2
0

[SWDEV-555807] TestCudaMallocAsync test power draw failing (#755)

* Clarified comments regarding power limit retrieval and its support on virtualized systems.
* Change unsupported comment to UINT32_MAX

---------

Signed-off-by: josnarlo <Joseph.Narlo@amd.com>
Signed-off-by: Arif, Maisam <Maisam.Arif@amd.com>

[ROCm/amdsmi commit: 460cfcba1f]
Este cometimento está contido em:
Narlo, Joseph
2025-10-17 08:57:57 -05:00
cometido por GitHub
ascendente baa8fa3042
cometimento 5ec7b213e4
2 ficheiros modificados com 9 adições e 3 eliminações
+4 -1
Ver ficheiro
@@ -1085,6 +1085,8 @@ typedef struct {
/**
* @brief Power Information
*
* Unsupported struct members are set to UINT32_MAX
*
* @cond @tag{gpu_bm_linux} @endcond
*/
typedef struct {
@@ -6259,9 +6261,10 @@ amdsmi_get_gpu_activity(amdsmi_processor_handle processor_handle, amdsmi_engine_
*
* @ingroup tagGPUMonitor
*
* @platform{gpu_bm_linux} @platform{host} @platform{guest_windows}
* @platform{gpu_bm_linux} @platform{host} @platform{guest_windows} @platform{guest_1vf}
*
* @note amdsmi_power_info_t::socket_power metric can rarely spike above the socket power limit in some cases
* @note unsupported struct members are set to UINT32_MAX
*
* @param[in] processor_handle PF of a processor for which to query
*
+5 -2
Ver ficheiro
@@ -4468,11 +4468,14 @@ amdsmi_get_power_info(amdsmi_processor_handle processor_handle, amdsmi_power_inf
}
int power_limit = 0;
status = smi_amdgpu_get_power_cap(gpu_device, &power_limit);
if (status == AMDSMI_STATUS_SUCCESS) {
amdsmi_status_t status2 = smi_amdgpu_get_power_cap(gpu_device, &power_limit);
if (status2 == AMDSMI_STATUS_SUCCESS) {
info->power_limit = power_limit;
}
// Returning status from amdsmi_get_gpu_metrics_info() which should return SUCCESS
// Getting power cap values may not be supported on all virtualized systems and should
// not return a failure when the metrics values are ascertainable.
return status;
}