AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
2,825 Commits 3 Branches 0 Tags
bbcb3fa0c0258faa692502ba82af9a2dbb60b288
Commit Graph

2825 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Aditya Atluri 98c4221dc2 changed copyright year from 2016 to 2017 in include directory 
Change-Id: Ib5935a84fb51a04b3446df31cc2287101f791b83
 2017-01-11 18:09:33 -06:00
Aditya Atluri 73fcce26f9 changed copyright year from 2016 to 2017 in src directory 
Change-Id: Idb97db509b2b4b1656b2df7a14a62ade38c9d574
 2017-01-11 18:05:41 -06:00
Aditya Atluri 57294ce461 added test for vector data types 
Change-Id: I0b6624886e474601cb1ef003c5f10adf399a21c9
 2017-01-11 18:02:30 -06:00
Aditya Atluri e30887dc69 fixed compilation issues with operator overloading device data types 
Change-Id: I6a60282f0c04a3c0d382cdf2d67ad8d9156880ad
 2017-01-11 17:53:32 -06:00
Aditya Atluri 39910029a6 Added proper device data types 
Change-Id: I42029635ff68c3c13a764a3eda6447e6c77878c6
 2017-01-11 15:06:25 -06:00
Evgeny Mankov 9a0780001b [HIPIFY] cudaDataType_t and libraryPropertyType_t support (CUDA 8.0.44 only) 
All are marked as HIP_UNSUPPORTED.
IMPORTANT:
1. libraryPropertyType_t has no cuda prefix. => TO_DO: new matcher is needed.
2. all libraries (cublas, cufft, cusolver, cusparse, nvgraph) have started to use these types (since 8.0).
 2017-01-10 20:24:27 +03:00
Evgeny Mankov 3a99536ed5 [HIPIFY] cudaDeviceAttr (RT API) support up to CUDA 8.0.44 
Attributes, which are not yet supported by HIP, are marked as HIP_UNSUPPORTED.
 2017-01-10 19:29:33 +03:00
Evgeny Mankov 7ed2b163de [HIPIFY] CUdevice_attribute support up to CUDA 8.0.44 
Attributes, which are not yet supported by HIP, are marked as HIP_UNSUPPORTED.
 2017-01-10 17:54:22 +03:00
Ben Sander a15d236de3 Fix delete[] 2017-01-09 21:03:11 -06:00
Ben Sander a3e0012567 Add HIP_MAX_QUEUES feature. 
Includes some tricky manipulation of the locks for contexts and streams.
issue is that stealing a stream requires we lock the context to
walk the streams to find a victim.  To avoid deadlock, we can't
have a stream locked when we lock the context.  This implementation
releases the stream lock, then acquires the context and selects the
victim.
A more stable implemenation might be to copy the stream list
from a context so that a lock is not required to walk all streams.
Smart shared_ptr could be used to prevent the streams from being
deallocated during the walk.
 2017-01-09 21:02:56 -06:00
Ben Sander 93fbc9cf7b First pass at virtualized queue support. 
Also updated stream debug messages to consistently use trace_helper.
 2017-01-09 21:02:53 -06:00
Ben Sander fd209f37d9 Add more notes on debugging HIP apps. 2017-01-09 21:02:50 -06:00
Ben Sander 3a42a7642a tolerate spaces in hip args 2017-01-09 20:57:13 -06:00
Rahul Garg 5fb09879c7 Added state for hipDevice. 
Change-Id: Idbc3c04cd054a01b634856a1e0a23ff172e991aa
 2017-01-09 23:54:01 +05:30
Maneesh Gupta 9199f952f9 Merge branch 'amd-develop' into amd-master 
Change-Id: I81429e5f3f55a71498da6cece9d08a8b1c170057
 2017-01-06 12:55:36 +05:30
scchan 4fd48084a6 [cmake] add library dependencies to hip_hcc libraries 2017-01-05 18:26:54 -05:00
Maneesh Gupta a42da10c44 hipcc: Link to shared HIP runtime by default 
Change-Id: I5030e3245e4afb6863b401656ca5d1ad9ae84310
 2017-01-04 12:39:09 +05:30
Evgeny Mankov 14e9cf7e62 [HIPIFY] Elapsed time is added to statistics. 2016-12-28 20:44:05 +03:00
Evgeny Mankov 6ceb85a03a [HIPIFY] Added the rest of cuBlas API. 
CUBLAS API 7.5 now is supported by hipify;
API calls, which are not yet supported by hcblas/hipblas, are listed as HIP_UNSUPPORTED.
 2016-12-28 18:08:10 +03:00
Evgeny Mankov d7d3fcc77d [HIPIFY] Formatting, no functional changes. 2016-12-27 19:48:59 +03:00
Evgeny Mankov 5ec0488ce8 [HIPIFY] [Fix] An argument of a function used as macro argument is not hipified. 
https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/issues/35
 2016-12-27 18:54:02 +03:00
Evgeny Mankov 24703944de [HIPIFY] Pointer to typedef declaration is not hipified 
Fix for https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/issues/60
 2016-12-26 19:03:50 +03:00
Evgeny Mankov bcbbc32fa6 [HIPIFY] Add hipconvertinplace2.sh and hipexamine2.sh scripts for hipify-clang. 
The differences from the similar scripts for hipify.pl:
1. CSV file with extended statistics is produced.
2. scripts' arguments are changed a bit:
DIRNAME [hipify options] [--] [clang options]

