MIOpen

MIOpen copied to clipboard

[conv] miopenFindConvolution: Officially allow smaller workspace than GWSS returns.

2

This essentially re-implements "_Consider workspace constraints when loading solutions from DB (#2888)_" but in a bit stricter manner and with (hopefully) almost all necessary bells and whistles. ### Functional differences...

atamazov

Experiment with XNACK+ on MI210

13

junliume

Gtest nchw nhwc

2

randyspauldingamd

[softmax] Store full info about unpacked tensors in the problem descriptor

1

_Originated from: https://github.com/ROCm/MIOpen/pull/2782#discussion_r1515323515_ Currently we do not store full information about unpacked tensors, only the flag. Also we do not store information about tensor layouts, at all. The above means...

atamazov

Implement MIOPEN_BACKEND_EXECUTION_PLAN_DESCRIPTOR - [x] #2922 - [x] #2925 - [ ] MIOPEN_BACKEND_EXECUTION_PLAN_DESCRIPTOR Serialization

scerzh

Implement MIOPEN_BACKEND_ENGINECFG_DESCRIPTOR - [x] #2921 - [ ] #2924

scerzh

The following descriptors are to be implemented. MIOPEN_BACKEND_ENGINE_DESCRIPTOR - [x] #2920 - [ ] #2923

scerzh

[XNACK] GPU is asleep during the copy and not waking back up when it should.

7

The cause appears to be that the GPU is asleep during the copy and not waking back up when it should. Changing the grub options allowed these tests to pass...

junliume

Basic mechanism for tuning in ExecutePrimitive

1

This allows to add tuning support for more primitives. As originally batchnorm was requested, I have also included a perfdb definition for that one. For such primitives tuning behaviour can...

DrizztDoUrden

Unable to builld MIOpen from source on SLES

1

When trying to build MIOpen from source, the compilation fails with the following error: ```shell MIOpen/src/include/miopen/handle.hpp:242:17: error: no member named 'has_value' in 'boost::optional' 242 | if(algo.has_value()) ``` I have confirmed...

Zakhrov