master
csv 1,272 lines 103 KB
Raw
1 # Netdata recommended dcgm-exporter profile.
2 #
3 # Usage:
4 # dcgm-exporter -f /path/to/dcgm-exporter-netdata.csv
5 #
6 # Keep exporter interval aligned with Netdata update_every.
7 # Example:
8 # dcgm-exporter -f /path/to/dcgm-exporter-netdata.csv -c 30000
9 #
10 # This file documents all known DCGM fields (623 in current source dataset).
11 # Exactly 123 fields are enabled by default below (127 max per field group).
12 # To enable a different field, uncomment one line and comment another enabled line.
13 #
14 # IMPORTANT: Fields with DCGM_FT_BINARY type (structs/blobs) MUST remain disabled.
15 # They return non-numeric values that break Prometheus metric parsing.
16 # Known BINARY fields: ACCOUNTING_DATA, COMPUTE_PIDS, CREATABLE_VGPU_TYPE_IDS,
17 # ENC_STATS, FBC_SESSIONS_INFO, FBC_STATS, GRAPHICS_PIDS, SUPPORTED_CLOCKS,
18 # SUPPORTED_TYPE_INFO, VGPU_ENC_SESSIONS_INFO, VGPU_ENC_STATS,
19 # VGPU_FBC_SESSIONS_INFO, VGPU_FBC_STATS, VGPU_INSTANCE_IDS,
20 # VGPU_PER_PROCESS_UTILIZATION, VGPU_UTILIZATIONS, GPU_TOPOLOGY_AFFINITY,
21 # GPU_TOPOLOGY_NVLINK, GPU_TOPOLOGY_PCI, SYNC_BOOST
22 #
23 # Format:
24 # DCGM FIELD, Prometheus metric type, help message
25
26 # context=dcgm.exporter.health.status family=exporter health dimension=bind_unbind_event
27 # DCGM_FI_BIND_UNBIND_EVENT, gauge, Bind unbind event
28 # context=dcgm.exporter.inventory.software family=exporter inventory dimension=cuda_driver_version
29 DCGM_FI_CUDA_DRIVER_VERSION, label, Cuda driver version
30 # context=dcgm.gpu.workload.sessions family=gpu workload dimension=accounting_data
31 # DCGM_FI_DEV_ACCOUNTING_DATA, gauge, Accounting data
32 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=app_mem_clock
33 DCGM_FI_DEV_APP_MEM_CLOCK, gauge, App mem clock
34 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=app_sm_clock
35 DCGM_FI_DEV_APP_SM_CLOCK, gauge, App sm clock
36 # context=dcgm.gpu.state.configuration family=gpu state dimension=autoboost
37 # DCGM_FI_DEV_AUTOBOOST, gauge, Autoboost
38 # context=dcgm.gpu.reliability.memory_health family=gpu reliability dimension=banks_remap_rows_avail_high
39 # DCGM_FI_DEV_BANKS_REMAP_ROWS_AVAIL_HIGH, gauge, Banks remap rows avail high
40 # context=dcgm.gpu.reliability.memory_health family=gpu reliability dimension=banks_remap_rows_avail_low
41 # DCGM_FI_DEV_BANKS_REMAP_ROWS_AVAIL_LOW, gauge, Banks remap rows avail low
42 # context=dcgm.gpu.reliability.memory_health family=gpu reliability dimension=banks_remap_rows_avail_max
43 # DCGM_FI_DEV_BANKS_REMAP_ROWS_AVAIL_MAX, gauge, Banks remap rows avail max
44 # context=dcgm.gpu.reliability.memory_health family=gpu reliability dimension=banks_remap_rows_avail_none
45 # DCGM_FI_DEV_BANKS_REMAP_ROWS_AVAIL_NONE, gauge, Banks remap rows avail none
46 # context=dcgm.gpu.reliability.memory_health family=gpu reliability dimension=banks_remap_rows_avail_partial
47 # DCGM_FI_DEV_BANKS_REMAP_ROWS_AVAIL_PARTIAL, gauge, Banks remap rows avail partial
48 # context=dcgm.gpu.memory.bar1_usage family=gpu memory dimension=free
49 DCGM_FI_DEV_BAR1_FREE, gauge, Bar1 free
50 # context=dcgm.gpu.memory.bar1_capacity family=gpu memory dimension=total
51 DCGM_FI_DEV_BAR1_TOTAL, gauge, Bar1 total
52 # context=dcgm.gpu.memory.bar1_usage family=gpu memory dimension=used
53 DCGM_FI_DEV_BAR1_USED, gauge, Bar1 used
54 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=board_limit_violation
55 # DCGM_FI_DEV_BOARD_LIMIT_VIOLATION, counter, Board limit violation
56 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=brand
57 DCGM_FI_DEV_BRAND, label, Brand
58 # context=dcgm.gpu.interconnect.state family=gpu interconnect/overview dimension=c2c_link
59 # DCGM_FI_DEV_C2C_LINK_COUNT, counter, C2c link count
60 # context=dcgm.gpu.interconnect.error_rate family=gpu interconnect/overview dimension=c2c_link_error_intr
61 # DCGM_FI_DEV_C2C_LINK_ERROR_INTR, counter, C2c link error intr
62 # context=dcgm.gpu.interconnect.error_rate family=gpu interconnect/overview dimension=c2c_link_error_replay
63 # DCGM_FI_DEV_C2C_LINK_ERROR_REPLAY, counter, C2c link error replay
64 # context=dcgm.gpu.interconnect.error_rate family=gpu interconnect/overview dimension=c2c_link_error_replay_b2b
65 # DCGM_FI_DEV_C2C_LINK_ERROR_REPLAY_B2B, counter, C2c link error replay b2b
66 # context=dcgm.gpu.interconnect.state family=gpu interconnect/overview dimension=c2c_link_power_state
67 # DCGM_FI_DEV_C2C_LINK_POWER_STATE, gauge, C2c link power state
68 # context=dcgm.gpu.interconnect.state family=gpu interconnect/overview dimension=c2c_link_status
69 # DCGM_FI_DEV_C2C_LINK_STATUS, gauge, C2c link status
70 # context=dcgm.gpu.interconnect.throughput family=gpu interconnect/overview dimension=c2c_max_bandwidth
71 # DCGM_FI_DEV_C2C_MAX_BANDWIDTH, gauge, C2c max bandwidth
72 # context=dcgm.gpu.capability.support family=gpu capability dimension=cc_mode
73 # DCGM_FI_DEV_CC_MODE, gauge, Cc mode
74 # Legacy name in older stacks: DCGM_FI_DEV_CLOCK_THROTTLE_REASONS.
75 # context=dcgm.gpu.throttle.reasons family=gpu throttle dimension=clocks_event_reasons
76 DCGM_FI_DEV_CLOCKS_EVENT_REASONS, gauge, Clocks event reasons
77 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=hw_power_brake_slowdown
78 DCGM_FI_DEV_CLOCKS_EVENT_REASON_HW_POWER_BRAKE_SLOWDOWN_NS, counter, Clocks event reason hw power brake slowdown ns
79 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=hw_therm_slowdown
80 DCGM_FI_DEV_CLOCKS_EVENT_REASON_HW_THERM_SLOWDOWN_NS, counter, Clocks event reason hw therm slowdown ns
81 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=sw_power_cap
82 DCGM_FI_DEV_CLOCKS_EVENT_REASON_SW_POWER_CAP_NS, counter, Clocks event reason sw power cap ns
83 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=sw_therm_slowdown
84 DCGM_FI_DEV_CLOCKS_EVENT_REASON_SW_THERM_SLOWDOWN_NS, counter, Clocks event reason sw therm slowdown ns
85 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=sync_boost
86 # DCGM_FI_DEV_CLOCKS_EVENT_REASON_SYNC_BOOST_NS, counter, Clocks event reason sync boost ns
87 # context=dcgm.gpu.state.configuration family=gpu state dimension=compute_mode
88 DCGM_FI_DEV_COMPUTE_MODE, gauge, Compute mode
89 # context=dcgm.gpu.interconnect.connectx.link family=gpu interconnect/pcie dimension=connectx_active_pcie_link_speed
90 # DCGM_FI_DEV_CONNECTX_ACTIVE_PCIE_LINK_SPEED, gauge, Connectx active pcie link speed
91 # context=dcgm.gpu.interconnect.pcie.link.width family=gpu interconnect/pcie dimension=connectx_active_pcie_link_width
92 # DCGM_FI_DEV_CONNECTX_ACTIVE_PCIE_LINK_WIDTH, gauge, Connectx active pcie link width
93 # context=dcgm.gpu.interconnect.connectx.error_status family=gpu interconnect/pcie dimension=connectx_correctable_err_mask
94 # DCGM_FI_DEV_CONNECTX_CORRECTABLE_ERR_MASK, gauge, Connectx correctable err mask
95 # context=dcgm.gpu.interconnect.connectx.error_status family=gpu interconnect/pcie dimension=connectx_correctable_err_status
96 # DCGM_FI_DEV_CONNECTX_CORRECTABLE_ERR_STATUS, gauge, Connectx correctable err status
97 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=connectx_device_temperature
98 # DCGM_FI_DEV_CONNECTX_DEVICE_TEMPERATURE, gauge, Connectx device temperature
99 # context=dcgm.gpu.interconnect.connectx.link family=gpu interconnect/pcie dimension=connectx_expect_pcie_link_speed
100 # DCGM_FI_DEV_CONNECTX_EXPECT_PCIE_LINK_SPEED, gauge, Connectx expect pcie link speed
101 # context=dcgm.gpu.interconnect.pcie.link.width family=gpu interconnect/pcie dimension=connectx_expect_pcie_link_width
102 # DCGM_FI_DEV_CONNECTX_EXPECT_PCIE_LINK_WIDTH, gauge, Connectx expect pcie link width
103 # context=dcgm.gpu.interconnect.connectx.status family=gpu interconnect/pcie dimension=connectx_health
104 # DCGM_FI_DEV_CONNECTX_HEALTH, gauge, Connectx health
105 # context=dcgm.gpu.interconnect.connectx.error_status family=gpu interconnect/pcie dimension=connectx_uncorrectable_err_mask
106 # DCGM_FI_DEV_CONNECTX_UNCORRECTABLE_ERR_MASK, gauge, Connectx uncorrectable err mask
107 # context=dcgm.gpu.interconnect.connectx.error_status family=gpu interconnect/pcie dimension=connectx_uncorrectable_err_severity
108 # DCGM_FI_DEV_CONNECTX_UNCORRECTABLE_ERR_SEVERITY, gauge, Connectx uncorrectable err severity
109 # context=dcgm.gpu.interconnect.connectx.error_status family=gpu interconnect/pcie dimension=connectx_uncorrectable_err_status
110 # DCGM_FI_DEV_CONNECTX_UNCORRECTABLE_ERR_STATUS, gauge, Connectx uncorrectable err status
111 # context=dcgm.gpu.reliability.row_remap_events family=gpu reliability dimension=correctable_remapped_rows
112 DCGM_FI_DEV_CORRECTABLE_REMAPPED_ROWS, counter, Correctable remapped rows
113 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=count
114 DCGM_FI_DEV_COUNT, gauge, Count
115 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=cpu_affinity_0
116 # DCGM_FI_DEV_CPU_AFFINITY_0, gauge, Cpu affinity 0
117 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=cpu_affinity_1
118 # DCGM_FI_DEV_CPU_AFFINITY_1, gauge, Cpu affinity 1
119 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=cpu_affinity_2
120 # DCGM_FI_DEV_CPU_AFFINITY_2, gauge, Cpu affinity 2
121 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=cpu_affinity_3
122 # DCGM_FI_DEV_CPU_AFFINITY_3, gauge, Cpu affinity 3
123 # context=dcgm.cpu.clock.frequency family=cpu clock dimension=cpu_clock_current
124 # DCGM_FI_DEV_CPU_CLOCK_CURRENT, gauge, Cpu clock current
125 # context=dcgm.gpu.cpu.info family=gpu cpu dimension=cpu_model
126 # DCGM_FI_DEV_CPU_MODEL, label, Cpu model
127 # context=dcgm.cpu.cpu.power family=cpu cpu dimension=cpu_power_limit
128 # DCGM_FI_DEV_CPU_POWER_LIMIT, gauge, Cpu power limit
129 # context=dcgm.cpu.cpu.power family=cpu cpu dimension=cpu_power_util_current
130 # DCGM_FI_DEV_CPU_POWER_UTIL_CURRENT, gauge, Cpu power util current
131 # context=dcgm.cpu.cpu.temperature family=cpu cpu dimension=cpu_temp_critical
132 # DCGM_FI_DEV_CPU_TEMP_CRITICAL, gauge, Cpu temp critical
133 # context=dcgm.cpu.cpu.temperature family=cpu cpu dimension=cpu_temp_current
134 # DCGM_FI_DEV_CPU_TEMP_CURRENT, gauge, Cpu temp current
135 # context=dcgm.cpu.cpu.temperature family=cpu cpu dimension=cpu_temp_warning
136 # DCGM_FI_DEV_CPU_TEMP_WARNING, gauge, Cpu temp warning
137 # context=dcgm.cpu.cpu.utilization family=cpu cpu dimension=cpu_util_irq
138 # DCGM_FI_DEV_CPU_UTIL_IRQ, gauge, Cpu util irq
139 # context=dcgm.cpu.cpu.utilization family=cpu cpu dimension=cpu_util_nice
140 # DCGM_FI_DEV_CPU_UTIL_NICE, gauge, Cpu util nice
141 # context=dcgm.cpu.cpu.utilization family=cpu cpu dimension=cpu_util_sys
142 # DCGM_FI_DEV_CPU_UTIL_SYS, gauge, Cpu util sys
143 # context=dcgm.cpu.cpu.utilization family=cpu cpu dimension=cpu_util
144 # DCGM_FI_DEV_CPU_UTIL_TOTAL, gauge, Cpu util total
145 # context=dcgm.cpu.cpu.utilization family=cpu cpu dimension=cpu_util_user
146 # DCGM_FI_DEV_CPU_UTIL_USER, gauge, Cpu util user
147 # context=dcgm.gpu.cpu.info family=gpu cpu dimension=cpu_vendor
148 # DCGM_FI_DEV_CPU_VENDOR, label, Cpu vendor
149 # context=dcgm.gpu.virtualization.vgpu.type family=gpu virtualization dimension=creatable_vgpu_type_ids
150 # DCGM_FI_DEV_CREATABLE_VGPU_TYPE_IDS, gauge, Creatable vgpu type ids
151 # context=dcgm.gpu.capability.support family=gpu capability dimension=cuda_compute_capability
152 # DCGM_FI_DEV_CUDA_COMPUTE_CAPABILITY, gauge, Cuda compute capability
153 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=cuda_visible_devices_str
154 # DCGM_FI_DEV_CUDA_VISIBLE_DEVICES_STR, label, Cuda visible devices str
155 # context=dcgm.gpu.compute.utilization family=gpu compute dimension=decoder
156 DCGM_FI_DEV_DEC_UTIL, gauge, Dec util
157 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_cpu_eud_result
158 # DCGM_FI_DEV_DIAG_CPU_EUD_RESULT, gauge, Diag cpu eud result
159 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_diagnostic_result
160 # DCGM_FI_DEV_DIAG_DIAGNOSTIC_RESULT, gauge, Diag diagnostic result
161 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_eud_result
162 # DCGM_FI_DEV_DIAG_EUD_RESULT, gauge, Diag eud result
163 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_memory_bandwidth_result
164 # DCGM_FI_DEV_DIAG_MEMORY_BANDWIDTH_RESULT, gauge, Diag memory bandwidth result
165 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_memory_result
