af586bbbf2
Checkin to fix bugs in math functions.
...
This change fixes the following bugs that were discovered while debuggnig TF unit test failures (cwise_ops_test)
1. __hisinf and __hisnan routines
Both had incorrect implementations.
2. abs
A "long long" (64bit int) version was missing, resulting in the 32bit version being used for 64bit ints (which resulted in incorrect results, when the value passed in was outside the 32bit int range)
3. lgamma
We seemed to have a custom version for the 'double' datatype (which was giving incorrect results). Replaced it with a call to the 'double' version of the underlying 'hc::precision_math::lgamma'
2018-04-24 18:10:07 +00:00
3cfb9c0d40
Added hipMemset2DAsync support
2018-04-17 18:27:27 +05:30
bda18af06d
Merge pull request #395 from lmoriche/master
...
Add missing __device__ function attributes.
2018-04-12 14:51:32 +05:30
211c18e674
Add missing __device__ function attributes.
2018-04-11 09:29:37 -07:00
ac7c5dfd0b
Merge pull request #388 from lmoriche/master
...
Add support for clang offload bundles and <<<>>> kernel launch.
2018-04-09 12:40:28 +05:30
f48e471dd4
adjust the hcc work week for the texture fix
2018-04-06 15:46:55 -04:00
bc6fa2bc62
Don't include __clang_cuda_builtin_vars.h
2018-04-05 16:27:05 -07:00
0c1893ea62
Add support for clang offload bundles and <<<>>> kernel launch.
2018-04-04 13:32:33 -07:00
62fc6c85e0
Fix texture driver api TRFS flags
2018-04-02 21:47:20 +05:30
0d8b97274a
Change constant address space to 4
...
when compiling with a newer hcc. This is
due to a recent change to address space
mapping in the amdgpu compiler backend.
2018-03-26 18:13:59 -04:00
a5e04c9706
Merge remote-tracking branch 'upstream/master'
2018-03-21 11:17:52 -07:00
a06b21c500
Add HIP_KERNEL_NAME/HIP_SYMBOL definitions for HIP-clang, and rename hipLaunch->hipLaunchByPtr.
2018-03-21 11:07:21 -07:00
1ba06f63c4
Apply .clangformat to all repo source files
...
Change-Id: I7e79c6058f0303f9a98911e3b7dd2e8596079344
2018-03-12 11:29:03 +05:30
18e70b1e6b
Merge pull request #363 from gargrahul/surface_object_api
...
Added surface object support
2018-03-09 11:50:30 +05:30
1e13ebd992
Merge pull request #288 from AlexVlx/feature_purge_genco
...
Purge hsagenco.sh
2018-03-05 09:59:56 +05:30
d697739d57
Fixed byte offset issue
...
Added HIP/NVCC support
2018-03-04 19:05:37 +05:30
0d45c1d2d3
__CUDA__ and __HIP__ are both defined when compiling with 'clang -x hip', so make sure __HIP__ is not defined in the case of __HIP_PLATFORM_NVCC__.
2018-03-02 06:51:51 -08:00
6536164914
Use __HIP__ instead of __HIPCC__, and __CUDA__ instead of __CUDACC__
2018-02-28 14:20:55 -08:00
2bc7adfa62
Fixes a build error with hcc
2018-02-28 14:12:09 -08:00
f8b0c96462
Add initial support for clang
2018-02-28 12:31:26 -08:00
5c3b91a0a1
Added surface object support
2018-02-26 11:59:03 +05:30
9526e05732
Missing merge.
2018-02-12 20:21:37 +00:00
baf50a5311
Re-sync with upstream.
2018-02-12 20:20:24 +00:00
4b8ae78891
Merge pull request #321 from gargrahul/hipMemcpyArray_Functions
...
Added support for hipMemcpy Array functions-
2018-02-12 10:36:38 +05:30
24ab820a11
Fixed host allocated globals address lookup for host usage
...
Fixed texture driver APIs failure
2018-01-30 18:06:31 +05:30
72d3e679e4
Merge branch 'master' into feature_purge_genco
2018-01-29 16:02:03 +05:30
44de3f8309
Merge pull request #346 from scchan/fix_rhel_build
...
use assign rather than insert
2018-01-26 06:59:25 +05:30
8acc8365fa
use assign rather than insert
2018-01-25 16:51:29 +00:00
0497424978
Fixing rocblas build failure
...
with ::Bundled_code_header constructor
Disabling hipPrintfKernel test from CI
2018-01-25 10:29:40 -06:00
ed75522ba5
Merge branch 'master' of https://github.com/ROCm-Developer-Tools/HIP into feature_purge_genco
2018-01-17 14:02:19 +00:00
487a430b5a
Added support for -
...
