1e6b79ea64
Fix broken CUDA path in hipcc
2016-04-07 10:48:16 +05:30
3304092b49
Corrected corner cases
2016-04-06 14:50:27 -05:00
1dcdfc93f1
Added feature for cross-compilation and seperated hcc api from hip api
2016-04-06 14:18:51 -05:00
8e3465f6ae
Add c++abi to linker options for HCC
2016-04-02 22:32:05 +05:30
bad947ae8b
Initial (non-default) support for shared library HIP
2016-04-01 16:08:49 +05:30
9de4c04a0a
Revert to multiple output files than single .a lib
2016-03-30 07:54:18 -05:00
2348091b62
include codexl marker path, if found
2016-03-29 17:26:44 -05:00
373251aca0
Improve automated platform detection.
...
If AMD GPU is installed and detected by driver, default
HIP_PLATFORM to hcc.
2016-03-25 17:08:34 -05:00
814c3d30fc
report linux distro if possible
2016-03-24 22:12:41 -05:00
f3ad047194
Provcide more flexiblity in using HIP_PATH to control installation.
...
Also improve docs for developers / contributing.
2016-03-24 13:33:07 -05:00
eb33395045
Add --hipcc_explicit_lib option
...
Tells hipcc not to add /opt/hip or -lhip_hcc to command-line
(since it has already been done).
2016-03-24 11:53:33 -05:00
7920fd9a47
moved variables to headers
2016-03-23 12:59:52 -05:00
89bfc0e375
WIP added modular feature for device apis
2016-03-23 07:54:01 -05:00
3a5f964c4f
Only include activity logger if CodeXL installed.
...
Fix hipHostMalloc in hipBusBandwidth.
2016-03-22 09:27:10 -05:00
7715fe424c
remove unneeded files
2016-03-23 03:41:01 -05:00
b6b3fea05b
use codexl marker interface to mark HIP function/begin end.
...
- Creates markers in HIP group and they show up in CodeXL trace
- Marker text includes HIP functioin arguments
- (Add trace_helper to convert arguments to strings)
- Still need to add HIP_INIT_API for ~30 HIP functions.
2016-03-23 01:17:53 -05:00
ab910efb96
hipHostRegister and hipHostMalloc refactor.
...
Note hipHostMalloc (not hipHostAlloc or hipMallocHost).
- the hipHost* is used for all HIP APIs dealing with Host memory.
(including hipHostMalloc, hipHostFree, hipHostRegister,
hipHostUnregister, hipHostGetFlags, hipHostGetDevicePointer).
- hipMallocHost is consistent with "hipMalloc" for allocating device
memory. Enumerations hipHostMalloc* also used as optional
flags parm to hipHostMalloc.
2016-03-22 02:30:10 -05:00
d3ba2b9782
corrected hipDeviceGetProperties to hipGetDeviceProperties - not docs
2016-03-06 08:31:04 -06:00
f479531be5
Added hipHostAlloc feature for CUDA
2016-03-05 13:58:56 -06:00
5013884d3d
Fix hipcc from trying to compile object files
2016-03-03 09:47:37 +05:30
352c056ea6
add check for hip platform configuration
2016-02-28 21:08:53 -06:00
af22d056e0
touchup
2016-02-28 21:08:53 -06:00
ea09557e1b
disable rocrv2, properly
2016-02-27 03:31:30 -06:00
3238185276
Null argument handling HIPCC
...
This change makes hipcc handle no-arguments passed to it. It solves https://github.com/AMDComputeLibraries/HIP-privatestaging/issues/1
2016-02-26 13:47:58 -06:00
a30018e166
Add PATH and LD_LIBRARY_FLAGS
2016-02-25 23:07:18 -06:00
4adab7b7ef
Merge branch 'memtracker' into privatestaging
...
Conflicts:
src/hip_hcc.cpp
2016-02-25 19:38:46 -06:00
b23f9776ca
A support of the following device properties is added to legacy hipify.pl: hipDeviceAttributeConcurrentKernels, hipDeviceAttributeMemoryClockRate & hipDeviceAttributeMemoryBusWidth.
2016-02-19 13:36:37 +03:00
311a728d25
Search multiple dirs.
2016-02-18 21:07:14 -06:00
dffe573d49
Tweak full formatting
2016-02-17 21:22:31 -06:00
59379ffb44
more work on async copies
2016-02-17 00:59:12 -06:00
4ee2a5229b
Create address tracker for am_alloc.
...
Tracks device where memory is allocated, pinned-host or device, and
more.
Uses memory-range-based lookups - so pointers that exist anywhere in
the range of hostPtr + size will find the associated AmPointerInfo.
The insertions and lookups use a self-balancing binary tree and
should support O(logN) lookup speed.
2016-02-12 18:24:08 -06:00
ea8f99702d
Fix typo: maxThreadsPerMultiProcessor -> MaxSharedMemoryPerMultiprocessor
...
Device property MaxSharedMemoryPerMultiprocessor set equal to totalGlobalMem (HIP path).
Reason: MaxSharedMemoryPerMultiprocessor should be as the same as group memory size. Group memory will not be paged out, so, the physical memory size = total shared memory size = group region size. NVCC path remains untouched: CUDA's device property MaxSharedMemoryPerMultiprocessor is reported.
hipify is updated as well.
2016-02-12 01:29:20 +03:00
7cc7f88717
which_hip -> hipconfig
2016-02-09 11:51:26 +05:30
2a33197a06
Add hcc-config info to --full
2016-02-08 21:44:55 -06:00
7cb5412ff4
iScript cleanup, add --full
2016-02-08 21:44:55 -06:00
0a27507208
Implementation of hipDeviceGetAttribute()
2016-02-04 17:39:27 +03:00
1327d3e03d
Merge branch 'master' into privatestaging
2016-02-03 09:39:19 -06:00
3898f6af3c
Remove warning on ballot/any/all and pop/clz.
...
Since these are supported in HIP no reason to emit warnings.
2016-02-02 10:02:48 -06:00
1e7642c569
Suppress linker warnings in case HCC distribution contains OpenCL/SPIR symbols
2016-02-02 16:27:42 +08:00
1e2fb8a1ab
Fix typo in hipStreamWaitEvent. Fixes#9
2016-01-28 09:51:11 -06:00
f38e63ff18
Initial commit for GPUOpen Launch
2016-01-26 20:14:33 -06:00
Maneesh Gupta
Aditya Atluri
Ben Sander
Evgeny Mankov
Sam Kolton
Jack Chung