166 # DCGM_FI_DEV_DIAG_MEMORY_RESULT, gauge, Diag memory result
167 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_memtest_result
168 # DCGM_FI_DEV_DIAG_MEMTEST_RESULT, gauge, Diag memtest result
169 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_nccl_tests_result
170 # DCGM_FI_DEV_DIAG_NCCL_TESTS_RESULT, gauge, Diag nccl tests result
171 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_nvbandwidth_result
172 # DCGM_FI_DEV_DIAG_NVBANDWIDTH_RESULT, gauge, Diag nvbandwidth result
173 # context=dcgm.gpu.interconnect.pcie.state family=gpu interconnect/pcie dimension=diag_pcie_result
174 # DCGM_FI_DEV_DIAG_PCIE_RESULT, gauge, Diag pcie result
175 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_pulse_test_result
176 # DCGM_FI_DEV_DIAG_PULSE_TEST_RESULT, gauge, Diag pulse test result
177 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_software_result
178 # DCGM_FI_DEV_DIAG_SOFTWARE_RESULT, gauge, Diag software result
179 # context=dcgm.gpu.diagnostics.status family=gpu diagnostics dimension=diag_status
180 # DCGM_FI_DEV_DIAG_STATUS, gauge, Diag status
181 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_targeted_power_result
182 # DCGM_FI_DEV_DIAG_TARGETED_POWER_RESULT, gauge, Diag targeted power result
183 # context=dcgm.gpu.diagnostics.results family=gpu diagnostics dimension=diag_targeted_stress_result
184 # DCGM_FI_DEV_DIAG_TARGETED_STRESS_RESULT, gauge, Diag targeted stress result
185 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_current
186 DCGM_FI_DEV_ECC_CURRENT, gauge, Ecc current
187 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_cbu
188 # DCGM_FI_DEV_ECC_DBE_AGG_CBU, gauge, Ecc dbe agg cbu
189 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_dev
190 # DCGM_FI_DEV_ECC_DBE_AGG_DEV, gauge, Ecc dbe agg dev
191 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_l1
192 # DCGM_FI_DEV_ECC_DBE_AGG_L1, gauge, Ecc dbe agg l1
193 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_l2
194 DCGM_FI_DEV_ECC_DBE_AGG_L2, gauge, Ecc dbe agg l2
195 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_reg
196 # DCGM_FI_DEV_ECC_DBE_AGG_REG, gauge, Ecc dbe agg reg
197 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_shm
198 # DCGM_FI_DEV_ECC_DBE_AGG_SHM, gauge, Ecc dbe agg shm
199 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_srm
200 # DCGM_FI_DEV_ECC_DBE_AGG_SRM, gauge, Ecc dbe agg srm
201 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_agg_tex
202 # DCGM_FI_DEV_ECC_DBE_AGG_TEX, gauge, Ecc dbe agg tex
203 # context=dcgm.gpu.memory.ecc_error_rate family=gpu memory dimension=ecc_dbe_agg
204 DCGM_FI_DEV_ECC_DBE_AGG_TOTAL, counter, Ecc dbe agg total
205 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_cbu
206 # DCGM_FI_DEV_ECC_DBE_VOL_CBU, gauge, Ecc dbe vol cbu
207 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_dev
208 DCGM_FI_DEV_ECC_DBE_VOL_DEV, gauge, Ecc dbe vol dev
209 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_l1
210 # DCGM_FI_DEV_ECC_DBE_VOL_L1, gauge, Ecc dbe vol l1
211 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_l2
212 DCGM_FI_DEV_ECC_DBE_VOL_L2, gauge, Ecc dbe vol l2
213 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_reg
214 # DCGM_FI_DEV_ECC_DBE_VOL_REG, gauge, Ecc dbe vol reg
215 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_shm
216 # DCGM_FI_DEV_ECC_DBE_VOL_SHM, gauge, Ecc dbe vol shm
217 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_srm
218 # DCGM_FI_DEV_ECC_DBE_VOL_SRM, gauge, Ecc dbe vol srm
219 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_dbe_vol_tex
220 # DCGM_FI_DEV_ECC_DBE_VOL_TEX, gauge, Ecc dbe vol tex
221 # context=dcgm.gpu.memory.ecc_error_rate family=gpu memory dimension=ecc_dbe_vol
222 DCGM_FI_DEV_ECC_DBE_VOL_TOTAL, counter, Ecc dbe vol total
223 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_inforom_ver
224 # DCGM_FI_DEV_ECC_INFOROM_VER, label, Ecc inforom ver
225 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_pending
226 DCGM_FI_DEV_ECC_PENDING, gauge, Ecc pending
227 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_cbu
228 # DCGM_FI_DEV_ECC_SBE_AGG_CBU, gauge, Ecc sbe agg cbu
229 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_dev
230 # DCGM_FI_DEV_ECC_SBE_AGG_DEV, gauge, Ecc sbe agg dev
231 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_l1
232 # DCGM_FI_DEV_ECC_SBE_AGG_L1, gauge, Ecc sbe agg l1
233 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_l2
234 DCGM_FI_DEV_ECC_SBE_AGG_L2, gauge, Ecc sbe agg l2
235 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_reg
236 # DCGM_FI_DEV_ECC_SBE_AGG_REG, gauge, Ecc sbe agg reg
237 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_shm
238 # DCGM_FI_DEV_ECC_SBE_AGG_SHM, gauge, Ecc sbe agg shm
239 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_srm
240 # DCGM_FI_DEV_ECC_SBE_AGG_SRM, gauge, Ecc sbe agg srm
241 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_agg_tex
242 # DCGM_FI_DEV_ECC_SBE_AGG_TEX, gauge, Ecc sbe agg tex
243 # context=dcgm.gpu.memory.ecc_error_rate family=gpu memory dimension=ecc_sbe_agg
244 DCGM_FI_DEV_ECC_SBE_AGG_TOTAL, counter, Ecc sbe agg total
245 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_cbu
246 # DCGM_FI_DEV_ECC_SBE_VOL_CBU, gauge, Ecc sbe vol cbu
247 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_dev
248 DCGM_FI_DEV_ECC_SBE_VOL_DEV, gauge, Ecc sbe vol dev
249 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_l1
250 # DCGM_FI_DEV_ECC_SBE_VOL_L1, gauge, Ecc sbe vol l1
251 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_l2
252 DCGM_FI_DEV_ECC_SBE_VOL_L2, gauge, Ecc sbe vol l2
253 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_reg
254 # DCGM_FI_DEV_ECC_SBE_VOL_REG, gauge, Ecc sbe vol reg
255 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_shm
256 # DCGM_FI_DEV_ECC_SBE_VOL_SHM, gauge, Ecc sbe vol shm
257 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_srm
258 # DCGM_FI_DEV_ECC_SBE_VOL_SRM, gauge, Ecc sbe vol srm
259 # context=dcgm.gpu.memory.ecc_errors family=gpu memory dimension=ecc_sbe_vol_tex
260 # DCGM_FI_DEV_ECC_SBE_VOL_TEX, gauge, Ecc sbe vol tex
261 # context=dcgm.gpu.memory.ecc_error_rate family=gpu memory dimension=ecc_sbe_vol
262 DCGM_FI_DEV_ECC_SBE_VOL_TOTAL, counter, Ecc sbe vol total
263 # context=dcgm.gpu.workload.sessions family=gpu workload dimension=enc_stats
264 # DCGM_FI_DEV_ENC_STATS, gauge, Enc stats
265 # context=dcgm.gpu.compute.utilization family=gpu compute dimension=encoder
266 DCGM_FI_DEV_ENC_UTIL, gauge, Enc util
267 # context=dcgm.gpu.power.usage family=gpu power dimension=enforced_limit
268 DCGM_FI_DEV_ENFORCED_POWER_LIMIT, gauge, Enforced power limit
269 # context=dcgm.gpu.power.profiles family=gpu power dimension=enforced_power_profile_mask
270 # DCGM_FI_DEV_ENFORCED_POWER_PROFILE_MASK, counter, Enforced power profile mask
271 # context=dcgm.gpu.interconnect.fabric family=gpu interconnect/overview dimension=fabric_clique_id
272 # DCGM_FI_DEV_FABRIC_CLIQUE_ID, gauge, Fabric clique id
273 # context=dcgm.gpu.interconnect.fabric family=gpu interconnect/overview dimension=fabric_cluster_uuid
274 # DCGM_FI_DEV_FABRIC_CLUSTER_UUID, label, Fabric cluster uuid
275 # context=dcgm.gpu.interconnect.fabric family=gpu interconnect/overview dimension=fabric_health_mask
276 # DCGM_FI_DEV_FABRIC_HEALTH_MASK, gauge, Fabric health mask
277 # context=dcgm.gpu.interconnect.fabric family=gpu interconnect/overview dimension=fabric_manager_error_code
278 DCGM_FI_DEV_FABRIC_MANAGER_ERROR_CODE, counter, Fabric manager error code
279 # context=dcgm.gpu.interconnect.fabric family=gpu interconnect/overview dimension=fabric_manager_status
280 DCGM_FI_DEV_FABRIC_MANAGER_STATUS, gauge, Fabric manager status
281 # context=dcgm.gpu.thermal.fan_speed family=gpu thermal dimension=fan_speed
282 DCGM_FI_DEV_FAN_SPEED, gauge, Fan speed
283 # context=dcgm.gpu.workload.sessions family=gpu workload dimension=fbc_sessions_info
284 # DCGM_FI_DEV_FBC_SESSIONS_INFO, gauge, Fbc sessions info
285 # context=dcgm.gpu.workload.sessions family=gpu workload dimension=fbc_stats
286 # DCGM_FI_DEV_FBC_STATS, gauge, Fbc stats
287 # context=dcgm.gpu.memory.usage family=gpu memory dimension=free
288 DCGM_FI_DEV_FB_FREE, gauge, Fb free
289 # context=dcgm.gpu.memory.usage family=gpu memory dimension=reserved
290 DCGM_FI_DEV_FB_RESERVED, gauge, Fb reserved
291 # context=dcgm.gpu.memory.capacity family=gpu memory dimension=total
292 DCGM_FI_DEV_FB_TOTAL, gauge, Fb total
293 # context=dcgm.gpu.memory.usage family=gpu memory dimension=used
294 DCGM_FI_DEV_FB_USED, gauge, Fb used
295 # context=dcgm.gpu.memory.utilization family=gpu memory dimension=used_percent
296 DCGM_FI_DEV_FB_USED_PERCENT, gauge, Fb used percent
297 # context=dcgm.gpu.internal.boundary family=gpu internal dimension=first_connectx_field_id
298 # DCGM_FI_DEV_FIRST_CONNECTX_FIELD_ID, gauge, First connectx field id
299 # context=dcgm.gpu.reliability.recovery_action family=gpu reliability dimension=get_gpu_recovery_action
300 # DCGM_FI_DEV_GET_GPU_RECOVERY_ACTION, counter, Get gpu recovery action
301 # context=dcgm.gpu.capability.support family=gpu capability dimension=gpm_support
302 # DCGM_FI_DEV_GPM_SUPPORT, gauge, Gpm support
303 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=gpu_max_op_temp
304 DCGM_FI_DEV_GPU_MAX_OP_TEMP, gauge, Gpu max op temp
305 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=gpu_nvlink_errors
306 DCGM_FI_DEV_GPU_NVLINK_ERRORS, counter, Gpu nvlink errors
307 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=gpu
308 DCGM_FI_DEV_GPU_TEMP, gauge, Gpu temp
309 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=gpu_temp_limit
310 DCGM_FI_DEV_GPU_TEMP_LIMIT, gauge, Gpu temp limit
311 # context=dcgm.gpu.compute.utilization family=gpu compute dimension=gpu
312 DCGM_FI_DEV_GPU_UTIL, gauge, Gpu util
313 # context=dcgm.gpu.inventory.software family=gpu inventory dimension=inforom_config_check
314 # DCGM_FI_DEV_INFOROM_CONFIG_CHECK, label, Inforom config check
315 # context=dcgm.gpu.inventory.software family=gpu inventory dimension=inforom_config_valid
316 # DCGM_FI_DEV_INFOROM_CONFIG_VALID, label, Inforom config valid
317 # context=dcgm.gpu.inventory.software family=gpu inventory dimension=inforom_image_ver
318 # DCGM_FI_DEV_INFOROM_IMAGE_VER, label, Inforom image ver
319 # context=dcgm.gpu.internal.boundary family=gpu internal dimension=last_connectx_field_id
320 # DCGM_FI_DEV_LAST_CONNECTX_FIELD_ID, gauge, Last connectx field id
321 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=low_utilization_violation
322 # DCGM_FI_DEV_LOW_UTIL_VIOLATION, gauge, Low util violation
323 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=max_mem_clock
324 DCGM_FI_DEV_MAX_MEM_CLOCK, gauge, Max mem clock
325 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=max_sm_clock
326 DCGM_FI_DEV_MAX_SM_CLOCK, gauge, Max sm clock
327 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=max_video_clock
328 DCGM_FI_DEV_MAX_VIDEO_CLOCK, gauge, Max video clock
329 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=memory
330 DCGM_FI_DEV_MEMORY_TEMP, gauge, Memory temp
331 # context=dcgm.gpu.reliability.memory_health family=gpu reliability dimension=memory_unrepairable_flag
332 # DCGM_FI_DEV_MEMORY_UNREPAIRABLE_FLAG, gauge, Memory unrepairable flag
333 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=mem_affinity_0
334 # DCGM_FI_DEV_MEM_AFFINITY_0, gauge, Mem affinity 0
335 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=mem_affinity_1
336 # DCGM_FI_DEV_MEM_AFFINITY_1, gauge, Mem affinity 1
337 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=mem_affinity_2
338 # DCGM_FI_DEV_MEM_AFFINITY_2, gauge, Mem affinity 2
339 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=mem_affinity_3
340 # DCGM_FI_DEV_MEM_AFFINITY_3, gauge, Mem affinity 3
341 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=memory
342 DCGM_FI_DEV_MEM_CLOCK, gauge, Mem clock
343 # context=dcgm.gpu.compute.utilization family=gpu compute dimension=memory_copy
344 DCGM_FI_DEV_MEM_COPY_UTIL, gauge, Mem copy util
345 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=mem_max_op_temp
346 DCGM_FI_DEV_MEM_MAX_OP_TEMP, gauge, Mem max op temp
347 # context=dcgm.gpu.capability.support family=gpu capability dimension=mig_attributes
348 # DCGM_FI_DEV_MIG_ATTRIBUTES, gauge, Mig attributes
