Coverage for cuda/core/_utils/runtime_cuda_error_explanations

2# SPDX-License-Identifier: Apache-2.0

4# CUDA Toolkit v13.1.1

5_FALLBACK_EXPLANATIONS = {

6 0: (

7 "The API call returned with no errors. In the case of query calls, this"

8 " also means that the operation being queried is complete (see"

9 " ::cudaEventQuery() and ::cudaStreamQuery())."

10 ),

11 1: (

12 "This indicates that one or more of the parameters passed to the API call"

13 " is not within an acceptable range of values."

14 ),

15 2: (

16 "The API call failed because it was unable to allocate enough memory or"

17 " other resources to perform the requested operation."

18 ),

19 3: ("The API call failed because the CUDA driver and runtime could not be initialized."),

20 4: (

21 "This indicates that a CUDA Runtime API call cannot be executed because"

22 " it is being called during process shut down, at a point in time after"

23 " CUDA driver has been unloaded."

24 ),

25 5: (

26 "This indicates profiler is not initialized for this run. This can"

27 " happen when the application is running with external profiling tools"

28 " like visual profiler."

29 ),

30 6: (

31 "This error return is deprecated as of CUDA 5.0. It is no longer an error"

32 " to attempt to enable/disable the profiling via ::cudaProfilerStart or"

33 " ::cudaProfilerStop without initialization."

34 ),

35 7: (

36 "This error return is deprecated as of CUDA 5.0. It is no longer an error"

37 " to call cudaProfilerStart() when profiling is already enabled."

38 ),

39 8: (

40 "This error return is deprecated as of CUDA 5.0. It is no longer an error"

41 " to call cudaProfilerStop() when profiling is already disabled."

42 ),

43 9: (

44 "This indicates that a kernel launch is requesting resources that can"

45 " never be satisfied by the current device. Requesting more shared memory"

46 " per block than the device supports will trigger this error, as will"

47 " requesting too many threads or blocks. See ::cudaDeviceProp for more"

48 " device limitations."

49 ),

50 12: (

51 "This indicates that one or more of the pitch-related parameters passed"

52 " to the API call is not within the acceptable range for pitch."

53 ),

54 13: ("This indicates that the symbol name/identifier passed to the API call is not a valid name or identifier."),

55 16: (

56 "This indicates that at least one host pointer passed to the API call is"

57 " not a valid host pointer."

58 " This error return is deprecated as of CUDA 10.1."

59 ),

60 17: (

61 "This indicates that at least one device pointer passed to the API call is"

62 " not a valid device pointer."

63 " This error return is deprecated as of CUDA 10.1."

64 ),

65 18: ("This indicates that the texture passed to the API call is not a valid texture."),

66 19: (

67 "This indicates that the texture binding is not valid. This occurs if you"

68 " call ::cudaGetTextureAlignmentOffset() with an unbound texture."

69 ),

70 20: (

71 "This indicates that the channel descriptor passed to the API call is not"

72 " valid. This occurs if the format is not one of the formats specified by"

73 " ::cudaChannelFormatKind, or if one of the dimensions is invalid."

74 ),

75 21: (

76 "This indicates that the direction of the memcpy passed to the API call is"

77 " not one of the types specified by ::cudaMemcpyKind."

78 ),

79 22: (

80 "This indicated that the user has taken the address of a constant variable,"

81 " which was forbidden up until the CUDA 3.1 release."

82 " This error return is deprecated as of CUDA 3.1. Variables in constant"

83 " memory may now have their address taken by the runtime via"

84 " ::cudaGetSymbolAddress()."

85 ),

86 23: (

87 "This indicated that a texture fetch was not able to be performed."

88 " This was previously used for device emulation of texture operations."

89 " This error return is deprecated as of CUDA 3.1. Device emulation mode was"

90 " removed with the CUDA 3.1 release."

91 ),

92 24: (

93 "This indicated that a texture was not bound for access."

94 " This was previously used for device emulation of texture operations."

95 " This error return is deprecated as of CUDA 3.1. Device emulation mode was"

96 " removed with the CUDA 3.1 release."

97 ),

98 25: (

99 "This indicated that a synchronization operation had failed."

100 " This was previously used for some device emulation functions."

