AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
5,703 Commits 3 Branches 0 Tags
79807462cb2fe383de44342cf7c73efe2f2c7621
Commit Graph

40 Commits

Author SHA1 Message Date
Sourabh Betigeri 19f45f83f3 SWDEV-251091: Added null checks and empty file check in hip_module and hip_platform 
Change-Id: Iad99f996dcf90ffb86f62e79bf4cbd53b35b7e3b


[ROCm/clr commit: b65314d168]
 2020-10-20 19:55:01 -04:00
kjayapra-amd 78fd06679b SWDEV-254363 - Checks to make sure kernel is launched on the same device as it was launched. 
Change-Id: I2f273a70b1a5fc0e9fc9c6144eabd32466ff4e59


[ROCm/clr commit: 40322a9f1c]
 2020-10-06 08:38:09 -04:00
Vladislav Sytchenko 7b635ae18c Revert "Revert "Merge branch 'amd-master-next' into amd-npi-next"" 
This reverts commit 0627e8783b.

Reason for revert: <INSERT REASONING HERE>

Change-Id: I92ceb171e31026ed1864704cef2fc1497b883ef9


[ROCm/clr commit: d5aa0f607d]
 2020-10-05 13:20:58 -04:00
Vladislav Sytchenko 0627e8783b Revert "Merge branch 'amd-master-next' into amd-npi-next" 
This reverts commit 6c6299acb2.

Reason for revert: <INSERT REASONING HERE>

Change-Id: I53322718dadde2c98f96140b8e260ec7ee9ef721


[ROCm/clr commit: 9142674ef0]
 2020-10-05 13:02:39 -04:00
kjayapra-amd c15524b728 SWDEV-241896 - Remove extern support in ROCclr/HIP since compiler added Texture reference support. 
Change-Id: I1b0a7377b803b4e2b94ffef6ffd91e376fdb2b2f


[ROCm/clr commit: 19acb0bfe5]
 2020-09-24 14:28:16 -04:00
Jatin Chaudhary c04f949c1d Adding checks for return type compatibility with CUDA 
Change-Id: I09b6bd9a7299f0a97ac9011ea0903a1f81e28251


[ROCm/clr commit: d752b6d14b]
 2020-09-21 00:36:02 -04:00
Vladislav Sytchenko 9734ad0e68 Fix Windows build 
Change-Id: I8769bafba199b9258c610e51b58e454fe10a75e6


[ROCm/clr commit: 1602db4114]
 2020-09-18 18:51:25 -04:00
kjayapra-amd 1c03aaaa95 SWDEV-241902 - Changes to pass file descriptor and offset to load code object. 
Change-Id: I96add27f4fde1b5ee5501e206b7e85ab77e87dfc


[ROCm/clr commit: 095ce0cc02]
 2020-09-15 07:54:28 -04:00
Rahul Garg d8db826498 Reinsert __gnu_h2f_ieee and __gnu_f2h_ieee. 
Fixes SWDEV-251676

Reverts Ide24d245d851e20961020323e52f33322a33fff9 and adds
__attribute__((weak)) to __gnu_h2f_ieee and __gnu_f2h_ieee.

Change-Id: If4f9e2f867b56a3f0121a907a49ae4032220d9af


[ROCm/clr commit: f4282cbdf9]
 2020-09-15 01:39:14 -04:00
Tao Sang d21b94e2b0 Replace private libelf with elfio 
Change-Id: I59ae33f7cc55e73f6519af14be91dd9863b03af3


[ROCm/clr commit: 2332174c8a]
 2020-09-01 10:29:11 -04:00
kjayapra-amd 4ea2412701 SWDEV-249499 - Add message to Guarantee assert code in hip runtime. 
Change-Id: I1850463675bbc3e6ff5e3cef8211b1fd40e55f96


[ROCm/clr commit: b9f8ab96ce]
 2020-08-24 17:35:39 -04:00
Jason Tang 91a2f86826 SWDEV-232197 - Remove refering to gfxipVersion_ 
Change-Id: I91b6b972e26b18dabf6deda5a142bff491c8bb45


[ROCm/clr commit: d5c97e0c31]
 2020-08-12 16:34:18 -04:00
Anusha Godavarthy Surya 66be9e86d5 SWDEV-243961 Added overflow check for globalWorkSize 
Change-Id: I153d99398eb4619baad8c0a52466e0d5d5f898ca


[ROCm/clr commit: 8481524416]
 2020-08-05 11:18:07 -04:00
Tao Sang d7c4f9f492 Remove __gnu_h2f_ieee and __gnu_f2h_ieee 
Change-Id: Ide24d245d851e20961020323e52f33322a33fff9