349 # context=dcgm.gpu.capability.support family=gpu capability dimension=mig_ci_info
350 # DCGM_FI_DEV_MIG_CI_INFO, gauge, Mig ci info
351 # context=dcgm.gpu.capability.support family=gpu capability dimension=mig_gi_info
352 # DCGM_FI_DEV_MIG_GI_INFO, gauge, Mig gi info
353 # context=dcgm.gpu.capability.support family=gpu capability dimension=mig_max_slices
354 DCGM_FI_DEV_MIG_MAX_SLICES, gauge, Mig max slices
355 # context=dcgm.gpu.state.virtualization family=gpu state dimension=mig_mode
356 DCGM_FI_DEV_MIG_MODE, gauge, Mig mode
357 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=minor_number
358 # DCGM_FI_DEV_MINOR_NUMBER, label, Minor number
359 # context=dcgm.gpu.cpu.power family=gpu cpu dimension=module_power_util_current
360 # DCGM_FI_DEV_MODULE_POWER_UTIL_CURRENT, gauge, Module power util current
361 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=name
362 # DCGM_FI_DEV_NAME, label, Name
363 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l0
364 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L0, counter, Nvlink bandwidth l0
365 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l1
366 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L1, counter, Nvlink bandwidth l1
367 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l10
368 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L10, counter, Nvlink bandwidth l10
369 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l11
370 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L11, counter, Nvlink bandwidth l11
371 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l12
372 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L12, counter, Nvlink bandwidth l12
373 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l13
374 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L13, counter, Nvlink bandwidth l13
375 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l14
376 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L14, counter, Nvlink bandwidth l14
377 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l15
378 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L15, counter, Nvlink bandwidth l15
379 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l16
380 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L16, counter, Nvlink bandwidth l16
381 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l17
382 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L17, counter, Nvlink bandwidth l17
383 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l2
384 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L2, counter, Nvlink bandwidth l2
385 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l3
386 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L3, counter, Nvlink bandwidth l3
387 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l4
388 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L4, counter, Nvlink bandwidth l4
389 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l5
390 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L5, counter, Nvlink bandwidth l5
391 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l6
392 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L6, counter, Nvlink bandwidth l6
393 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l7
394 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L7, counter, Nvlink bandwidth l7
395 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l8
396 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L8, counter, Nvlink bandwidth l8
397 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth_l9
398 # DCGM_FI_DEV_NVLINK_BANDWIDTH_L9, counter, Nvlink bandwidth l9
399 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_bandwidth
400 DCGM_FI_DEV_NVLINK_BANDWIDTH_TOTAL, counter, Nvlink bandwidth total
401 # context=dcgm.nvlink.interconnect.ber family=nvlink interconnect dimension=nvlink_count_effective_ber
402 # DCGM_FI_DEV_NVLINK_COUNT_EFFECTIVE_BER, counter, Nvlink count effective ber
403 # context=dcgm.nvlink.interconnect.ber family=nvlink interconnect dimension=nvlink_count_effective_ber_float
404 # DCGM_FI_DEV_NVLINK_COUNT_EFFECTIVE_BER_FLOAT, counter, Nvlink count effective ber float
405 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_effective_errors
406 # DCGM_FI_DEV_NVLINK_COUNT_EFFECTIVE_ERRORS, counter, Nvlink count effective errors
407 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_0
408 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_0, counter, Nvlink count fec history 0
409 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_1
410 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_1, counter, Nvlink count fec history 1
411 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_10
412 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_10, counter, Nvlink count fec history 10
413 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_11
414 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_11, counter, Nvlink count fec history 11
415 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_12
416 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_12, counter, Nvlink count fec history 12
417 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_13
418 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_13, counter, Nvlink count fec history 13
419 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_14
420 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_14, counter, Nvlink count fec history 14
421 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_15
422 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_15, counter, Nvlink count fec history 15
423 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_2
424 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_2, counter, Nvlink count fec history 2
425 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_3
426 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_3, counter, Nvlink count fec history 3
427 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_4
428 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_4, counter, Nvlink count fec history 4
429 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_5
430 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_5, counter, Nvlink count fec history 5
431 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_6
432 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_6, counter, Nvlink count fec history 6
433 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_7
434 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_7, counter, Nvlink count fec history 7
435 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_8
436 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_8, counter, Nvlink count fec history 8
437 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_fec_history_9
438 # DCGM_FI_DEV_NVLINK_COUNT_FEC_HISTORY_9, counter, Nvlink count fec history 9
439 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_link_recovery_events
440 # DCGM_FI_DEV_NVLINK_COUNT_LINK_RECOVERY_EVENTS, counter, Nvlink count link recovery events
441 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_link_recovery_failed_events
442 # DCGM_FI_DEV_NVLINK_COUNT_LINK_RECOVERY_FAILED_EVENTS, counter, Nvlink count link recovery failed events
443 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_link_recovery_successful_events
444 # DCGM_FI_DEV_NVLINK_COUNT_LINK_RECOVERY_SUCCESSFUL_EVENTS, counter, Nvlink count link recovery successful events
445 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_local_link_integrity_errors
446 # DCGM_FI_DEV_NVLINK_COUNT_LOCAL_LINK_INTEGRITY_ERRORS, counter, Nvlink count local link integrity errors
447 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_rx_buffer_overrun_errors
448 # DCGM_FI_DEV_NVLINK_COUNT_RX_BUFFER_OVERRUN_ERRORS, counter, Nvlink count rx buffer overrun errors
449 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_count_rx
450 DCGM_FI_DEV_NVLINK_COUNT_RX_BYTES, counter, Nvlink count rx bytes
451 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_rx_errors
452 # DCGM_FI_DEV_NVLINK_COUNT_RX_ERRORS, counter, Nvlink count rx errors
453 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_rx_general_errors
454 # DCGM_FI_DEV_NVLINK_COUNT_RX_GENERAL_ERRORS, counter, Nvlink count rx general errors
455 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_rx_malformed_packet_errors
456 # DCGM_FI_DEV_NVLINK_COUNT_RX_MALFORMED_PACKET_ERRORS, counter, Nvlink count rx malformed packet errors
457 # context=dcgm.nvlink.interconnect.traffic family=nvlink interconnect dimension=nvlink_count_rx_packets
458 # DCGM_FI_DEV_NVLINK_COUNT_RX_PACKETS, counter, Nvlink count rx packets
459 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_rx_remote_errors
460 # DCGM_FI_DEV_NVLINK_COUNT_RX_REMOTE_ERRORS, counter, Nvlink count rx remote errors
461 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_rx_symbol_errors
462 # DCGM_FI_DEV_NVLINK_COUNT_RX_SYMBOL_ERRORS, counter, Nvlink count rx symbol errors
463 # context=dcgm.nvlink.interconnect.ber family=nvlink interconnect dimension=nvlink_count_symbol_ber
464 # DCGM_FI_DEV_NVLINK_COUNT_SYMBOL_BER, counter, Nvlink count symbol ber
465 # context=dcgm.nvlink.interconnect.ber family=nvlink interconnect dimension=nvlink_count_symbol_ber_float
466 # DCGM_FI_DEV_NVLINK_COUNT_SYMBOL_BER_FLOAT, counter, Nvlink count symbol ber float
467 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_count_tx
468 DCGM_FI_DEV_NVLINK_COUNT_TX_BYTES, counter, Nvlink count tx bytes
469 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_count_tx_discards
470 # DCGM_FI_DEV_NVLINK_COUNT_TX_DISCARDS, counter, Nvlink count tx discards
471 # context=dcgm.nvlink.interconnect.traffic family=nvlink interconnect dimension=nvlink_count_tx_packets
472 # DCGM_FI_DEV_NVLINK_COUNT_TX_PACKETS, counter, Nvlink count tx packets
473 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l0
474 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L0, counter, Nvlink crc data error count l0
475 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l1
476 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L1, counter, Nvlink crc data error count l1
477 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l10
478 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L10, counter, Nvlink crc data error count l10
479 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l11
480 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L11, counter, Nvlink crc data error count l11
481 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l12
482 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L12, counter, Nvlink crc data error count l12
483 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l13
484 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L13, counter, Nvlink crc data error count l13
485 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l14
486 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L14, counter, Nvlink crc data error count l14
487 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l15
488 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L15, counter, Nvlink crc data error count l15
489 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l16
490 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L16, counter, Nvlink crc data error count l16
491 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l17
492 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L17, counter, Nvlink crc data error count l17
493 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l2
494 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L2, counter, Nvlink crc data error count l2
495 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l3
496 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L3, counter, Nvlink crc data error count l3
497 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l4
498 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L4, counter, Nvlink crc data error count l4
499 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l5
500 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L5, counter, Nvlink crc data error count l5
501 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l6
502 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L6, counter, Nvlink crc data error count l6
503 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l7