101 " This error return is deprecated as of CUDA 3.1. Device emulation mode was"

102 " removed with the CUDA 3.1 release."

103 ),

104 26: (

105 "This indicates that a non-float texture was being accessed with linear"

106 " filtering. This is not supported by CUDA."

107 ),

108 27: (

109 "This indicates that an attempt was made to read an unsupported data type as a"

110 " normalized float. This is not supported by CUDA."

111 ),

112 28: (

113 "Mixing of device and device emulation code was not allowed."

114 " This error return is deprecated as of CUDA 3.1. Device emulation mode was"

115 " removed with the CUDA 3.1 release."

116 ),

117 31: (

118 "This indicates that the API call is not yet implemented. Production"

119 " releases of CUDA will never return this error."

120 " This error return is deprecated as of CUDA 4.1."

121 ),

122 32: (

123 "This indicated that an emulated device pointer exceeded the 32-bit address"

124 " range."

125 " This error return is deprecated as of CUDA 3.1. Device emulation mode was"

126 " removed with the CUDA 3.1 release."

127 ),

128 34: (

129 "This indicates that the CUDA driver that the application has loaded is a"

130 " stub library. Applications that run with the stub rather than a real"

131 " driver loaded will result in CUDA API returning this error."

132 ),

133 35: (

134 "This indicates that the installed NVIDIA CUDA driver is older than the"

135 " CUDA runtime library. This is not a supported configuration. Users should"

136 " install an updated NVIDIA display driver to allow the application to run."

137 ),

138 36: (

139 "This indicates that the API call requires a newer CUDA driver than the one"

140 " currently installed. Users should install an updated NVIDIA CUDA driver"

141 " to allow the API call to succeed."

142 ),

143 37: ("This indicates that the surface passed to the API call is not a valid surface."),

144 43: (

145 "This indicates that multiple global or constant variables (across separate"

146 " CUDA source files in the application) share the same string name."

147 ),

148 44: (

149 "This indicates that multiple textures (across separate CUDA source"

150 " files in the application) share the same string name."

151 ),

152 45: (

153 "This indicates that multiple surfaces (across separate CUDA source"

154 " files in the application) share the same string name."

155 ),

156 46: (

157 "This indicates that all CUDA devices are busy or unavailable at the current"

158 " time. Devices are often busy/unavailable due to use of"

159 " ::cudaComputeModeProhibited, ::cudaComputeModeExclusiveProcess, or when long"

160 " running CUDA kernels have filled up the GPU and are blocking new work"

161 " from starting. They can also be unavailable due to memory constraints"

162 " on a device that already has active CUDA work being performed."

163 ),

164 49: (

165 "This indicates that the current context is not compatible with this"

166 " the CUDA Runtime. This can only occur if you are using CUDA"

167 " Runtime/Driver interoperability and have created an existing Driver"

168 " context using the driver API. The Driver context may be incompatible"

169 " either because the Driver context was created using an older version"

170 " of the API, because the Runtime API call expects a primary driver"

171 " context and the Driver context is not primary, or because the Driver"

172 ' context has been destroyed. Please see CUDART_DRIVER "Interactions'

173 ' with the CUDA Driver API" for more information.'

174 ),

175 52: (

176 "The device function being invoked (usually via ::cudaLaunchKernel()) was not"

177 " previously configured via the ::cudaConfigureCall() function."

178 ),

179 53: (

180 "This indicated that a previous kernel launch failed. This was previously"

181 " used for device emulation of kernel launches."

182 " This error return is deprecated as of CUDA 3.1. Device emulation mode was"

183 " removed with the CUDA 3.1 release."

184 ),

185 65: (

186 "This error indicates that a device runtime grid launch did not occur"

187 " because the depth of the child grid would exceed the maximum supported"

188 " number of nested grid launches."

189 ),

190 66: (

191 "This error indicates that a grid launch did not occur because the kernel"

192 " uses file-scoped textures which are unsupported by the device runtime."

193 " Kernels launched via the device runtime only support textures created with"

194 " the Texture Object API's."

195 ),

196 67: (

197 "This error indicates that a grid launch did not occur because the kernel"

198 " uses file-scoped surfaces which are unsupported by the device runtime."

199 " Kernels launched via the device runtime only support surfaces created with"

200 " the Surface Object API's."