where -- is a delimiter; all the arguments are optional, except DIRNAME.

Usage example:
./hipexamine2.sh ./tmp -o-stats ./tmp/stats.csv -- -I/usr/local/cuda-7.5/include -I/usr/local/hipify-clang/hipblas/include 2>&1 | tee log
 2016-12-23 22:06:20 +03:00
Evgeny Mankov ab00e2a627 [HIPIFY] Fix line endings. 2016-12-23 18:01:26 +03:00
Evgeny Mankov 6882057fd2 [HIPIFY] Stats: Calculation of changed code amount, based on actually replaced bytes. 
+ REPLACED bytes, TOTAL bytes & CODE CHANGED are added to statistics.
+ -o-stats option for specifying the file with statistic.
 2016-12-23 17:40:06 +03:00
Maneesh Gupta f6e9f6f0bf hipcc: link to hip runtime using absolute path 
Change-Id: I714b3e9da0bc1d49665b079d9c4cec1c1a2efa80
 2016-12-23 11:49:00 +05:30
Ben Sander c325c988b1 Support size_t in memset kernel. 
Add disable for HSA_AMD_AGENT_INFO_MAX_WAVES_PER_CU
Remove one copy of completion_future in memset.
 2016-12-22 12:25:09 -06:00
Maneesh Gupta 9b6d1588ba hip_hcc package changes 
- updated hip_hcc package creation dependencies
- support build hip_hcc package for HCC-1.0

Change-Id: Idf23e415eff8cb352a8906191c79bd822c7618e7
 2016-12-22 15:30:38 +05:30
Ben Sander 37d8cafb12 Increment API sequence number. 
Change name to tls_tidInfo
 2016-12-21 15:30:36 -06:00
Evgeny Mankov 4bb8bf8dab [HIPIFY] Statistics in CSV file. 
+ Stats by CUDA ref name.
+ Conversion %.

TODO: Calculation of changed code amount, based on actually replaced bytes.
 2016-12-21 23:08:01 +03:00
Rahul Garg 4704547bab Removed redundant GetPCIBusID int version function 
Change-Id: I37f2ff87d09fcfb1e3b104c44c51f606fcb83c01
 2016-12-20 23:25:16 +05:30
Evgeny Mankov 3dd32e969d [HIPIFY] Reflect unsupported CUDA API refs in statistics 
https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/issues/53

+ Unsupported refs (by HIP) are now might be listed along with the supported ones.
+ Warnings are added for the unhandled (by HIPIFY) refs, for instance:
  "warning: the following reference is not handled: 'cublasContext' [param decl ptr]."
+ Reflect unsupported CUDA API refs in statistics.
+ Occupancy API [HIP_UNSUPPORTED].
+ A few CUBLAS refs are listed as HIP_UNSUPPORTED.