504 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L7, counter, Nvlink crc data error count l7
505 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l8
506 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L8, counter, Nvlink crc data error count l8
507 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error_count_l9
508 # DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_L9, counter, Nvlink crc data error count l9
509 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_data_error
510 DCGM_FI_DEV_NVLINK_CRC_DATA_ERROR_COUNT_TOTAL, counter, Nvlink crc data error count total
511 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l0
512 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L0, counter, Nvlink crc flit error count l0
513 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l1
514 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L1, counter, Nvlink crc flit error count l1
515 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l10
516 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L10, counter, Nvlink crc flit error count l10
517 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l11
518 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L11, counter, Nvlink crc flit error count l11
519 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l12
520 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L12, counter, Nvlink crc flit error count l12
521 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l13
522 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L13, counter, Nvlink crc flit error count l13
523 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l14
524 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L14, counter, Nvlink crc flit error count l14
525 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l15
526 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L15, counter, Nvlink crc flit error count l15
527 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l16
528 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L16, counter, Nvlink crc flit error count l16
529 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l17
530 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L17, counter, Nvlink crc flit error count l17
531 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l2
532 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L2, counter, Nvlink crc flit error count l2
533 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l3
534 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L3, counter, Nvlink crc flit error count l3
535 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l4
536 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L4, counter, Nvlink crc flit error count l4
537 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l5
538 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L5, counter, Nvlink crc flit error count l5
539 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l6
540 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L6, counter, Nvlink crc flit error count l6
541 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l7
542 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L7, counter, Nvlink crc flit error count l7
543 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l8
544 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L8, counter, Nvlink crc flit error count l8
545 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error_count_l9
546 # DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_L9, counter, Nvlink crc flit error count l9
547 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_crc_flit_error
548 DCGM_FI_DEV_NVLINK_CRC_FLIT_ERROR_COUNT_TOTAL, counter, Nvlink crc flit error count total
549 # context=dcgm.nvlink.memory.ecc_error_rate family=nvlink memory dimension=nvlink_ecc_data_error
550 # DCGM_FI_DEV_NVLINK_ECC_DATA_ERROR_COUNT_TOTAL, counter, Nvlink ecc data error count total
551 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_error_dl_crc
552 DCGM_FI_DEV_NVLINK_ERROR_DL_CRC, counter, Nvlink error dl crc
553 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_error_dl_recovery
554 DCGM_FI_DEV_NVLINK_ERROR_DL_RECOVERY, counter, Nvlink error dl recovery
555 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_error_dl_replay
556 DCGM_FI_DEV_NVLINK_ERROR_DL_REPLAY, counter, Nvlink error dl replay
557 # context=dcgm.nvlink.interconnect.state family=nvlink interconnect dimension=nvlink_get_state
558 # DCGM_FI_DEV_NVLINK_GET_STATE, gauge, Nvlink get state
559 # context=dcgm.nvlink.interconnect.congestion family=nvlink interconnect dimension=nvlink_ppcnt_ibpc_port_xmit_wait
560 # DCGM_FI_DEV_NVLINK_PPCNT_IBPC_PORT_XMIT_WAIT, gauge, Nvlink ppcnt ibpc port xmit wait
561 # context=dcgm.nvlink.interconnect.state family=nvlink interconnect dimension=nvlink_ppcnt_physical_link_down_counter
562 # DCGM_FI_DEV_NVLINK_PPCNT_PHYSICAL_LINK_DOWN_COUNTER, counter, Nvlink ppcnt physical link down counter
563 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_ppcnt_physical_successful_recovery_events
564 # DCGM_FI_DEV_NVLINK_PPCNT_PHYSICAL_SUCCESSFUL_RECOVERY_EVENTS, counter, Nvlink ppcnt physical successful recovery events
565 # context=dcgm.nvlink.interconnect.traffic family=nvlink interconnect dimension=nvlink_ppcnt_plr_rcv_codes
566 # DCGM_FI_DEV_NVLINK_PPCNT_PLR_RCV_CODES, counter, Nvlink ppcnt plr rcv codes
567 # context=dcgm.nvlink.interconnect.state family=nvlink interconnect dimension=nvlink_ppcnt_plr_rcv_code_err
568 # DCGM_FI_DEV_NVLINK_PPCNT_PLR_RCV_CODE_ERR, gauge, Nvlink ppcnt plr rcv code err
569 # context=dcgm.nvlink.interconnect.errors family=nvlink interconnect dimension=nvlink_ppcnt_plr_rcv_uncorrectable_code
570 # DCGM_FI_DEV_NVLINK_PPCNT_PLR_RCV_UNCORRECTABLE_CODE, gauge, Nvlink ppcnt plr rcv uncorrectable code
571 # context=dcgm.nvlink.interconnect.state family=nvlink interconnect dimension=nvlink_ppcnt_plr_sync_events
572 # DCGM_FI_DEV_NVLINK_PPCNT_PLR_SYNC_EVENTS, gauge, Nvlink ppcnt plr sync events
573 # context=dcgm.nvlink.interconnect.traffic family=nvlink interconnect dimension=nvlink_ppcnt_plr_xmit_codes
574 # DCGM_FI_DEV_NVLINK_PPCNT_PLR_XMIT_CODES, counter, Nvlink ppcnt plr xmit codes
575 # context=dcgm.nvlink.interconnect.traffic family=nvlink interconnect dimension=nvlink_ppcnt_plr_xmit_retry_codes
576 # DCGM_FI_DEV_NVLINK_PPCNT_PLR_XMIT_RETRY_CODES, counter, Nvlink ppcnt plr xmit retry codes
577 # context=dcgm.nvlink.interconnect.state family=nvlink interconnect dimension=nvlink_ppcnt_plr_xmit_retry_events
578 # DCGM_FI_DEV_NVLINK_PPCNT_PLR_XMIT_RETRY_EVENTS, gauge, Nvlink ppcnt plr xmit retry events
579 # context=dcgm.nvlink.internal.boundary family=nvlink internal dimension=nvlink_ppcnt_recovery_time_between_last_two
580 # DCGM_FI_DEV_NVLINK_PPCNT_RECOVERY_TIME_BETWEEN_LAST_TWO, counter, Nvlink ppcnt recovery time between last two
581 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_ppcnt_recovery_time_since_last
582 # DCGM_FI_DEV_NVLINK_PPCNT_RECOVERY_TIME_SINCE_LAST, counter, Nvlink ppcnt recovery time since last
583 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_ppcnt_recovery_total_successful_events
584 # DCGM_FI_DEV_NVLINK_PPCNT_RECOVERY_TOTAL_SUCCESSFUL_EVENTS, counter, Nvlink ppcnt recovery total successful events
585 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_pprm_oper_recovery
586 # DCGM_FI_DEV_NVLINK_PPRM_OPER_RECOVERY, counter, Nvlink pprm oper recovery
587 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l0
588 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L0, counter, Nvlink recovery error count l0
589 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l1
590 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L1, counter, Nvlink recovery error count l1
591 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l10
592 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L10, counter, Nvlink recovery error count l10
593 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l11
594 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L11, counter, Nvlink recovery error count l11
595 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l12
596 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L12, counter, Nvlink recovery error count l12
597 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l13
598 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L13, counter, Nvlink recovery error count l13
599 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l14
600 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L14, counter, Nvlink recovery error count l14
601 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l15
602 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L15, counter, Nvlink recovery error count l15
603 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l16
604 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L16, counter, Nvlink recovery error count l16
605 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l17
606 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L17, counter, Nvlink recovery error count l17
607 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l2
608 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L2, counter, Nvlink recovery error count l2
609 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l3
610 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L3, counter, Nvlink recovery error count l3
611 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l4
612 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L4, counter, Nvlink recovery error count l4
613 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l5
614 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L5, counter, Nvlink recovery error count l5
615 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l6
616 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L6, counter, Nvlink recovery error count l6
617 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l7
618 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L7, counter, Nvlink recovery error count l7
619 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l8
620 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L8, counter, Nvlink recovery error count l8
621 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error_count_l9
622 # DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_L9, counter, Nvlink recovery error count l9
623 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_recovery_error
624 DCGM_FI_DEV_NVLINK_RECOVERY_ERROR_COUNT_TOTAL, counter, Nvlink recovery error count total
625 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l0
626 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L0, counter, Nvlink replay error count l0
627 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l1
628 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L1, counter, Nvlink replay error count l1
629 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l10
630 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L10, counter, Nvlink replay error count l10
631 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l11
632 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L11, counter, Nvlink replay error count l11
633 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l12
634 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L12, counter, Nvlink replay error count l12
635 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l13
636 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L13, counter, Nvlink replay error count l13
637 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l14
638 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L14, counter, Nvlink replay error count l14
639 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l15
640 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L15, counter, Nvlink replay error count l15
641 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l16
642 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L16, counter, Nvlink replay error count l16
643 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l17
644 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L17, counter, Nvlink replay error count l17
645 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l2
646 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L2, counter, Nvlink replay error count l2