201 ),

202 68: (

203 "This error indicates that a call to ::cudaDeviceSynchronize made from"

204 " the device runtime failed because the call was made at grid depth greater"

205 " than than either the default (2 levels of grids) or user specified device"

206 " limit ::cudaLimitDevRuntimeSyncDepth. To be able to synchronize on"

207 " launched grids at a greater depth successfully, the maximum nested"

208 " depth at which ::cudaDeviceSynchronize will be called must be specified"

209 " with the ::cudaLimitDevRuntimeSyncDepth limit to the ::cudaDeviceSetLimit"

210 " api before the host-side launch of a kernel using the device runtime."

211 " Keep in mind that additional levels of sync depth require the runtime"

212 " to reserve large amounts of device memory that cannot be used for"

213 " user allocations. Note that ::cudaDeviceSynchronize made from device"

214 " runtime is only supported on devices of compute capability < 9.0."

215 ),

216 69: (

217 "This error indicates that a device runtime grid launch failed because"

218 " the launch would exceed the limit ::cudaLimitDevRuntimePendingLaunchCount."

219 " For this launch to proceed successfully, ::cudaDeviceSetLimit must be"

220 " called to set the ::cudaLimitDevRuntimePendingLaunchCount to be higher"

221 " than the upper bound of outstanding launches that can be issued to the"

222 " device runtime. Keep in mind that raising the limit of pending device"

223 " runtime launches will require the runtime to reserve device memory that"

224 " cannot be used for user allocations."

225 ),

226 98: ("The requested device function does not exist or is not compiled for the proper device architecture."),

227 100: ("This indicates that no CUDA-capable devices were detected by the installed CUDA driver."),

228 101: (

229 "This indicates that the device ordinal supplied by the user does not"

230 " correspond to a valid CUDA device or that the action requested is"

231 " invalid for the specified device."

232 ),

233 102: "This indicates that the device doesn't have a valid Grid License.",

234 103: (

235 "By default, the CUDA runtime may perform a minimal set of self-tests,"

236 " as well as CUDA driver tests, to establish the validity of both."

237 " Introduced in CUDA 11.2, this error return indicates that at least one"

238 " of these tests has failed and the validity of either the runtime"

239 " or the driver could not be established."

240 ),

241 127: "This indicates an internal startup failure in the CUDA runtime.",

242 200: "This indicates that the device kernel image is invalid.",

243 201: (

244 "This most frequently indicates that there is no context bound to the"

245 " current thread. This can also be returned if the context passed to an"

246 " API call is not a valid handle (such as a context that has had"

247 " ::cuCtxDestroy() invoked on it). This can also be returned if a user"

248 " mixes different API versions (i.e. 3010 context with 3020 API calls)."

249 " See ::cuCtxGetApiVersion() for more details."

250 ),

251 205: "This indicates that the buffer object could not be mapped.",

252 206: "This indicates that the buffer object could not be unmapped.",

253 207: ("This indicates that the specified array is currently mapped and thus cannot be destroyed."),

254 208: "This indicates that the resource is already mapped.",

255 209: (

256 "This indicates that there is no kernel image available that is suitable"

257 " for the device. This can occur when a user specifies code generation"

258 " options for a particular CUDA source file that do not include the"

259 " corresponding device configuration."

260 ),

261 210: "This indicates that a resource has already been acquired.",

262 211: "This indicates that a resource is not mapped.",

263 212: ("This indicates that a mapped resource is not available for access as an array."),

264 213: ("This indicates that a mapped resource is not available for access as a pointer."),

265 214: ("This indicates that an uncorrectable ECC error was detected during execution."),

266 215: ("This indicates that the ::cudaLimit passed to the API call is not supported by the active device."),

267 216: (

268 "This indicates that a call tried to access an exclusive-thread device that"

269 " is already in use by a different thread."

270 ),

271 217: ("This error indicates that P2P access is not supported across the given devices."),

272 218: (

273 "A PTX compilation failed. The runtime may fall back to compiling PTX if"

274 " an application does not contain a suitable binary for the current device."

275 ),

276 219: "This indicates an error with the OpenGL or DirectX context.",

277 220: ("This indicates that an uncorrectable NVLink error was detected during the execution."),

278 221: (

279 "This indicates that the PTX JIT compiler library was not found. The JIT Compiler"

280 " library is used for PTX compilation. The runtime may fall back to compiling PTX"

281 " if an application does not contain a suitable binary for the current device."