[ROCm/clr commit: e9cc570be8]
 2020-07-23 14:50:24 -04:00
kjayapra-amd db1fbf23d2 SWDEV-236178 - Remove use of old routines functions()/modules(). 
Change-Id: I3a1ed967227c91b0d8cdf39e1360ade685e0bc73


[ROCm/clr commit: fd783c1e88]
 2020-07-22 14:55:02 -04:00
Saleel Kudchadker 2bf550d3c1 Print devPtr and size for hip*Symbol* api 
Change-Id: I07be6949cca005700ee1d1708d86220c793c9728


[ROCm/clr commit: 603546cd45]
 2020-07-21 13:49:10 -04:00
Tao Sang 5225aff583 Implement HIP_ENABLE_DEFERRED_LOADING in right way 
Change-Id: Ia28dbdea5cde93b3ca69f44c6a3c18ec8a730d42


[ROCm/clr commit: 19d6a646bb]
 2020-07-11 18:14:03 -04:00
kjayapra-amd 5cf8c4331b SWDEV-236178 - Fixing func_attributes that was over written on redesign. 
Change-Id: I157913ae4fa4a3a149817e901f10475b7750188b


[ROCm/clr commit: 2f963728f7]
 2020-07-01 11:52:45 -04:00
Tao Sang 3113312e03 Fix static lib crashing issue 
Initialize HIP_ENABLE_DEFERRED_LOADING = 1

Change-Id: I432ccf1926b953c4391fdd086cde7da2380962bb


[ROCm/clr commit: d536a661ac]
 2020-06-22 13:24:40 -04:00
kjayapra-amd 1ed6d978a8 SWDEV-236178 - Store texture reference metadata for dynamically loaded modules. 
Change-Id: I99ecc80da7e29c691341a01a09e4532972f1e3e5


[ROCm/clr commit: 0b788c4c67]
 2020-06-11 22:34:50 -04:00
kjayapra-amd d6ccf1209c SWDEV-236178 - Reorganizing Platform/Modules code for easy access. 
Change-Id: Ie8920260ffc4ff01e44b48af8cec9ea5aed1aa9b


[ROCm/clr commit: 840347f0d0]
 2020-06-11 10:11:20 -04:00
kjayapra-amd edcdf621b9 SWDEV-239327 - Remove amd_mem_obj during unregistervar 
Change-Id: I2130eaa21369b9634a9459680061138c61eaaaa4


[ROCm/clr commit: ee2ff4bc5e]
 2020-06-05 23:24:38 -04:00
kjayapra-amd db1b21b1fa SWDEV-235295 - Move addDeviceProgram() to lazy loading 
Change-Id: I8fe07e370e58844496e18c858bb528393556854f


[ROCm/clr commit: 1dc24194a3]
 2020-06-05 18:03:32 -04:00
kjayapra-amd 70e90419a9 SWDEV-234295 - Pass flag to ROCclr to not clear device programs during program::build() 
Change-Id: I50b9fa1a96da6895f73fdf4a7c0d3f096b1188da


[ROCm/clr commit: 9261a35be9]
 2020-06-05 09:53:11 -04:00
Jatin 216873e9c8 Adding changes for hipExtLaunchKernel for rocCLR 
Change-Id: Iba52bc3bde7c37f3fb375a55ba0947e87b3cdc9b


[ROCm/clr commit: 126573df4c]
 2020-06-02 14:16:41 -04:00
kjayapra-amd 776af02dfe SWDEV-229840 - fixing return HIP_RETURN instances in hip. 
Change-Id: I48763d7268bf5649bf2242c962c185f5f4af159c


[ROCm/clr commit: ab17b43d45]
 2020-05-29 09:43:58 -04:00
kjayapra-amd b91ab37e8d SWDEV-233927 - Crash if binary for current device is not found. 
Change-Id: I57281ae6c09110f39155664fca5a83ea57bb62b4


[ROCm/clr commit: f2899243f3]
 2020-05-28 16:18:27 -04:00
Joseph Greathouse 9b1faa7db0 Fix a build error on signed/unsigned comparison 
Change-Id: Ic79eb4c3ec5c6fd36cea7c4810d990619f08b9e1


[ROCm/clr commit: 766e708535]
 2020-05-28 14:27:16 -05:00
kjayapra-amd 640559bdcb SWDEV-236465 - Return error code as soon as global creation fails. 
Change-Id: I790b8b4fdd6ab8818bc5b6b9a79e6900b840372d


[ROCm/clr commit: 55cdef8e45]
 2020-05-28 13:28:23 -04:00
Joseph Greathouse db6de71b54 Fix occupancy calculation functions in ROCclr path 
The hipOccupancyMaxPotentialBlockSize API is meant to return the
number of threads for the highest-occupancy workgroup, and the number
of those workgroups. It was previously calculating the number of
maximum-sized workgroups that would fit on a single CU. This is
a mixture of the API we wanted (to calculate max potential block size)
and the MaxBlocksPerMultiprocessor function.