647 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l3
648 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L3, counter, Nvlink replay error count l3
649 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l4
650 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L4, counter, Nvlink replay error count l4
651 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l5
652 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L5, counter, Nvlink replay error count l5
653 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l6
654 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L6, counter, Nvlink replay error count l6
655 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l7
656 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L7, counter, Nvlink replay error count l7
657 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l8
658 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L8, counter, Nvlink replay error count l8
659 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error_count_l9
660 # DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_L9, counter, Nvlink replay error count l9
661 # context=dcgm.nvlink.interconnect.error_rate family=nvlink interconnect dimension=nvlink_replay_error
662 DCGM_FI_DEV_NVLINK_REPLAY_ERROR_COUNT_TOTAL, counter, Nvlink replay error count total
663 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l0
664 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L0, counter, Nvlink rx bandwidth l0
665 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l1
666 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L1, counter, Nvlink rx bandwidth l1
667 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l10
668 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L10, counter, Nvlink rx bandwidth l10
669 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l11
670 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L11, counter, Nvlink rx bandwidth l11
671 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l12
672 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L12, counter, Nvlink rx bandwidth l12
673 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l13
674 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L13, counter, Nvlink rx bandwidth l13
675 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l14
676 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L14, counter, Nvlink rx bandwidth l14
677 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l15
678 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L15, counter, Nvlink rx bandwidth l15
679 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l16
680 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L16, counter, Nvlink rx bandwidth l16
681 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l17
682 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L17, counter, Nvlink rx bandwidth l17
683 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l2
684 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L2, counter, Nvlink rx bandwidth l2
685 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l3
686 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L3, counter, Nvlink rx bandwidth l3
687 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l4
688 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L4, counter, Nvlink rx bandwidth l4
689 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l5
690 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L5, counter, Nvlink rx bandwidth l5
691 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l6
692 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L6, counter, Nvlink rx bandwidth l6
693 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l7
694 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L7, counter, Nvlink rx bandwidth l7
695 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l8
696 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L8, counter, Nvlink rx bandwidth l8
697 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth_l9
698 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_L9, counter, Nvlink rx bandwidth l9
699 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx_bandwidth
700 # DCGM_FI_DEV_NVLINK_RX_BANDWIDTH_TOTAL, counter, Nvlink rx bandwidth total
701 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l0
702 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L0, counter, Nvlink tx bandwidth l0
703 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l1
704 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L1, counter, Nvlink tx bandwidth l1
705 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l10
706 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L10, counter, Nvlink tx bandwidth l10
707 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l11
708 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L11, counter, Nvlink tx bandwidth l11
709 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l12
710 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L12, counter, Nvlink tx bandwidth l12
711 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l13
712 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L13, counter, Nvlink tx bandwidth l13
713 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l14
714 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L14, counter, Nvlink tx bandwidth l14
715 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l15
716 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L15, counter, Nvlink tx bandwidth l15
717 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l16
718 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L16, counter, Nvlink tx bandwidth l16
719 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l17
720 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L17, counter, Nvlink tx bandwidth l17
721 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l2
722 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L2, counter, Nvlink tx bandwidth l2
723 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l3
724 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L3, counter, Nvlink tx bandwidth l3
725 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l4
726 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L4, counter, Nvlink tx bandwidth l4
727 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l5
728 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L5, counter, Nvlink tx bandwidth l5
729 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l6
730 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L6, counter, Nvlink tx bandwidth l6
731 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l7
732 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L7, counter, Nvlink tx bandwidth l7
733 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l8
734 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L8, counter, Nvlink tx bandwidth l8
735 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth_l9
736 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_L9, counter, Nvlink tx bandwidth l9
737 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx_bandwidth
738 # DCGM_FI_DEV_NVLINK_TX_BANDWIDTH_TOTAL, counter, Nvlink tx bandwidth total
739 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=nvml_index
740 # DCGM_FI_DEV_NVML_INDEX, gauge, Nvml index
741 # context=dcgm.nvswitch.interconnect.nvswitch.current family=nvswitch interconnect dimension=nvswitch_current_iddq
742 # DCGM_FI_DEV_NVSWITCH_CURRENT_IDDQ, gauge, Nvswitch current iddq
743 # context=dcgm.nvswitch.interconnect.nvswitch.current family=nvswitch interconnect dimension=nvswitch_current_iddq_dvdd
744 # DCGM_FI_DEV_NVSWITCH_CURRENT_IDDQ_DVDD, gauge, Nvswitch current iddq dvdd
745 # context=dcgm.nvswitch.interconnect.nvswitch.current family=nvswitch interconnect dimension=nvswitch_current_iddq_rev
746 # DCGM_FI_DEV_NVSWITCH_CURRENT_IDDQ_REV, gauge, Nvswitch current iddq rev
747 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_device_uuid
748 # DCGM_FI_DEV_NVSWITCH_DEVICE_UUID, label, Nvswitch device uuid
749 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_fatal_errors
750 DCGM_FI_DEV_NVSWITCH_FATAL_ERRORS, counter, Nvswitch fatal errors
751 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors
752 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS, counter, Nvswitch link crc errors
753 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane0
754 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE0, counter, Nvswitch link crc errors lane0
755 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane1
756 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE1, counter, Nvswitch link crc errors lane1
757 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane2
758 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE2, counter, Nvswitch link crc errors lane2
759 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane3
760 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE3, counter, Nvswitch link crc errors lane3
761 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane4
762 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE4, counter, Nvswitch link crc errors lane4
763 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane5
764 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE5, counter, Nvswitch link crc errors lane5
765 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane6
766 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE6, counter, Nvswitch link crc errors lane6
767 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_crc_errors_lane7
768 # DCGM_FI_DEV_NVSWITCH_LINK_CRC_ERRORS_LANE7, counter, Nvswitch link crc errors lane7
769 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_link_device_link_id
770 # DCGM_FI_DEV_NVSWITCH_LINK_DEVICE_LINK_ID, gauge, Nvswitch link device link id
771 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_link_device_link_sid
772 # DCGM_FI_DEV_NVSWITCH_LINK_DEVICE_LINK_SID, gauge, Nvswitch link device link sid
773 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors
774 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS, counter, Nvswitch link ecc errors
775 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane0
776 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE0, counter, Nvswitch link ecc errors lane0
777 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane1
778 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE1, counter, Nvswitch link ecc errors lane1
779 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane2
780 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE2, counter, Nvswitch link ecc errors lane2
781 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane3
782 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE3, counter, Nvswitch link ecc errors lane3
783 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane4
784 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE4, counter, Nvswitch link ecc errors lane4
785 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane5
786 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE5, counter, Nvswitch link ecc errors lane5
787 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane6
788 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE6, counter, Nvswitch link ecc errors lane6
789 # context=dcgm.nvswitch.memory.ecc_error_rate family=nvswitch memory dimension=nvswitch_link_ecc_errors_lane7
790 # DCGM_FI_DEV_NVSWITCH_LINK_ECC_ERRORS_LANE7, counter, Nvswitch link ecc errors lane7
791 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_fatal_errors
792 DCGM_FI_DEV_NVSWITCH_LINK_FATAL_ERRORS, counter, Nvswitch link fatal errors
793 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_flit_errors
794 # DCGM_FI_DEV_NVSWITCH_LINK_FLIT_ERRORS, counter, Nvswitch link flit errors
795 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_link_id
796 # DCGM_FI_DEV_NVSWITCH_LINK_ID, gauge, Nvswitch link id
797 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_count_vc0
798 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_COUNT_VC0, counter, Nvswitch link latency count vc0
799 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_count_vc1
800 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_COUNT_VC1, counter, Nvswitch link latency count vc1
