Adds rocprofiler_load_counter_definition. This function allows a counter definition file to be supplied to rocprofiler-sdk directly. Takes in a string containing the counter definition YAML, its size (in bytes), and a flag value to state whether this is an append operation or not.
---------
Co-authored-by: Benjamin Welton <ben@amd.com>
Co-authored-by: Jonathan R. Madsen <jrmadsen@users.noreply.github.com>
Co-authored-by: usrihari123 <srihari.u@amd.com>
[ROCm/rocprofiler-sdk commit: 7ddc72ad45]
Adds rocprofiler_load_counter_definition. This function allows a counter definition file to be supplied to rocprofiler-sdk directly. Takes in a string containing the counter definition YAML, its size (in bytes), and a flag value to state whether this is an append operation or not.
---------
Co-authored-by: Benjamin Welton <ben@amd.com>
Co-authored-by: Jonathan R. Madsen <jrmadsen@users.noreply.github.com>
Co-authored-by: usrihari123 <srihari.u@amd.com>
SWDEV-475242
For the description of "FP32 Engine Activity" and "FP64 Engine Activity" in dcgm,
It seems that we do not have an equivalent to these pipe-utilizations on our hardware.
In rocprofiler, I think VALU Utilization is the closest to what we want.
Change-Id: Ibce8835ef4757084cdfd73258de6fc1606ca0158
Signed-off-by: Chen Gong <curry.gong@amd.com>
SWDEV-475242
For the description of "FP32 Engine Activity" and "FP64 Engine Activity" in dcgm,
It seems that we do not have an equivalent to these pipe-utilizations on our hardware.
In rocprofiler, I think VALU Utilization is the closest to what we want.
Change-Id: Ibce8835ef4757084cdfd73258de6fc1606ca0158
Signed-off-by: Chen Gong <curry.gong@amd.com>
[ROCm/rdc commit: 251fcbe49d]
This reverts commit 84fb57e7f9.
Reason for revert: Even though this change is valid, this would break backward compatibility.
Change-Id: I9c7cab83198c8d5c8485b11194099162e3e7a874
[ROCm/clr commit: ec6f83b544]
This reverts commit efce2f77c4.
Reason for revert: Even though this change is valid, this would break backward compatibility.
Change-Id: I9c7cab83198c8d5c8485b11194099162e3e7a874
* Rebased optizations for rocprofv3 tool
* Fixing merge conflicts
* Formatting
* Open from within mutex
* Small name changes
* Added operator
[ROCm/rocprofiler-sdk commit: 6ae441f785]
* Host trap PC sampling uses new record type
* removing redundant field
* formatting
* simplifying templates in the parser - no need for HostTrap boolean
* reviving some parser tests
* hw_id decoding on GFX9
* HW id parser test
* parser CID test
* Parser multigpu test
* removing rocprofiler_pc_sampling_record_t and some fields from hw_id
* simplifying parser context
* keep bench test internally
* initializing gfx9_hw_id_t differently
* anonymous struct first
* avoiding inlining initialization of struct
[ROCm/rocprofiler-sdk commit: bc52c17e64]
* Host trap PC sampling uses new record type
* removing redundant field
* formatting
* simplifying templates in the parser - no need for HostTrap boolean
* reviving some parser tests
* hw_id decoding on GFX9
* HW id parser test
* parser CID test
* Parser multigpu test
* removing rocprofiler_pc_sampling_record_t and some fields from hw_id
* simplifying parser context
* keep bench test internally
* initializing gfx9_hw_id_t differently
* anonymous struct first
* avoiding inlining initialization of struct
GFX 9.0.8 may not properly support pipe reset capabilities so disable
test for now.
Change-Id: I3061cdad87eb979ba884c194f4229c0cbb144ee2
[ROCm/ROCR-Runtime commit: 0f02ed6ffb]
Currently amd::Monitor can work in FILO mode for the active waits
and cause a delay in wakeup of some threads. That may have a problem
with the current sysmem pool design.
Change-Id: I145081478d1e0b282d8838855c5718f09cf54b69
[ROCm/clr commit: 9473f143c2]
Currently amd::Monitor can work in FILO mode for the active waits
and cause a delay in wakeup of some threads. That may have a problem
with the current sysmem pool design.
Change-Id: I145081478d1e0b282d8838855c5718f09cf54b69
Support zero width and height for hipTextObjectCreate to align
with cuda.
Change-Id: I5d4c48625faf5f060ed2a7e634ec65e4ecac9da5
[ROCm/clr commit: efce2f77c4]
Ioannis Assiouras
Benjamin Welton
Gopesh Bhardwaj
Saleel Kudchadker
Vladimir Indic
Chen Gong
Maneesh Gupta
Giovanni Lenzi Baraldi
Jonathan Kim
German Andryeyev
Rahul Manocha