. Added a new file driver/runtime/rs_quaternion.c to contain the implementations. Signed-off-by: Verena Beckham <verena@codeplay.com> (cherry picked from commit 336fc4b3fb2df22623d51d9bbb996d11cd69f16c) Change-Id: Ib37fad0a911d05640f7b2b9e88b25cd0053108c1
* Fix function mangling of rs_quaternion* arguments in stubs white list.Dean De Leo2016-02-113-18/+51
| | | | | | | | | | | | | | | | | Fix the generator of the RenderScript header files to allow functions that take rs_quaternion arguments. Previously they were wrongly mangled in the RenderScript stubs white list (float* instead of float4*). Factor out the calculation of vector length from the runtime header generator, and call it after expanding a typedef to update the vector length. This way it can calculate the right mangling for a function taking a pointer to rs_quaternion (which is not itself a vector but the mangling does need to output a vector length, i.e. Dv4_f). This fix is required by a following change list, which moves the bodies of the quaternion builtins to libclcore, so their mangled names are then added to the white list. (cherry picked from commit 1f08801f894a04d3f9221e7a85a93de1e3215def) Change-Id: If90d402830631e144afa91d6004f710fead89036
* Add fract, frexp, modf for halfPirama Arumuga Nainar2016-02-081-2/+39
| | | | | | | | http://b/23536224 (cherry picked from commit 03ad5f9d481843c6884e7fabcd085bf19fba64a7) Change-Id: I11ac519865f8f2dc595ca8d356c5df807d0e1ed5
* Merge changes from topic \'fp16-math-update\' am: 20807a72d7Pirama Arumuga Nainar2016-02-051-3/+32
|\ | | | | | | | | | | | | | | | | | | am: 9084aebf5c * commit '9084aebf5c4047f61cb02602de071714da286360': Enable half tests for copysign, ilogb, nextafter Implement copysign, ilogb and nextafter for half Update RS headers with copysign, ilogb and nextafter Add copysign, ilogb and nextafter for half
| * Add copysign, ilogb and nextafter for halfPirama Arumuga Nainar2016-02-041-3/+32
| | | | | | | | | | | | http://b/23536224 Change-Id: Ib760a4120ec3ca3b1ad0459bd0e35398fd83bc4f
* | Update the legal notice: change year to 2016Miao Wang2016-02-021-1/+1
| | | | | | | | Change-Id: I2ed106fa112ba57baba10ce7208de45876cf713f
* | Auto generate .rs script that has all the APIs in N for CTS tests.Miao Wang2016-02-021-1/+6
| | | | | | | | Change-Id: I8aebea2205a23ea70f32d6f1f19a1d5cba705bcf
* | Merge "Avoid casting varargs" am: 9233f7323aYang Ni2016-02-021-1/+4
|\| | | | | | | | | | | | | am: 60c8a9c1fa * commit '60c8a9c1fadc323c4d458738ceae20916f45f6e1': Avoid casting varargs
| * Avoid casting varargsYang Ni2016-02-011-1/+4
| | | | | | | | | | | | | | | | | | | | Bug: 26912071 The auto generating tool for API should not cast arguments of varargs. Doing so would be incorrect code in the auto-generated test calls for RS API. Change-Id: I88637b306cd4224e2018f5a0dbc2683b59b86f1f
* | Merge "Fixed rsQuaternionAdd() implementation" am: c17bf3ff97Yang Ni2016-02-011-4/+4
|\| | | | | | | | | | | | | am: f772e7893c * commit 'f772e7893ce71f43dc5df2a214213a2e919b610c': Fixed rsQuaternionAdd() implementation
| * Fixed rsQuaternionAdd() implementationYang Ni2016-02-011-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Bug: 26272685 We use float4 to represent a quaternion, in which a float4 q = (x,y,z,w) represents quaternion wi + xj + yk + zl. According to Wikipedia and Wolfram, the addition of two quaternions q and r should be defined as the following (q.w+r.w)i + (q.x+r.x)j + (q.y+q.y)k + (q.z+r.z)l, using the above representation of quaternions. https://en.wikipedia.org/wiki/Quaternion http://mathworld.wolfram.com/Quaternion.html Change-Id: I07837b266f51b664bed2dc50b9ff71ba375f5aaf
* | Merge "Move implementations of matrix functions into driver/runtime." am: ↵Stephen Hines2016-01-311-0/+31
|\| | | | | | | | | | | | | | | | | e961168a5f am: eabdaf0133 * commit 'eabdaf01332a6194c01cedcf1f53902d50a78eae': Move implementations of matrix functions into driver/runtime.
| * Move implementations of matrix functions into driver/runtime.Verena Beckham2016-01-301-0/+31
| | | | | | | | | | | | | | | | | | | | Move the implementation of rsExtractFrustumPlanes and rsIsSphereInFrustum from rs_math.rsh into driver/runtime. The old implementation remains in place for older versions, and the move occurs only for API level 24 and up. Change-Id: I7661e71eb5d344e39fa84e0aa502c621936e68a0 Signed-off-by: Verena Beckham <verena@codeplay.com>
* | Merge "Move implementation of fract(float) into driver/runtime." am: 0298e7a7a7Stephen Hines2016-01-311-1/+9
|\| | | | | | | | | | | | | am: 6659c1873b * commit '6659c1873b9a4234c401ee5ac941f288e20c51c0': Move implementation of fract(float) into driver/runtime.
| * Move implementation of fract(float) into driver/runtime.Verena Beckham2016-01-301-1/+9
| | | | | | | | | | | | | | | | | | Move the implementation of the float variant of fract from rs_math.rsh into driver/runtime. The old implementation remains in place for older versions, and the move occurs only for API level >= 24. Signed-off-by: Verena Beckham <verena@codeplay.com> Change-Id: Id94e7191604d5ccaee41b6634731ee997bf01983
* | Merge "Remove the always_inline attribute on all RS builtins." am: 3b414d79c2Stephen Hines2016-01-303-6/+1
|\| | | | | | | | | | | | | am: 8871aaed33 * commit '8871aaed33ff8b1764ecdde06ecc35107f7b462c': Remove the always_inline attribute on all RS builtins.
| * Remove the always_inline attribute on all RS builtins.Verena Beckham2016-01-303-6/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Remove the always_inline attribute on all RenderScript builtins in the specification files. This leads to the removal of that attribute from the declarations of all those functions in the RenderScript header files. Also, updates a comment in the unrelated file rs_for_each.rsh. I presume someone forgot to update that when they changed the corresponding spec file. Signed-off-by: Verena Beckham <verena@codeplay.com> Change-Id: Ia4cdffdbdc9866a5b961356218dd01732bf2e830
* | Assign API level 24 to single-source RS APIYang Ni2016-01-223-23/+23
|/ | | | | | | | | | | Updated the RenderScript API spec files, and reran the auto-generating script for the single-source RS API. Also updated LatencyBenchmark makefile to set its target-api to 23, since it uses the deprecated rsForEach() function that takes a script parameter. Change-Id: I75d36010be49a0cdcbe1d6ec1fc7eed3a6d7b12b
generated by cgit v1.2.3 (git 2.25.1) at 2025-12-22 08:19:26 +0000