801 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_count_vc2
802 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_COUNT_VC2, counter, Nvswitch link latency count vc2
803 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_count_vc3
804 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_COUNT_VC3, counter, Nvswitch link latency count vc3
805 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_high_vc0
806 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_HIGH_VC0, counter, Nvswitch link latency high vc0
807 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_high_vc1
808 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_HIGH_VC1, counter, Nvswitch link latency high vc1
809 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_high_vc2
810 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_HIGH_VC2, counter, Nvswitch link latency high vc2
811 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_high_vc3
812 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_HIGH_VC3, counter, Nvswitch link latency high vc3
813 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_low_vc0
814 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_LOW_VC0, counter, Nvswitch link latency low vc0
815 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_low_vc1
816 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_LOW_VC1, counter, Nvswitch link latency low vc1
817 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_low_vc2
818 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_LOW_VC2, counter, Nvswitch link latency low vc2
819 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_low_vc3
820 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_LOW_VC3, counter, Nvswitch link latency low vc3
821 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_medium_vc0
822 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_MEDIUM_VC0, counter, Nvswitch link latency medium vc0
823 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_medium_vc1
824 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_MEDIUM_VC1, counter, Nvswitch link latency medium vc1
825 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_medium_vc2
826 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_MEDIUM_VC2, counter, Nvswitch link latency medium vc2
827 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_medium_vc3
828 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_MEDIUM_VC3, counter, Nvswitch link latency medium vc3
829 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_panic_vc0
830 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_PANIC_VC0, counter, Nvswitch link latency panic vc0
831 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_panic_vc1
832 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_PANIC_VC1, counter, Nvswitch link latency panic vc1
833 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_panic_vc2
834 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_PANIC_VC2, counter, Nvswitch link latency panic vc2
835 # context=dcgm.nvswitch.interconnect.nvswitch.latency family=nvswitch interconnect dimension=nvswitch_link_latency_panic_vc3
836 # DCGM_FI_DEV_NVSWITCH_LINK_LATENCY_PANIC_VC3, counter, Nvswitch link latency panic vc3
837 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_non_fatal_errors
838 DCGM_FI_DEV_NVSWITCH_LINK_NON_FATAL_ERRORS, counter, Nvswitch link non fatal errors
839 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_recovery_errors
840 # DCGM_FI_DEV_NVSWITCH_LINK_RECOVERY_ERRORS, counter, Nvswitch link recovery errors
841 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_link_remote_pcie_bus
842 # DCGM_FI_DEV_NVSWITCH_LINK_REMOTE_PCIE_BUS, gauge, Nvswitch link remote pcie bus
843 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_link_remote_pcie_device
844 # DCGM_FI_DEV_NVSWITCH_LINK_REMOTE_PCIE_DEVICE, gauge, Nvswitch link remote pcie device
845 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_link_remote_pcie_domain
846 # DCGM_FI_DEV_NVSWITCH_LINK_REMOTE_PCIE_DOMAIN, gauge, Nvswitch link remote pcie domain
847 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_link_remote_pcie_function
848 # DCGM_FI_DEV_NVSWITCH_LINK_REMOTE_PCIE_FUNCTION, gauge, Nvswitch link remote pcie function
849 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_link_replay_errors
850 DCGM_FI_DEV_NVSWITCH_LINK_REPLAY_ERRORS, counter, Nvswitch link replay errors
851 # context=dcgm.nvswitch.interconnect.nvswitch.status family=nvswitch interconnect dimension=nvswitch_link_status
852 # DCGM_FI_DEV_NVSWITCH_LINK_STATUS, gauge, Nvswitch link status
853 # context=dcgm.nvswitch.interconnect.nvswitch.throughput family=nvswitch interconnect dimension=nvswitch_link_throughput_rx
854 DCGM_FI_DEV_NVSWITCH_LINK_THROUGHPUT_RX, counter, Nvswitch link throughput rx
855 # context=dcgm.nvswitch.interconnect.nvswitch.throughput family=nvswitch interconnect dimension=nvswitch_link_throughput_tx
856 DCGM_FI_DEV_NVSWITCH_LINK_THROUGHPUT_TX, counter, Nvswitch link throughput tx
857 # context=dcgm.nvswitch.interconnect.nvswitch.status family=nvswitch interconnect dimension=nvswitch_link_type
858 # DCGM_FI_DEV_NVSWITCH_LINK_TYPE, gauge, Nvswitch link type
859 # context=dcgm.nvswitch.interconnect.nvswitch.errors family=nvswitch interconnect dimension=nvswitch_non_fatal_errors
860 DCGM_FI_DEV_NVSWITCH_NON_FATAL_ERRORS, counter, Nvswitch non fatal errors
861 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_pcie_bus
862 # DCGM_FI_DEV_NVSWITCH_PCIE_BUS, gauge, Nvswitch pcie bus
863 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_pcie_device
864 # DCGM_FI_DEV_NVSWITCH_PCIE_DEVICE, gauge, Nvswitch pcie device
865 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_pcie_domain
866 # DCGM_FI_DEV_NVSWITCH_PCIE_DOMAIN, gauge, Nvswitch pcie domain
867 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_pcie_function
868 # DCGM_FI_DEV_NVSWITCH_PCIE_FUNCTION, gauge, Nvswitch pcie function
869 # context=dcgm.nvswitch.interconnect.nvswitch.topology family=nvswitch interconnect dimension=nvswitch_phys_id
870 # DCGM_FI_DEV_NVSWITCH_PHYS_ID, gauge, Nvswitch phys id
871 # context=dcgm.nvswitch.interconnect.nvswitch.power family=nvswitch interconnect dimension=nvswitch_power_dvdd
872 # DCGM_FI_DEV_NVSWITCH_POWER_DVDD, gauge, Nvswitch power dvdd
873 # context=dcgm.nvswitch.interconnect.nvswitch.power family=nvswitch interconnect dimension=nvswitch_power_hvdd
874 # DCGM_FI_DEV_NVSWITCH_POWER_HVDD, gauge, Nvswitch power hvdd
875 # context=dcgm.nvswitch.interconnect.nvswitch.power family=nvswitch interconnect dimension=nvswitch_power_vdd
876 # DCGM_FI_DEV_NVSWITCH_POWER_VDD, gauge, Nvswitch power vdd
877 # context=dcgm.nvswitch.interconnect.nvswitch.status family=nvswitch interconnect dimension=nvswitch_reset_required
878 # DCGM_FI_DEV_NVSWITCH_RESET_REQUIRED, gauge, Nvswitch reset required
879 # context=dcgm.nvswitch.thermal.temperature family=nvswitch thermal dimension=nvswitch_temperature_current
880 DCGM_FI_DEV_NVSWITCH_TEMPERATURE_CURRENT, gauge, Nvswitch temperature current
881 # context=dcgm.nvswitch.thermal.temperature family=nvswitch thermal dimension=nvswitch_temperature_limit_shutdown
882 # DCGM_FI_DEV_NVSWITCH_TEMPERATURE_LIMIT_SHUTDOWN, gauge, Nvswitch temperature limit shutdown
883 # context=dcgm.nvswitch.thermal.temperature family=nvswitch thermal dimension=nvswitch_temperature_limit_slowdown
884 # DCGM_FI_DEV_NVSWITCH_TEMPERATURE_LIMIT_SLOWDOWN, gauge, Nvswitch temperature limit slowdown
885 # context=dcgm.nvswitch.interconnect.nvswitch.throughput family=nvswitch interconnect dimension=nvswitch_throughput_rx
886 # DCGM_FI_DEV_NVSWITCH_THROUGHPUT_RX, counter, Nvswitch throughput rx
887 # context=dcgm.nvswitch.interconnect.nvswitch.throughput family=nvswitch interconnect dimension=nvswitch_throughput_tx
888 # DCGM_FI_DEV_NVSWITCH_THROUGHPUT_TX, counter, Nvswitch throughput tx
889 # context=dcgm.nvswitch.interconnect.nvswitch.voltage family=nvswitch interconnect dimension=nvswitch_voltage_mvolt
890 # DCGM_FI_DEV_NVSWITCH_VOLTAGE_MVOLT, gauge, Nvswitch voltage mvolt
891 # context=dcgm.gpu.inventory.software family=gpu inventory dimension=oem_inforom_ver
892 # DCGM_FI_DEV_OEM_INFOROM_VER, label, Oem inforom ver
893 # context=dcgm.nvlink.interconnect.state family=nvlink interconnect dimension=p2p_nvlink_status
894 # DCGM_FI_DEV_P2P_NVLINK_STATUS, gauge, P2p nvlink status
895 # context=dcgm.gpu.interconnect.pcie.error_rate family=gpu interconnect/pcie dimension=pcie_count_correctable_errors
896 # DCGM_FI_DEV_PCIE_COUNT_CORRECTABLE_ERRORS, counter, Pcie count correctable errors
897 # context=dcgm.gpu.interconnect.pcie.link.generation family=gpu interconnect/pcie dimension=link_gen
898 DCGM_FI_DEV_PCIE_LINK_GEN, gauge, Pcie link gen
899 # context=dcgm.gpu.interconnect.pcie.link.width family=gpu interconnect/pcie dimension=link_width
900 DCGM_FI_DEV_PCIE_LINK_WIDTH, gauge, Pcie link width
901 # context=dcgm.gpu.interconnect.pcie.link.generation family=gpu interconnect/pcie dimension=max_link_gen
902 # DCGM_FI_DEV_PCIE_MAX_LINK_GEN, gauge, Pcie max link gen
903 # context=dcgm.gpu.interconnect.pcie.link.width family=gpu interconnect/pcie dimension=max_link_width
904 # DCGM_FI_DEV_PCIE_MAX_LINK_WIDTH, gauge, Pcie max link width
905 # context=dcgm.gpu.interconnect.pcie.error_rate family=gpu interconnect/pcie dimension=pcie_replay
906 DCGM_FI_DEV_PCIE_REPLAY_COUNTER, counter, Pcie replay counter
907 # context=dcgm.gpu.interconnect.pcie.throughput family=gpu interconnect/pcie dimension=pcie_rx_throughput
908 DCGM_FI_DEV_PCIE_RX_THROUGHPUT, counter, Pcie rx throughput
909 # context=dcgm.gpu.interconnect.pcie.throughput family=gpu interconnect/pcie dimension=pcie_tx_throughput
910 DCGM_FI_DEV_PCIE_TX_THROUGHPUT, counter, Pcie tx throughput
911 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=pci_busid
912 # DCGM_FI_DEV_PCI_BUSID, label, Pci busid
913 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=pci_combined_id
914 # DCGM_FI_DEV_PCI_COMBINED_ID, gauge, Pci combined id
915 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=pci_subsys_id
916 # DCGM_FI_DEV_PCI_SUBSYS_ID, gauge, Pci subsys id
917 # context=dcgm.gpu.state.configuration family=gpu state dimension=persistence_mode
918 DCGM_FI_DEV_PERSISTENCE_MODE, gauge, Persistence mode
919 # context=dcgm.gpu.inventory.platform family=gpu inventory dimension=platform_chassis_serial_number
920 # DCGM_FI_DEV_PLATFORM_CHASSIS_SERIAL_NUMBER, label, Platform chassis serial number
921 # context=dcgm.gpu.inventory.platform family=gpu inventory dimension=platform_chassis_slot_number
922 # DCGM_FI_DEV_PLATFORM_CHASSIS_SLOT_NUMBER, gauge, Platform chassis slot number
923 # context=dcgm.gpu.inventory.platform family=gpu inventory dimension=platform_host_id
924 # DCGM_FI_DEV_PLATFORM_HOST_ID, gauge, Platform host id
925 # context=dcgm.gpu.inventory.platform family=gpu inventory dimension=platform_infiniband_guid
926 # DCGM_FI_DEV_PLATFORM_INFINIBAND_GUID, gauge, Platform infiniband guid
927 # context=dcgm.gpu.inventory.platform family=gpu inventory dimension=platform_module_id
928 # DCGM_FI_DEV_PLATFORM_MODULE_ID, gauge, Platform module id
929 # context=dcgm.gpu.inventory.platform family=gpu inventory dimension=platform_peer_type
930 # DCGM_FI_DEV_PLATFORM_PEER_TYPE, gauge, Platform peer type
931 # context=dcgm.gpu.inventory.platform family=gpu inventory dimension=platform_tray_index
932 # DCGM_FI_DEV_PLATFORM_TRAY_INDEX, gauge, Platform tray index
933 # context=dcgm.gpu.inventory.software family=gpu inventory dimension=power_inforom_ver
934 # DCGM_FI_DEV_POWER_INFOROM_VER, label, Power inforom ver
935 # context=dcgm.gpu.power.usage family=gpu power dimension=power_mgmt_limit
936 DCGM_FI_DEV_POWER_MGMT_LIMIT, gauge, Power mgmt limit
937 # context=dcgm.gpu.power.usage family=gpu power dimension=power_mgmt_limit_def
938 DCGM_FI_DEV_POWER_MGMT_LIMIT_DEF, gauge, Power mgmt limit def
939 # context=dcgm.gpu.power.usage family=gpu power dimension=power_mgmt_limit_max
940 DCGM_FI_DEV_POWER_MGMT_LIMIT_MAX, gauge, Power mgmt limit max
941 # context=dcgm.gpu.power.usage family=gpu power dimension=power_mgmt_limit_min
942 DCGM_FI_DEV_POWER_MGMT_LIMIT_MIN, gauge, Power mgmt limit min
943 # context=dcgm.gpu.power.usage family=gpu power dimension=draw
944 DCGM_FI_DEV_POWER_USAGE, gauge, Power usage