TODO: Statistics in CSV file.
 2016-12-19 14:38:19 +03:00
Maneesh Gupta 8f31ad6ce0 Merge branch 'amd-develop' into amd-master 
Change-Id: I77fa88b460be549bfcf9e18d3212e732ffc045f5
 2016-12-19 16:20:38 +05:30
Rahul Garg fbf7ed63a8 Fix for HCSWAP-67 
Change-Id: I0b2ce5ab933237947fb41d89769db3da16e5be6a

Conflicts:
	src/hip_hcc.cpp
 2016-12-19 16:19:51 +05:30
Maneesh Gupta f052f43b3b Updated doxygen documentation 
Change-Id: If04d1155173fba8d3e050f3259da8b3edc60e076
 2016-12-19 04:04:06 +00:00
Ben Sander 90c69e14bb Add name for function 2016-12-17 08:54:09 -06:00
Ben Sander 8bf4bd2f7d Remove HSA dependency from hipFunction_t 
Place _groupSegmentSize and _privateSegmentSize inside Function,
remove hsa_executable_symbol_t.
 2016-12-17 07:22:56 -06:00
Ben Sander 6ed7e1c1c1 Remove USE_DISPATCH_HSA_KERNEL=0 path. 2016-12-17 07:22:56 -06:00
Ben Sander 4d29885be3 Refactor Module and Function APIs. 
- hipFunction_t is now returned by value.  This eliminates dynamic
      allocation / memory management complexity in the module.  Removed
the kernel
      name so the structure is just 16 bytes now.

    - Moved the hsa_executable_load_module and hsa_executable_freeze
      calls to the hipModuleLoad and hipModuleLoadData calls.

    - Apply sharedMemBytes in hipModuleLaunchKernel to group segment
      size (not private).
 2016-12-17 07:22:33 -06:00
Rahul Garg 263a9614ff Mapped hipDevice_t to int 
Change-Id: I6cfa56c42b7cd04aa0e0bce510c0d72d34ea211a
 2016-12-17 16:53:03 +05:30
Aditya Atluri 2665ad2762 disabled half native support as inline asm is not working 
Change-Id: I3073d8ae39eed321987f0f2f0e689eec4cdbb48c
 2016-12-16 09:24:59 -06:00
Ben Sander 43635f51dc Print limits on CUDA devices 2016-12-16 08:55:11 -06:00
Ben Sander bd19bb4074 Fix typo 2016-12-15 14:42:52 -06:00
Ben Sander 8ed38bae69 fix copyright 2016-12-15 14:42:52 -06:00
Ben Sander 4080fe209d remove TODO file 2016-12-15 14:42:52 -06:00
Evgeny Mankov 2383d9bc1a [HIPIFY] nested macro is not hipified, when it isAnyIdentifier 
Fix for https://github.com/GPUOpen-ProfessionalCompute-Tools/HIP/issues/55
 2016-12-15 21:00:34 +03:00
Maneesh Gupta 617737e0df Merge branch 'amd-develop' into amd-master 
Change-Id: I52830df409da1f021c32ea569d4ae671aeb57b03
 2016-12-15 16:25:33 +05:30
Brecht Carlier 05b3e2928d Update hip_faq.md 
Fixed navigation and list.
 2016-12-15 14:08:34 +05:30
Martin Schleiss b1eace4348 Fix another typo 2016-12-15 14:07:20 +05:30
Martin Schleiss 263dcfed83 Fix various typos 
Conflicts:
	README.md

Change-Id: Ie296d503d16121a62fed1a208352ec2b81c97fd9
 2016-12-15 14:06:38 +05:30