This patch fixes it up so that the internal occupancy calculation
function works for two uses: the traditional function that calculates
the maximum blocks per multiprocessor when a user passes in a fixed
block size (used for hipMaxBlocksPerMultiprocessor style functions)
and a function that calculates the size of a block that would lead
to maximum occupancy, and how many blocks of that size would be
needed to fill the whole GPU (for hipOccupancyMaxPotentialBlockSize
style functions).

This also updates the occupancy calculation function to prepare for
gfx10, which does not have SGPR-based occupancy limits.

Change-Id: Ie007b3f9d5ebc4e166b50a3a051498af35650f35


[ROCm/clr commit: 90453b68d3]
 2020-05-28 10:22:10 -05:00
Laurent Morichetti 39491f0ddc load functions when registered if lazy load is disabled 
Change-Id: I08c9684ca4ab86a6294ed9e670f3125c1a0cc6c4


[ROCm/clr commit: 161c14c1de]
 2020-05-21 09:19:20 -04:00
Payam bfde2ce8f8 load functions when registered if lazy load is disabled 
Change-Id: I6f10a72f7643e476407af03b04ef52e59741be45


[ROCm/clr commit: 0a9d550d9b]
 2020-05-15 12:15:32 -04:00
Evgeny a68bf95ab5 API tracing instrumentation 
Change-Id: I257409b9fe299b009ded3e3a43287322d5f93a70


[ROCm/clr commit: 10cb7645dc]
 2020-05-14 11:03:09 -05:00
Payam 0790fed712 adding support to disable lazy loading, HIP_ENABLE_LAZY_KERNEL_LOADING 
Change-Id: Iafe0f396570ee2bccf642b537cbf9aad967e1370


[ROCm/clr commit: 6152b200cb]
 2020-05-13 17:18:48 -04:00
Christophe Paquot dd774dbc5c Make sure to remove the global var from amd::MemObjMap 
We need this otherwise ROCr can give us a matching address
for another allocation and doing "insert" in ROCclr will not
update the map with the newest object. We would then end up
using stale objects (yikes)
SWDEV-234992

Change-Id: I3475adf9781a9309d64a024fae45181d7e5afb04


[ROCm/clr commit: 2cc0fb9381]
 2020-05-13 00:26:19 -07:00
Christophe Paquot 5c506c70c3 Add lock to addFatBinary and removeFatBinary 
In case hipModule(Un)Load is called from different thread as hipInit we need to grab the lock
as both are going to modify modules_
Also add some logging for __hipExtractCodeObjectFromFatBinary in case binary isn't found for GPU
SWDEV-236032

Change-Id: Icbd72b412502df80d5066cea42a4fbcd5b0b8a98


[ROCm/clr commit: 374fd36f31]
 2020-05-12 16:04:08 -07:00
kjayapra-amd f5fd38679b SWDEV-227602 - Fixing hipFuncGetAttribute parameters 
Change-Id: I46bd079372f453cabfaa2c709e2b30c69400dd33


[ROCm/clr commit: aaf3002567]
 2020-05-12 09:15:43 -04:00
Vlad Sytchenko b2afbd99b3 Correct HIP_FUNC_ATTRIBUTE_MAX_THREADS_PER_BLOCK query 
We should be returning the max workgroup size calculated by the compiler.

Change-Id: If86590efbb9b291f470bdbe87e5df992e661c539


[ROCm/clr commit: b5f9d2f818]
 2020-05-08 14:36:47 -04:00
Vlad Sytchenko f3dffba0fe Fix confusion in hipFuncGetAttribute() 
Cuda shared == OpenCL local

Cuda local == OpenCL private

Change-Id: I5a204945ecde35919b9e9def20bbb2662fffea2b


[ROCm/clr commit: 276bfc9667]
 2020-05-08 14:36:36 -04:00
Payam 082b9e51ab name change vdi to rocclr 
Change-Id: I06d198bbb4a499e153b290b73a92afed3553b252


[ROCm/clr commit: dba0e72de2]
 2020-05-06 09:14:30 -04:00