945 # context=dcgm.gpu.power.usage family=gpu power dimension=power_usage_instant
946 DCGM_FI_DEV_POWER_USAGE_INSTANT, gauge, Power usage instant
947 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=power_violation
948 DCGM_FI_DEV_POWER_VIOLATION, counter, Power violation
949 # context=dcgm.gpu.state.performance family=gpu state dimension=pstate
950 DCGM_FI_DEV_PSTATE, gauge, Pstate
951 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_active_preset_profile
952 # DCGM_FI_DEV_PWR_SMOOTHING_ACTIVE_PRESET_PROFILE, gauge, Pwr smoothing active preset profile
953 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_admin_override_percent_tmp_floor
954 # DCGM_FI_DEV_PWR_SMOOTHING_ADMIN_OVERRIDE_PERCENT_TMP_FLOOR, gauge, Pwr smoothing admin override percent tmp floor
955 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_admin_override_ramp_down_hyst_val
956 # DCGM_FI_DEV_PWR_SMOOTHING_ADMIN_OVERRIDE_RAMP_DOWN_HYST_VAL, gauge, Pwr smoothing admin override ramp down hyst val
957 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_admin_override_ramp_down_rate
958 # DCGM_FI_DEV_PWR_SMOOTHING_ADMIN_OVERRIDE_RAMP_DOWN_RATE, gauge, Pwr smoothing admin override ramp down rate
959 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_admin_override_ramp_up_rate
960 # DCGM_FI_DEV_PWR_SMOOTHING_ADMIN_OVERRIDE_RAMP_UP_RATE, gauge, Pwr smoothing admin override ramp up rate
961 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_applied_tmp_ceil
962 # DCGM_FI_DEV_PWR_SMOOTHING_APPLIED_TMP_CEIL, gauge, Pwr smoothing applied tmp ceil
963 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_applied_tmp_floor
964 # DCGM_FI_DEV_PWR_SMOOTHING_APPLIED_TMP_FLOOR, gauge, Pwr smoothing applied tmp floor
965 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_enabled
966 # DCGM_FI_DEV_PWR_SMOOTHING_ENABLED, gauge, Pwr smoothing enabled
967 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_hw_circuitry_percent_lifetime_remaining
968 # DCGM_FI_DEV_PWR_SMOOTHING_HW_CIRCUITRY_PERCENT_LIFETIME_REMAINING, gauge, Pwr smoothing hw circuitry percent lifetime remaining
969 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_imm_ramp_down_enabled
970 # DCGM_FI_DEV_PWR_SMOOTHING_IMM_RAMP_DOWN_ENABLED, gauge, Pwr smoothing imm ramp down enabled
971 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_max_num_preset_profiles
972 # DCGM_FI_DEV_PWR_SMOOTHING_MAX_NUM_PRESET_PROFILES, gauge, Pwr smoothing max num preset profiles
973 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_max_percent_tmp_floor_setting
974 # DCGM_FI_DEV_PWR_SMOOTHING_MAX_PERCENT_TMP_FLOOR_SETTING, gauge, Pwr smoothing max percent tmp floor setting
975 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_min_percent_tmp_floor_setting
976 # DCGM_FI_DEV_PWR_SMOOTHING_MIN_PERCENT_TMP_FLOOR_SETTING, gauge, Pwr smoothing min percent tmp floor setting
977 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_priv_lvl
978 # DCGM_FI_DEV_PWR_SMOOTHING_PRIV_LVL, gauge, Pwr smoothing priv lvl
979 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_profile_percent_tmp_floor
980 # DCGM_FI_DEV_PWR_SMOOTHING_PROFILE_PERCENT_TMP_FLOOR, gauge, Pwr smoothing profile percent tmp floor
981 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_profile_ramp_down_hyst_val
982 # DCGM_FI_DEV_PWR_SMOOTHING_PROFILE_RAMP_DOWN_HYST_VAL, gauge, Pwr smoothing profile ramp down hyst val
983 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_profile_ramp_down_rate
984 # DCGM_FI_DEV_PWR_SMOOTHING_PROFILE_RAMP_DOWN_RATE, gauge, Pwr smoothing profile ramp down rate
985 # context=dcgm.gpu.power.smoothing family=gpu power dimension=pwr_smoothing_profile_ramp_up_rate
986 # DCGM_FI_DEV_PWR_SMOOTHING_PROFILE_RAMP_UP_RATE, gauge, Pwr smoothing profile ramp up rate
987 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=reliability_violation
988 # DCGM_FI_DEV_RELIABILITY_VIOLATION, counter, Reliability violation
989 # context=dcgm.gpu.power.profiles family=gpu power dimension=requested_power_profile_mask
990 # DCGM_FI_DEV_REQUESTED_POWER_PROFILE_MASK, counter, Requested power profile mask
991 # context=dcgm.gpu.memory.page_retirements family=gpu memory dimension=retired_dbe
992 DCGM_FI_DEV_RETIRED_DBE, counter, Retired dbe
993 # context=dcgm.gpu.memory.page_retirements family=gpu memory dimension=retired_pending
994 DCGM_FI_DEV_RETIRED_PENDING, counter, Retired pending
995 # context=dcgm.gpu.memory.page_retirements family=gpu memory dimension=retired_sbe
996 DCGM_FI_DEV_RETIRED_SBE, counter, Retired sbe
997 # context=dcgm.gpu.reliability.row_remap_status family=gpu reliability dimension=row_remap_failure
998 DCGM_FI_DEV_ROW_REMAP_FAILURE, gauge, Row remap failure
999 # context=dcgm.gpu.reliability.row_remap_status family=gpu reliability dimension=row_remap_pending
1000 DCGM_FI_DEV_ROW_REMAP_PENDING, gauge, Row remap pending
1001 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=serial
1002 DCGM_FI_DEV_SERIAL, label, Serial
1003 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=shutdown_temp
1004 DCGM_FI_DEV_SHUTDOWN_TEMP, gauge, Shutdown temp
1005 # context=dcgm.gpu.thermal.temperature family=gpu thermal dimension=slowdown_temp
1006 DCGM_FI_DEV_SLOWDOWN_TEMP, gauge, Slowdown temp
1007 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=sm
1008 DCGM_FI_DEV_SM_CLOCK, gauge, Sm clock
1009 # context=dcgm.gpu.capability.support family=gpu capability dimension=supported_clocks
1010 # DCGM_FI_DEV_SUPPORTED_CLOCKS, gauge, Supported clocks
1011 # context=dcgm.gpu.capability.support family=gpu capability dimension=supported_type_info
1012 # DCGM_FI_DEV_SUPPORTED_TYPE_INFO, label, Supported type info
1013 # context=dcgm.gpu.virtualization.vgpu.type family=gpu virtualization dimension=supported_vgpu_type_ids
1014 # DCGM_FI_DEV_SUPPORTED_VGPU_TYPE_IDS, gauge, Supported vgpu type ids
1015 # context=dcgm.gpu.state.configuration family=gpu state dimension=sync_boost_violation
1016 # DCGM_FI_DEV_SYNC_BOOST_VIOLATION, counter, Sync boost violation
1017 # context=dcgm.gpu.cpu.power family=gpu cpu dimension=sysio_power_util_current
1018 # DCGM_FI_DEV_SYSIO_POWER_UTIL_CURRENT, gauge, Sysio power util current
1019 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=thermal_violation
1020 DCGM_FI_DEV_THERMAL_VIOLATION, counter, Thermal violation
1021 # context=dcgm.gpu.reliability.memory_health family=gpu reliability dimension=threshold_srm
1022 DCGM_FI_DEV_THRESHOLD_SRM, gauge, Threshold srm
1023 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=total_app_clocks_violation
1024 # DCGM_FI_DEV_TOTAL_APP_CLOCKS_VIOLATION, counter, Total app clocks violation
1025 # context=dcgm.gpu.throttle.violations family=gpu throttle dimension=total_base_clocks_violation
1026 # DCGM_FI_DEV_TOTAL_BASE_CLOCKS_VIOLATION, counter, Total base clocks violation
1027 # context=dcgm.gpu.power.energy family=gpu power dimension=total
1028 DCGM_FI_DEV_TOTAL_ENERGY_CONSUMPTION, counter, Total energy consumption
1029 # context=dcgm.gpu.reliability.row_remap_events family=gpu reliability dimension=uncorrectable_remapped_rows
1030 DCGM_FI_DEV_UNCORRECTABLE_REMAPPED_ROWS, counter, Uncorrectable remapped rows
1031 # context=dcgm.gpu.inventory.identity family=gpu inventory dimension=uuid
1032 # DCGM_FI_DEV_UUID, label, Uuid
1033 # context=dcgm.gpu.power.profiles family=gpu power dimension=valid_power_profile_mask
1034 # DCGM_FI_DEV_VALID_POWER_PROFILE_MASK, counter, Valid power profile mask
1035 # context=dcgm.gpu.inventory.software family=gpu inventory dimension=vbios_version
1036 # DCGM_FI_DEV_VBIOS_VERSION, label, Vbios version
1037 # context=dcgm.gpu.virtualization.vgpu.software family=gpu virtualization dimension=vgpu_driver_version
1038 # DCGM_FI_DEV_VGPU_DRIVER_VERSION, label, Vgpu driver version
1039 # context=dcgm.gpu.virtualization.vgpu.sessions family=gpu virtualization dimension=vgpu_enc_sessions_info
1040 # DCGM_FI_DEV_VGPU_ENC_SESSIONS_INFO, gauge, Vgpu enc sessions info
1041 # context=dcgm.gpu.virtualization.vgpu.sessions family=gpu virtualization dimension=vgpu_enc_stats
1042 # DCGM_FI_DEV_VGPU_ENC_STATS, gauge, Vgpu enc stats
1043 # context=dcgm.gpu.virtualization.vgpu.sessions family=gpu virtualization dimension=vgpu_fbc_sessions_info
1044 # DCGM_FI_DEV_VGPU_FBC_SESSIONS_INFO, gauge, Vgpu fbc sessions info
1045 # context=dcgm.gpu.virtualization.vgpu.sessions family=gpu virtualization dimension=vgpu_fbc_stats
1046 # DCGM_FI_DEV_VGPU_FBC_STATS, gauge, Vgpu fbc stats
1047 # context=dcgm.gpu.virtualization.vgpu.frame_rate family=gpu virtualization dimension=vgpu_frame_rate_limit
1048 # DCGM_FI_DEV_VGPU_FRAME_RATE_LIMIT, gauge, Vgpu frame rate limit
1049 # context=dcgm.gpu.virtualization.vgpu.instance family=gpu virtualization dimension=vgpu_instance_ids
1050 # DCGM_FI_DEV_VGPU_INSTANCE_IDS, gauge, Vgpu instance ids
1051 # context=dcgm.gpu.virtualization.vgpu.license family=gpu virtualization dimension=vgpu_instance_license_state
1052 # DCGM_FI_DEV_VGPU_INSTANCE_LICENSE_STATE, gauge, Vgpu instance license state
1053 # context=dcgm.gpu.virtualization.vgpu.license family=gpu virtualization dimension=vgpu_license_status
1054 DCGM_FI_DEV_VGPU_LICENSE_STATUS, gauge, Vgpu license status
1055 # context=dcgm.gpu.virtualization.vgpu.memory family=gpu virtualization dimension=vgpu_memory_usage
1056 DCGM_FI_DEV_VGPU_MEMORY_USAGE, gauge, Vgpu memory usage
1057 # context=dcgm.gpu.virtualization.vgpu.instance family=gpu virtualization dimension=vgpu_pci_id
1058 # DCGM_FI_DEV_VGPU_PCI_ID, gauge, Vgpu pci id
1059 # context=dcgm.gpu.virtualization.vgpu.utilization family=gpu virtualization dimension=vgpu_per_process_utilization
1060 # DCGM_FI_DEV_VGPU_PER_PROCESS_UTILIZATION, gauge, Vgpu per process utilization
1061 # context=dcgm.gpu.virtualization.vgpu.type family=gpu virtualization dimension=vgpu_type
1062 # DCGM_FI_DEV_VGPU_TYPE, gauge, Vgpu type
1063 # context=dcgm.gpu.virtualization.vgpu.type family=gpu virtualization dimension=vgpu_type_class
1064 # DCGM_FI_DEV_VGPU_TYPE_CLASS, gauge, Vgpu type class
1065 # context=dcgm.gpu.virtualization.vgpu.type family=gpu virtualization dimension=vgpu_type_info
1066 # DCGM_FI_DEV_VGPU_TYPE_INFO, label, Vgpu type info
1067 # context=dcgm.gpu.virtualization.vgpu.license family=gpu virtualization dimension=vgpu_type_license
1068 # DCGM_FI_DEV_VGPU_TYPE_LICENSE, gauge, Vgpu type license
1069 # context=dcgm.gpu.virtualization.vgpu.type family=gpu virtualization dimension=vgpu_type_name
1070 # DCGM_FI_DEV_VGPU_TYPE_NAME, label, Vgpu type name
1071 # context=dcgm.gpu.virtualization.vgpu.utilization family=gpu virtualization dimension=vgpu_utilizations
1072 # DCGM_FI_DEV_VGPU_UTILIZATIONS, gauge, Vgpu utilizations
1073 # context=dcgm.gpu.virtualization.vgpu.instance family=gpu virtualization dimension=vgpu_uuid
1074 # DCGM_FI_DEV_VGPU_UUID, label, Vgpu uuid
1075 # context=dcgm.gpu.virtualization.vgpu.vm family=gpu virtualization dimension=vgpu_vm_gpu_instance_id
1076 # DCGM_FI_DEV_VGPU_VM_GPU_INSTANCE_ID, gauge, Vgpu vm gpu instance id
1077 # context=dcgm.gpu.virtualization.vgpu.vm family=gpu virtualization dimension=vgpu_vm_id
1078 # DCGM_FI_DEV_VGPU_VM_ID, label, Vgpu vm id
1079 # context=dcgm.gpu.virtualization.vgpu.vm family=gpu virtualization dimension=vgpu_vm_name
1080 # DCGM_FI_DEV_VGPU_VM_NAME, label, Vgpu vm name
1081 # context=dcgm.gpu.clock.frequency family=gpu clock dimension=video_clock
1082 DCGM_FI_DEV_VIDEO_CLOCK, gauge, Video clock
1083 # context=dcgm.gpu.state.virtualization family=gpu state dimension=virtual_mode
1084 DCGM_FI_DEV_VIRTUAL_MODE, gauge, Virtual mode
1085 # context=dcgm.gpu.reliability.xid family=gpu reliability dimension=xid
1086 DCGM_FI_DEV_XID_ERRORS, counter, Xid errors
1087 # context=dcgm.exporter.inventory.software family=exporter inventory dimension=driver_version
1088 DCGM_FI_DRIVER_VERSION, label, Driver version
1089 # context=dcgm.nvswitch.internal.boundary family=nvswitch internal dimension=first_nvswitch_field_id
1090 # DCGM_FI_FIRST_NVSWITCH_FIELD_ID, gauge, First nvswitch field id
1091 # context=dcgm.gpu.internal.boundary family=gpu internal dimension=first_vgpu_field_id
1092 # DCGM_FI_FIRST_VGPU_FIELD_ID, gauge, First vgpu field id
1093 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=gpu_topology_affinity
1094 # DCGM_FI_GPU_TOPOLOGY_AFFINITY, gauge, Gpu topology affinity
