5.10. Fabric definitions
Classes
Defines
- #define NVML_GPU_FABRIC_HEALTH_GET ( var, type )
- #define NVML_GPU_FABRIC_HEALTH_MASK_ACCESS_TIMEOUT_RECOVERY_FALSE 2
- Fabric Health Mask: Access Timeout Recovery not in progress.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ACCESS_TIMEOUT_RECOVERY_NOT_SUPPORTED 0
- Fabric Health Mask: Access Timeout Recovery not supported.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ACCESS_TIMEOUT_RECOVERY_TRUE 1
- Fabric Health Mask: Access Timeout Recovery in progress.
- #define NVML_GPU_FABRIC_HEALTH_MASK_DEGRADED_BW_FALSE 2
- Fabric Health Mask: Bandwidth not degraded.
- #define NVML_GPU_FABRIC_HEALTH_MASK_DEGRADED_BW_NOT_SUPPORTED 0
- Fabric Health Mask: Degraded Bandwidth not supported.
- #define NVML_GPU_FABRIC_HEALTH_MASK_DEGRADED_BW_TRUE 1
- Fabric Health Mask: Bandwidth degraded.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INCOMPATIBLE_GPU_FW 6
- Fabric Health Mask: Incorrect Configuration - Incompatible GPU Firmware.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INCORRECT_CHASSIS_SN 3
- Fabric Health Mask: Incorrect Configuration - Chassis Serial Number.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INCORRECT_SYSGUID 2
- Fabric Health Mask: Incorrect Configuration - SysGUID.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INSUFFICIENT_NVLINKS 5
- Fabric Health Mask: Incorrect Configuration - Insufficient Nvlinks.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INVALID_LOCATION 7
- Fabric Health Mask: Incorrect Configuration - Invalid Location.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_NONE 1
- Fabric Health Mask: Correct Configuration.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_NOT_SUPPORTED 0
- Fabric Health Mask: Incorrect Configuration not supported.
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_NO_PARTITION 4
- Fabric Health Mask: Incorrect Configuration - No Partition.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_RECOVERY_FALSE 2
- Fabric Health Mask: Route Recovery not in progress.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_RECOVERY_NOT_SUPPORTED 0
- Fabric Health Mask: Route Recovery not supported.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_RECOVERY_TRUE 1
- Fabric Health Mask: Route Recovery in progress.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_UNHEALTHY_FALSE 2
- Fabric Health Mask: Route is healthy.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_UNHEALTHY_NOT_SUPPORTED 0
- Fabric Health Mask: Route Unhealthy not supported.
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_UNHEALTHY_TRUE 1
- Fabric Health Mask: Route is unhealthy.
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_ACCESS_TIMEOUT_RECOVERY 6
- Fabric Health Mask Bit Shift for Access Timeout Recovery.
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_DEGRADED_BW 0
- Fabric Health Mask Bit Shift for Degraded Bandwidth.
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_INCORRECT_CONFIGURATION 8
- Fabric Health Mask Bit Shift for Incorrect Configuration.
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_ROUTE_RECOVERY 2
- Fabric Health Mask Bit Shift for Route Recovery.
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_ROUTE_UNHEALTHY 4
- Fabric Health Mask Bit Shift for Route Unhealthy.
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_ACCESS_TIMEOUT_RECOVERY 0x3
- Fabric Health Mask Width for Access Timeout Recovery.
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_DEGRADED_BW 0x3
- Fabric Health Mask Width for Degraded Bandwidth.
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_INCORRECT_CONFIGURATION 0xf
- Fabric Health Mask Width for Incorrect Configuration.
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_ROUTE_RECOVERY 0x3
- Fabric Health Mask Width for Route Recovery.
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_ROUTE_UNHEALTHY 0x3
- Fabric Health Mask Width for Route Unhealthy.
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_HEALTHY 1
- Fabric Health Summary: Healthy.
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_LIMITED_CAPACITY 3
- Fabric Health Summary: Limited Capacity.
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_NOT_SUPPORTED 0
- Fabric Health Summary: Not supported.
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_UNHEALTHY 2
- Fabric Health Summary: Unhealthy.
- #define NVML_GPU_FABRIC_HEALTH_TEST ( var, type, val )
- #define NVML_GPU_FABRIC_STATE_COMPLETED 3
- Fabric Probe State completed.
- #define NVML_GPU_FABRIC_STATE_IN_PROGRESS 2
- Fabric Probe in progress.
- #define NVML_GPU_FABRIC_STATE_NOT_STARTED 1
- Fabric Probe has not started.