282 ),

283 222: (

284 "This indicates that the provided PTX was compiled with an unsupported toolchain."

285 " The most common reason for this, is the PTX was generated by a compiler newer"

286 " than what is supported by the CUDA driver and PTX JIT compiler."

287 ),

288 223: (

289 "This indicates that the JIT compilation was disabled. The JIT compilation compiles"

290 " PTX. The runtime may fall back to compiling PTX if an application does not contain"

291 " a suitable binary for the current device."

292 ),

293 224: "This indicates that the provided execution affinity is not supported by the device.",

294 225: (

295 "This indicates that the code to be compiled by the PTX JIT contains unsupported call to cudaDeviceSynchronize."

296 ),

297 226: (

298 "This indicates that an exception occurred on the device that is now"

299 " contained by the GPU's error containment capability. Common causes are -"

300 " a. Certain types of invalid accesses of peer GPU memory over nvlink"

301 " b. Certain classes of hardware errors"

302 " This leaves the process in an inconsistent state and any further CUDA"

303 " work will return the same error. To continue using CUDA, the process must"

304 " be terminated and relaunched."

305 ),

306 300: "This indicates that the device kernel source is invalid.",

307 301: "This indicates that the file specified was not found.",

308 302: "This indicates that a link to a shared object failed to resolve.",

309 303: "This indicates that initialization of a shared object failed.",

310 304: "This error indicates that an OS call failed.",

311 400: (

312 "This indicates that a resource handle passed to the API call was not"

313 " valid. Resource handles are opaque types like ::cudaStream_t and"

314 " ::cudaEvent_t."

315 ),

316 401: (

317 "This indicates that a resource required by the API call is not in a"

318 " valid state to perform the requested operation."

319 ),

320 402: (

321 "This indicates an attempt was made to introspect an object in a way that"

322 " would discard semantically important information. This is either due to"

323 " the object using funtionality newer than the API version used to"

324 " introspect it or omission of optional return arguments."

325 ),

326 500: (

327 "This indicates that a named symbol was not found. Examples of symbols"

328 " are global/constant variable names, driver function names, texture names,"

329 " and surface names."

330 ),

331 600: (

332 "This indicates that asynchronous operations issued previously have not"

333 " completed yet. This result is not actually an error, but must be indicated"

334 " differently than ::cudaSuccess (which indicates completion). Calls that"

335 " may return this value include ::cudaEventQuery() and ::cudaStreamQuery()."

336 ),

337 700: (

338 "The device encountered a load or store instruction on an invalid memory address."

339 " This leaves the process in an inconsistent state and any further CUDA work"

340 " will return the same error. To continue using CUDA, the process must be terminated"

341 " and relaunched."

342 ),

343 701: (

344 "This indicates that a launch did not occur because it did not have"

345 " appropriate resources. Although this error is similar to"

346 " ::cudaErrorInvalidConfiguration, this error usually indicates that the"

347 " user has attempted to pass too many arguments to the device kernel, or the"

348 " kernel launch specifies too many threads for the kernel's register count."

349 ),

350 702: (

351 "This indicates that the device kernel took too long to execute. This can"

352 " only occur if timeouts are enabled - see the device attribute"

353 ' ::cudaDeviceAttr::cudaDevAttrKernelExecTimeout "cudaDevAttrKernelExecTimeout"'

354 " for more information."

355 " This leaves the process in an inconsistent state and any further CUDA work"

356 " will return the same error. To continue using CUDA, the process must be terminated"

357 " and relaunched."

358 ),

359 703: ("This error indicates a kernel launch that uses an incompatible texturing mode."),

360 704: (

361 "This error indicates that a call to ::cudaDeviceEnablePeerAccess() is"

362 " trying to re-enable peer addressing on from a context which has already"

363 " had peer addressing enabled."

364 ),

365 705: (

366 "This error indicates that ::cudaDeviceDisablePeerAccess() is trying to"

367 " disable peer addressing which has not been enabled yet via"

368 " ::cudaDeviceEnablePeerAccess()."