1095 # context=dcgm.nvlink.interconnect.state family=nvlink interconnect dimension=gpu_topology_nvlink
1096 # DCGM_FI_GPU_TOPOLOGY_NVLINK, gauge, Gpu topology nvlink
1097 # context=dcgm.gpu.topology.affinity family=gpu topology dimension=gpu_topology_pci
1098 # DCGM_FI_GPU_TOPOLOGY_PCI, gauge, Gpu topology pci
1099 # context=dcgm.gpu.health.status family=gpu health dimension=imex_daemon_status
1100 # DCGM_FI_IMEX_DAEMON_STATUS, gauge, Imex daemon status
1101 # context=dcgm.gpu.health.status family=gpu health dimension=imex_domain_status
1102 # DCGM_FI_IMEX_DOMAIN_STATUS, gauge, Imex domain status
1103 # context=dcgm.gpu.internal.boundary family=gpu internal dimension=internal_fields_0_end
1104 # DCGM_FI_INTERNAL_FIELDS_0_END, gauge, Internal fields 0 end
1105 # context=dcgm.gpu.internal.boundary family=gpu internal dimension=internal_fields_0_start
1106 # DCGM_FI_INTERNAL_FIELDS_0_START, gauge, Internal fields 0 start
1107 # context=dcgm.nvswitch.internal.boundary family=nvswitch internal dimension=last_nvswitch_field_id
1108 # DCGM_FI_LAST_NVSWITCH_FIELD_ID, gauge, Last nvswitch field id
1109 # context=dcgm.gpu.internal.boundary family=gpu internal dimension=last_vgpu_field_id
1110 # DCGM_FI_LAST_VGPU_FIELD_ID, gauge, Last vgpu field id
1111 # context=dcgm.exporter.inventory.software family=exporter inventory dimension=nvml_version
1112 # DCGM_FI_NVML_VERSION, label, Nvml version
1113 # context=dcgm.gpu.inventory.software family=gpu inventory dimension=process_name
1114 # DCGM_FI_PROCESS_NAME, label, Process name
1115 # context=dcgm.gpu.interconnect.throughput family=gpu interconnect/overview dimension=c2c_rx_all_bytes
1116 # DCGM_FI_PROF_C2C_RX_ALL_BYTES, gauge, C2c rx all bytes
1117 # context=dcgm.gpu.interconnect.throughput family=gpu interconnect/overview dimension=c2c_rx_data_bytes
1118 # DCGM_FI_PROF_C2C_RX_DATA_BYTES, gauge, C2c rx data bytes
1119 # context=dcgm.gpu.interconnect.throughput family=gpu interconnect/overview dimension=c2c_tx_all_bytes
1120 # DCGM_FI_PROF_C2C_TX_ALL_BYTES, gauge, C2c tx all bytes
1121 # context=dcgm.gpu.interconnect.throughput family=gpu interconnect/overview dimension=c2c_tx_data_bytes
1122 # DCGM_FI_PROF_C2C_TX_DATA_BYTES, gauge, C2c tx data bytes
1123 # context=dcgm.gpu.compute.activity family=gpu compute dimension=dram
1124 DCGM_FI_PROF_DRAM_ACTIVE, gauge, Dram active
1125 # context=dcgm.gpu.compute.activity family=gpu compute dimension=graphics_engine_active
1126 DCGM_FI_PROF_GR_ENGINE_ACTIVE, gauge, Gr engine active
1127 # context=dcgm.gpu.compute.cache.activity family=gpu compute dimension=hostmem_cache_hit
1128 DCGM_FI_PROF_HOSTMEM_CACHE_HIT, gauge, Hostmem cache hit
1129 # context=dcgm.gpu.compute.cache.activity family=gpu compute dimension=hostmem_cache_miss
1130 DCGM_FI_PROF_HOSTMEM_CACHE_MISS, gauge, Hostmem cache miss
1131 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec0_active
1132 # DCGM_FI_PROF_NVDEC0_ACTIVE, gauge, Nvdec0 active
1133 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec1_active
1134 # DCGM_FI_PROF_NVDEC1_ACTIVE, gauge, Nvdec1 active
1135 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec2_active
1136 # DCGM_FI_PROF_NVDEC2_ACTIVE, gauge, Nvdec2 active
1137 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec3_active
1138 # DCGM_FI_PROF_NVDEC3_ACTIVE, gauge, Nvdec3 active
1139 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec4_active
1140 # DCGM_FI_PROF_NVDEC4_ACTIVE, gauge, Nvdec4 active
1141 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec5_active
1142 # DCGM_FI_PROF_NVDEC5_ACTIVE, gauge, Nvdec5 active
1143 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec6_active
1144 # DCGM_FI_PROF_NVDEC6_ACTIVE, gauge, Nvdec6 active
1145 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvdec7_active
1146 # DCGM_FI_PROF_NVDEC7_ACTIVE, gauge, Nvdec7 active
1147 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg0_active
1148 # DCGM_FI_PROF_NVJPG0_ACTIVE, gauge, Nvjpg0 active
1149 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg1_active
1150 # DCGM_FI_PROF_NVJPG1_ACTIVE, gauge, Nvjpg1 active
1151 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg2_active
1152 # DCGM_FI_PROF_NVJPG2_ACTIVE, gauge, Nvjpg2 active
1153 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg3_active
1154 # DCGM_FI_PROF_NVJPG3_ACTIVE, gauge, Nvjpg3 active
1155 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg4_active
1156 # DCGM_FI_PROF_NVJPG4_ACTIVE, gauge, Nvjpg4 active
1157 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg5_active
1158 # DCGM_FI_PROF_NVJPG5_ACTIVE, gauge, Nvjpg5 active
1159 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg6_active
1160 # DCGM_FI_PROF_NVJPG6_ACTIVE, gauge, Nvjpg6 active
1161 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvjpg7_active
1162 # DCGM_FI_PROF_NVJPG7_ACTIVE, gauge, Nvjpg7 active
1163 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l0_rx
1164 # DCGM_FI_PROF_NVLINK_L0_RX_BYTES, gauge, Nvlink l0 rx bytes
1165 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l0_tx
1166 # DCGM_FI_PROF_NVLINK_L0_TX_BYTES, gauge, Nvlink l0 tx bytes
1167 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l10_rx
1168 # DCGM_FI_PROF_NVLINK_L10_RX_BYTES, gauge, Nvlink l10 rx bytes
1169 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l10_tx
1170 # DCGM_FI_PROF_NVLINK_L10_TX_BYTES, gauge, Nvlink l10 tx bytes
1171 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l11_rx
1172 # DCGM_FI_PROF_NVLINK_L11_RX_BYTES, gauge, Nvlink l11 rx bytes
1173 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l11_tx
1174 # DCGM_FI_PROF_NVLINK_L11_TX_BYTES, gauge, Nvlink l11 tx bytes
1175 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l12_rx
1176 # DCGM_FI_PROF_NVLINK_L12_RX_BYTES, gauge, Nvlink l12 rx bytes
1177 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l12_tx
1178 # DCGM_FI_PROF_NVLINK_L12_TX_BYTES, gauge, Nvlink l12 tx bytes
1179 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l13_rx
1180 # DCGM_FI_PROF_NVLINK_L13_RX_BYTES, gauge, Nvlink l13 rx bytes
1181 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l13_tx
1182 # DCGM_FI_PROF_NVLINK_L13_TX_BYTES, gauge, Nvlink l13 tx bytes
1183 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l14_rx
1184 # DCGM_FI_PROF_NVLINK_L14_RX_BYTES, gauge, Nvlink l14 rx bytes
1185 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l14_tx
1186 # DCGM_FI_PROF_NVLINK_L14_TX_BYTES, gauge, Nvlink l14 tx bytes
1187 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l15_rx
1188 # DCGM_FI_PROF_NVLINK_L15_RX_BYTES, gauge, Nvlink l15 rx bytes
1189 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l15_tx
1190 # DCGM_FI_PROF_NVLINK_L15_TX_BYTES, gauge, Nvlink l15 tx bytes
1191 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l16_rx
1192 # DCGM_FI_PROF_NVLINK_L16_RX_BYTES, gauge, Nvlink l16 rx bytes
1193 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l16_tx
1194 # DCGM_FI_PROF_NVLINK_L16_TX_BYTES, gauge, Nvlink l16 tx bytes
1195 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l17_rx
1196 # DCGM_FI_PROF_NVLINK_L17_RX_BYTES, gauge, Nvlink l17 rx bytes
1197 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l17_tx
1198 # DCGM_FI_PROF_NVLINK_L17_TX_BYTES, gauge, Nvlink l17 tx bytes
1199 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l1_rx
1200 # DCGM_FI_PROF_NVLINK_L1_RX_BYTES, gauge, Nvlink l1 rx bytes
1201 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l1_tx
1202 # DCGM_FI_PROF_NVLINK_L1_TX_BYTES, gauge, Nvlink l1 tx bytes
1203 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l2_rx
1204 # DCGM_FI_PROF_NVLINK_L2_RX_BYTES, gauge, Nvlink l2 rx bytes
1205 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l2_tx
1206 # DCGM_FI_PROF_NVLINK_L2_TX_BYTES, gauge, Nvlink l2 tx bytes
1207 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l3_rx
1208 # DCGM_FI_PROF_NVLINK_L3_RX_BYTES, gauge, Nvlink l3 rx bytes
1209 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l3_tx
1210 # DCGM_FI_PROF_NVLINK_L3_TX_BYTES, gauge, Nvlink l3 tx bytes
1211 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l4_rx
1212 # DCGM_FI_PROF_NVLINK_L4_RX_BYTES, gauge, Nvlink l4 rx bytes
1213 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l4_tx
1214 # DCGM_FI_PROF_NVLINK_L4_TX_BYTES, gauge, Nvlink l4 tx bytes
1215 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l5_rx
1216 # DCGM_FI_PROF_NVLINK_L5_RX_BYTES, gauge, Nvlink l5 rx bytes
1217 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l5_tx
1218 # DCGM_FI_PROF_NVLINK_L5_TX_BYTES, gauge, Nvlink l5 tx bytes
1219 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l6_rx
1220 # DCGM_FI_PROF_NVLINK_L6_RX_BYTES, gauge, Nvlink l6 rx bytes
1221 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l6_tx
1222 # DCGM_FI_PROF_NVLINK_L6_TX_BYTES, gauge, Nvlink l6 tx bytes
1223 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l7_rx
1224 # DCGM_FI_PROF_NVLINK_L7_RX_BYTES, gauge, Nvlink l7 rx bytes
1225 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l7_tx
1226 # DCGM_FI_PROF_NVLINK_L7_TX_BYTES, gauge, Nvlink l7 tx bytes
1227 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l8_rx
1228 # DCGM_FI_PROF_NVLINK_L8_RX_BYTES, gauge, Nvlink l8 rx bytes
1229 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l8_tx
1230 # DCGM_FI_PROF_NVLINK_L8_TX_BYTES, gauge, Nvlink l8 tx bytes
1231 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l9_rx
1232 # DCGM_FI_PROF_NVLINK_L9_RX_BYTES, gauge, Nvlink l9 rx bytes
1233 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_l9_tx
1234 # DCGM_FI_PROF_NVLINK_L9_TX_BYTES, gauge, Nvlink l9 tx bytes
1235 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_rx
1236 DCGM_FI_PROF_NVLINK_RX_BYTES, gauge, Nvlink rx bytes
1237 # context=dcgm.nvlink.interconnect.throughput family=nvlink interconnect dimension=nvlink_tx
1238 DCGM_FI_PROF_NVLINK_TX_BYTES, gauge, Nvlink tx bytes
1239 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvofa0_active
1240 # DCGM_FI_PROF_NVOFA0_ACTIVE, gauge, Nvofa0 active
1241 # context=dcgm.gpu.compute.media.activity family=gpu compute dimension=nvofa1_active
1242 # DCGM_FI_PROF_NVOFA1_ACTIVE, gauge, Nvofa1 active
1243 # context=dcgm.gpu.interconnect.pcie.throughput family=gpu interconnect/pcie dimension=pcie_rx
1244 DCGM_FI_PROF_PCIE_RX_BYTES, gauge, Pcie rx bytes
1245 # context=dcgm.gpu.interconnect.pcie.throughput family=gpu interconnect/pcie dimension=pcie_tx
1246 DCGM_FI_PROF_PCIE_TX_BYTES, gauge, Pcie tx bytes
1247 # context=dcgm.gpu.compute.cache.activity family=gpu compute dimension=peermem_cache_hit
1248 DCGM_FI_PROF_PEERMEM_CACHE_HIT, gauge, Peermem cache hit
1249 # context=dcgm.gpu.compute.cache.activity family=gpu compute dimension=peermem_cache_miss
1250 DCGM_FI_PROF_PEERMEM_CACHE_MISS, gauge, Peermem cache miss
1251 # context=dcgm.gpu.compute.activity family=gpu compute dimension=fp16
1252 DCGM_FI_PROF_PIPE_FP16_ACTIVE, gauge, Pipe fp16 active
1253 # context=dcgm.gpu.compute.activity family=gpu compute dimension=fp32
1254 DCGM_FI_PROF_PIPE_FP32_ACTIVE, gauge, Pipe fp32 active
1255 # context=dcgm.gpu.compute.activity family=gpu compute dimension=fp64
1256 DCGM_FI_PROF_PIPE_FP64_ACTIVE, gauge, Pipe fp64 active
1257 # context=dcgm.gpu.compute.activity family=gpu compute dimension=integer
1258 DCGM_FI_PROF_PIPE_INT_ACTIVE, gauge, Pipe int active
1259 # context=dcgm.gpu.compute.activity family=gpu compute dimension=tensor
1260 DCGM_FI_PROF_PIPE_TENSOR_ACTIVE, gauge, Pipe tensor active
1261 # context=dcgm.gpu.compute.tensor.activity family=gpu compute dimension=tensor_dfma
1262 DCGM_FI_PROF_PIPE_TENSOR_DFMA_ACTIVE, gauge, Pipe tensor dfma active
1263 # context=dcgm.gpu.compute.tensor.activity family=gpu compute dimension=tensor_hmma
1264 DCGM_FI_PROF_PIPE_TENSOR_HMMA_ACTIVE, gauge, Pipe tensor hmma active
1265 # context=dcgm.gpu.compute.tensor.activity family=gpu compute dimension=tensor_imma
1266 DCGM_FI_PROF_PIPE_TENSOR_IMMA_ACTIVE, gauge, Pipe tensor imma active
1267 # context=dcgm.gpu.compute.activity family=gpu compute dimension=sm_active
1268 DCGM_FI_PROF_SM_ACTIVE, gauge, Sm active
1269 # context=dcgm.gpu.compute.activity family=gpu compute dimension=sm_occupancy
1270 DCGM_FI_PROF_SM_OCCUPANCY, gauge, Sm occupancy
1271 # context=dcgm.gpu.state.configuration family=gpu state dimension=sync_boost
1272 # DCGM_FI_SYNC_BOOST, gauge, Sync boost