- hipMemcpyFromArray
- hipMemcpyAtoH
- hipMemcpyHtoA
2018-01-16 11:44:19 +05:30
814d55f046
Merge pull request #282 from gargrahul/texture_driver_3d_support
...
Added support for 3D texture driver apis
2017-12-18 15:39:26 +05:30
e14ac87474
Merge branch 'master' of https://github.com/ROCm-Developer-Tools/HIP into feature_purge_genco
2017-12-14 13:50:49 +00:00
6cb5e92d75
Merge pull request #286 from gargrahul/fix_hipDeviceGetAttribute_nvcc
...
Fix hipGetDeviceAttribute dtest for HIP/NVCC
2017-12-12 12:49:23 +05:30
6293d3f887
Merge pull request #285 from aaronenyeshi/fix-ilogb-unreachable
...
Fix ilogb/ilogbf functions to return int
2017-12-12 10:47:33 +05:30
b842394957
This introduces LipoProteinLipase (lpl), a simple tool for creating fat binaries. It represents a direct replacement of the creaky hccgenco.sh script, which had various issues. The format it uses is that of a code object bundle, generated by the Clang Offload Bundler. The output is always suffixed with the ".adipose" extension. It is shared with HCC. The hipcc script and associated tests are modified to use lpl. Help can be obtained by invoking lpl --help. A more computer-sciency / corporate friendly name is likely to be beneficial, which is a reason for choosing easily searchable/replaceable names such as lpl or adipose.
2017-12-08 04:22:57 +00:00
0782c52271
Fix hipGetDeviceAttribute dtest for HIP/NVCC
2017-12-06 15:49:06 +05:30
707aaef99d
Fix ilogb/ilogbf functions to return int
...
This patch will fix hipDoublePrecisionMathDevice test on ThinLTO, which uncovered that hip math_function's ilogb/ilogbf should return type int instead of double. This will match rocdl.
2017-12-05 23:14:10 +00:00
115c7f2b79
Added support for
...
- 3D texture driver APIs
- hipMalloc3D
- hipMemcpy3D for destination other than array
2017-12-05 14:11:13 +05:30
5127ce67e8
This is primarily intended as an additional cleanup of the module functionality, in the aftermath of adopting module based dispatch. The main effort was associated with refactoring the questionable ihipModuleGetSymbol. It was quaintly written and misleading, in that it had little to do with getting symbols, and was exactly retrieving a kernel object. Error handling is modified so as to reduce branching depth. Functions which serve as interfaces to the HSA RT are moved in a separate helper header. Code object readers are properly deleted. Some leftover dead functionality pertaining to associating namespace scope variables with their allocated memory is removed. Executable loading is changed to use a string which holds the ELF image of the code object being loaded, thus avoiding some corner cases where using a istream would fail.
2017-12-03 23:09:06 +00:00
7c0b9a005b
Fix legacy mode detection of the address of an agent allocated variable. In this mode, there exist two executables per each code object, one created by HCC and one created by HIP. Since we dispatch through HCC in legacy mode, we should obtain the address for an agent allocated variable from the latter's executable. Also add two omitted validity checks, whose absence could lead to segfaults when the current process had no .kernel section and / or when an invalid or empty blob was extracted from the latter.
2017-11-30 03:29:04 +00:00
20fc68c9a1
Add missing space (the final frontier).
2017-11-29 21:50:43 +00:00
32e11e7dc6
Revert "Revert adoption of CUDA indexing in general - this can only work with later versions of the compiler, just like module based dispatch, and thus must be guarded against usage in earlier (e.g. 1.6) versions."
...
This reverts commit d2fd1f5
2017-11-29 21:49:10 +00:00
7acb1e6ff6
Use a much simpler guard for version 1.6, which allows for direct CUDA indexing to be used.
2017-11-29 21:47:04 +00:00
d2fd1f5544
Revert adoption of CUDA indexing in general - this can only work with later versions of the compiler, just like module based dispatch, and thus must be guarded against usage in earlier (e.g. 1.6) versions.
2017-11-29 21:01:28 +00:00
b881cf713c
Fix compiler version check.
2017-11-29 03:05:53 +00:00
faa546d194
Fix oversight in selection mechanism which led to erroneous code to be compiled for the grid_launch_GGL component.
2017-11-29 01:37:52 +00:00
89e9399427
Choose whether or not to use functional grid_launch based on the version of HCC used to compile.
2017-11-29 00:17:44 +00:00
6e4ca3fbb4
Change memset kernel to use memcpy instead of placement new. Simplify indexers.
2017-11-28 19:45:47 +00:00
02c2bfc7ef
Re-sync with upstream and re-factor platform global management for texture references.
2017-11-28 19:15:29 +00:00
Deven Desai
Rahul Garg
Maneesh Gupta
Laurent Morichetti
Siu Chi Chan
Alex Voicu
Kent Knox
Aaron En Ye Shi