369 ),

370 708: (

371 "This indicates that the user has called ::cudaSetValidDevices(),"

372 " ::cudaSetDeviceFlags(), ::cudaD3D9SetDirect3DDevice(),"

373 " ::cudaD3D10SetDirect3DDevice, ::cudaD3D11SetDirect3DDevice(), or"

374 " ::cudaVDPAUSetVDPAUDevice() after initializing the CUDA runtime by"

375 " calling non-device management operations (allocating memory and"

376 " launching kernels are examples of non-device management operations)."

377 " This error can also be returned if using runtime/driver"

378 " interoperability and there is an existing ::CUcontext active on the"

379 " host thread."

380 ),

381 709: (

382 "This error indicates that the context current to the calling thread"

383 " has been destroyed using ::cuCtxDestroy, or is a primary context which"

384 " has not yet been initialized."

385 ),

386 710: (

387 "An assert triggered in device code during kernel execution. The device"

388 " cannot be used again. All existing allocations are invalid. To continue"

389 " using CUDA, the process must be terminated and relaunched."

390 ),

391 711: (

392 "This error indicates that the hardware resources required to enable"

393 " peer access have been exhausted for one or more of the devices"

394 " passed to ::cudaEnablePeerAccess()."

395 ),

396 712: ("This error indicates that the memory range passed to ::cudaHostRegister() has already been registered."),

397 713: (

398 "This error indicates that the pointer passed to ::cudaHostUnregister()"

399 " does not correspond to any currently registered memory region."

400 ),

401 714: (

402 "Device encountered an error in the call stack during kernel execution,"

403 " possibly due to stack corruption or exceeding the stack size limit."

404 " This leaves the process in an inconsistent state and any further CUDA work"

405 " will return the same error. To continue using CUDA, the process must be terminated"

406 " and relaunched."

407 ),

408 715: (

409 "The device encountered an illegal instruction during kernel execution"

410 " This leaves the process in an inconsistent state and any further CUDA work"

411 " will return the same error. To continue using CUDA, the process must be terminated"

412 " and relaunched."

413 ),

414 716: (

415 "The device encountered a load or store instruction"

416 " on a memory address which is not aligned."

417 " This leaves the process in an inconsistent state and any further CUDA work"

418 " will return the same error. To continue using CUDA, the process must be terminated"

419 " and relaunched."

420 ),

421 717: (

422 "While executing a kernel, the device encountered an instruction"

423 " which can only operate on memory locations in certain address spaces"

424 " (global, shared, or local), but was supplied a memory address not"

425 " belonging to an allowed address space."

426 " This leaves the process in an inconsistent state and any further CUDA work"

427 " will return the same error. To continue using CUDA, the process must be terminated"

428 " and relaunched."

429 ),

430 718: (

431 "The device encountered an invalid program counter."

432 " This leaves the process in an inconsistent state and any further CUDA work"

433 " will return the same error. To continue using CUDA, the process must be terminated"

434 " and relaunched."

435 ),

436 719: (

437 "An exception occurred on the device while executing a kernel. Common"

438 " causes include dereferencing an invalid device pointer and accessing"

439 " out of bounds shared memory. Less common cases can be system specific - more"

440 " information about these cases can be found in the system specific user guide."

441 " This leaves the process in an inconsistent state and any further CUDA work"

442 " will return the same error. To continue using CUDA, the process must be terminated"

443 " and relaunched."

444 ),

445 720: (

446 "This error indicates that the number of blocks launched per grid for a kernel that was"

447 " launched via either ::cudaLaunchCooperativeKernel"

448 " exceeds the maximum number of blocks as allowed by ::cudaOccupancyMaxActiveBlocksPerMultiprocessor"

449 " or ::cudaOccupancyMaxActiveBlocksPerMultiprocessorWithFlags times the number of multiprocessors"

450 " as specified by the device attribute ::cudaDevAttrMultiProcessorCount."

451 ),

452 721: (

453 "An exception occurred on the device while exiting a kernel using tensor memory: the"

454 " tensor memory was not completely deallocated. This leaves the process in an inconsistent"

455 " state and any further CUDA work will return the same error. To continue using CUDA, the"

456 " process must be terminated and relaunched."

457 ),

458 800: "This error indicates the attempted operation is not permitted.",

459 801: ("This error indicates the attempted operation is not supported on the current system or device."),

460 802: (

461 "This error indicates that the system is not yet ready to start any CUDA"

462 " work. To continue using CUDA, verify the system configuration is in a"

463 " valid state and all required driver daemons are actively running."