- #define NVML_GPU_FABRIC_STATE_NOT_SUPPORTED 0
- Fabric Probe State not supported.
- #define NVML_GPU_FABRIC_UUID_LEN 16
- Length of Fabric UUID.
- #define nvmlGpuFabricInfo_v2
- #define nvmlGpuFabricInfo_v3
Typedefs
- typedef unsigned char nvmlGpuFabricState_t
Defines
- #define NVML_GPU_FABRIC_HEALTH_GET ( var, type )
-
GPU Fabric Health Status Mask for various fields can be obtained using the below macro. Ex - NVML_GPU_FABRIC_HEALTH_GET( var, _DEGRADED_BW)
Value
(((var) >> NVML_GPU_FABRIC_HEALTH_MASK_SHIFT##type) & \ (NVML_GPU_FABRIC_HEALTH_MASK_WIDTH##type))
- #define NVML_GPU_FABRIC_HEALTH_MASK_ACCESS_TIMEOUT_RECOVERY_FALSE 2
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_ACCESS_TIMEOUT_RECOVERY_NOT_SUPPORTED 0
-
Fabric Access Timeout Recovery
- #define NVML_GPU_FABRIC_HEALTH_MASK_ACCESS_TIMEOUT_RECOVERY_TRUE 1
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_DEGRADED_BW_FALSE 2
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_DEGRADED_BW_NOT_SUPPORTED 0
-
Fabric Degraded BW
- #define NVML_GPU_FABRIC_HEALTH_MASK_DEGRADED_BW_TRUE 1
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INCOMPATIBLE_GPU_FW 6
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INCORRECT_CHASSIS_SN 3
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INCORRECT_SYSGUID 2
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INSUFFICIENT_NVLINKS 5
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_INVALID_LOCATION 7
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_NONE 1
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_NOT_SUPPORTED 0
-
Fabric Incorrect Configuration
- #define NVML_GPU_FABRIC_HEALTH_MASK_INCORRECT_CONFIGURATION_NO_PARTITION 4
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_RECOVERY_FALSE 2
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_RECOVERY_NOT_SUPPORTED 0
-
Fabric Route Recovery
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_RECOVERY_TRUE 1
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_UNHEALTHY_FALSE 2
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_UNHEALTHY_NOT_SUPPORTED 0
-
Nvlink Fabric Route Unhealthy
- #define NVML_GPU_FABRIC_HEALTH_MASK_ROUTE_UNHEALTHY_TRUE 1
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_ACCESS_TIMEOUT_RECOVERY 6
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_DEGRADED_BW 0
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_INCORRECT_CONFIGURATION 8
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_ROUTE_RECOVERY 2
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_SHIFT_ROUTE_UNHEALTHY 4
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_ACCESS_TIMEOUT_RECOVERY 0x3
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_DEGRADED_BW 0x3
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_INCORRECT_CONFIGURATION 0xf
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_ROUTE_RECOVERY 0x3
-
- #define NVML_GPU_FABRIC_HEALTH_MASK_WIDTH_ROUTE_UNHEALTHY 0x3
-
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_HEALTHY 1
-
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_LIMITED_CAPACITY 3
-
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_NOT_SUPPORTED 0
-
Fabric Health
- #define NVML_GPU_FABRIC_HEALTH_SUMMARY_UNHEALTHY 2
-
- #define NVML_GPU_FABRIC_HEALTH_TEST ( var, type, val )
-
GPU Fabric Health Status Mask for various fields can be tested using the below macro. Ex - NVML_GPU_FABRIC_HEALTH_TEST( var, _DEGRADED_BW, _TRUE)
Value
(NVML_GPU_FABRIC_HEALTH_GET(var, type) == \ NVML_GPU_FABRIC_HEALTH_MASK##type##val)
- #define NVML_GPU_FABRIC_STATE_COMPLETED 3
-
- #define NVML_GPU_FABRIC_STATE_IN_PROGRESS 2
-
- #define NVML_GPU_FABRIC_STATE_NOT_STARTED 1
-
- #define NVML_GPU_FABRIC_STATE_NOT_SUPPORTED 0
-
Fabric Probe States
- #define NVML_GPU_FABRIC_UUID_LEN 16
-
- #define nvmlGpuFabricInfo_v2
-
Version identifier value for nvmlGpuFabricInfo_v2_t::version.
Value
NVML_STRUCT_VERSION(GpuFabricInfo, 2)
- #define nvmlGpuFabricInfo_v3
-
Version identifier value for nvmlGpuFabricInfo_v3_t::version.
Value
NVML_STRUCT_VERSION(GpuFabricInfo, 3)