464 " More information about this error can be found in the system specific"

465 " user guide."

466 ),

467 803: (

468 "This error indicates that there is a mismatch between the versions of"

469 " the display driver and the CUDA driver. Refer to the compatibility documentation"

470 " for supported versions."

471 ),

472 804: (

473 "This error indicates that the system was upgraded to run with forward compatibility"

474 " but the visible hardware detected by CUDA does not support this configuration."

475 " Refer to the compatibility documentation for the supported hardware matrix or ensure"

476 " that only supported hardware is visible during initialization via the CUDA_VISIBLE_DEVICES"

477 " environment variable."

478 ),

479 805: "This error indicates that the MPS client failed to connect to the MPS control daemon or the MPS server.",

480 806: "This error indicates that the remote procedural call between the MPS server and the MPS client failed.",

481 807: (

482 "This error indicates that the MPS server is not ready to accept new MPS client requests."

483 " This error can be returned when the MPS server is in the process of recovering from a fatal failure."

484 ),

485 808: "This error indicates that the hardware resources required to create MPS client have been exhausted.",

486 809: "This error indicates the the hardware resources required to device connections have been exhausted.",

487 810: "This error indicates that the MPS client has been terminated by the server. To continue using CUDA, the process must be terminated and relaunched.",

488 811: "This error indicates, that the program is using CUDA Dynamic Parallelism, but the current configuration, like MPS, does not support it.",

489 812: "This error indicates, that the program contains an unsupported interaction between different versions of CUDA Dynamic Parallelism.",

490 900: "The operation is not permitted when the stream is capturing.",

491 901: ("The current capture sequence on the stream has been invalidated due to a previous error."),

492 902: ("The operation would have resulted in a merge of two independent capture sequences."),

493 903: "The capture was not initiated in this stream.",

494 904: ("The capture sequence contains a fork that was not joined to the primary stream."),

495 905: (

496 "A dependency would have been created which crosses the capture sequence"

497 " boundary. Only implicit in-stream ordering dependencies are allowed to"

498 " cross the boundary."

499 ),

500 906: (

501 "The operation would have resulted in a disallowed implicit dependency on"

502 " a current capture sequence from cudaStreamLegacy."

503 ),

504 907: ("The operation is not permitted on an event which was last recorded in a capturing stream."),

505 908: (

506 "A stream capture sequence not initiated with the ::cudaStreamCaptureModeRelaxed"

507 " argument to ::cudaStreamBeginCapture was passed to ::cudaStreamEndCapture in a"

508 " different thread."

509 ),

510 909: "This indicates that the wait operation has timed out.",

511 910: (

512 "This error indicates that the graph update was not performed because it included"

513 " changes which violated constraints specific to instantiated graph update."

514 ),

515 911: (

516 "This indicates that an async error has occurred in a device outside of CUDA."

517 " If CUDA was waiting for an external device's signal before consuming shared data,"

518 " the external device signaled an error indicating that the data is not valid for"

519 " consumption. This leaves the process in an inconsistent state and any further CUDA"

520 " work will return the same error. To continue using CUDA, the process must be"

521 " terminated and relaunched."

522 ),

523 912: ("This indicates that a kernel launch error has occurred due to cluster misconfiguration."),

524 913: ("Indiciates a function handle is not loaded when calling an API that requires a loaded function."),

525 914: ("This error indicates one or more resources passed in are not valid resource types for the operation."),

526 915: ("This error indicates one or more resources are insufficient or non-applicable for the operation."),

527 917: (

528 "This error indicates that the requested operation is not permitted because the"

529 " stream is in a detached state. This can occur if the green context associated"

530 " with the stream has been destroyed, limiting the stream's operational capabilities."

531 ),

532 999: "This indicates that an unknown internal error has occurred.",

533 10000: (

534 "Any unhandled CUDA driver error is added to this value and returned via"

535 " the runtime. Production releases of CUDA should not return such errors."

536 " This error return is deprecated as of CUDA 4.1."

537 ),

538}

Coverage for cuda / core / _utils / runtime_cuda_error_explanations_frozen.py: 0.00%

1 statements