Logging & debugging#

All logging messages are output to the standard error stream (stderr). You can use the following environmental variables to control logging. Both variables are disabled by default.

  • MIOPEN_ENABLE_LOGGING: Print the basic layer-by-layer MIOpen API call information with actual parameters (configurations). This information is important for debugging.

    • To enable feature: 1, on, yes, true, enable, enabled

    • To disable feature: 0, off, no, false, disable, disabled

  • MIOPEN_ENABLE_LOGGING_CMD: Output the associated MIOpenDriver command lines into the console.

    • To enable feature: 1, on, yes, true, enable, enabled

    • To disable feature: 0, off, no, false, disable, disabled

  • MIOPEN_LOG_LEVEL: In addition to API call information and driver commands, MIOpen prints information related to the progress of its internal operations. This information can be useful for debugging and understanding the library’s principles of operation. MIOPEN_LOG_LEVEL controls the verbosity of these messages. Allowed values are:

    • 0: Default. Works as level 4 for release builds and level 5 for debug builds.

    • 1: Quiet. No logging messages.

    • 2: Fatal errors only (not used yet).

    • 3: Errors and fatals.

    • 4: All errors and warnings.

    • 5: Info. All the preceding information, plus information for debugging purposes.

    • 6: Detailed information. All the preceding information, plus more detailed information for debugging.

    • 7: Trace. All the preceding information, plus additional details.

  • MIOPEN_ENABLE_LOGGING_MPMT: When enabled, each log line is prefixed with information you can use to identify records printed from different processes or threads. This is useful for debugging multi-process/multi-threaded applications.

  • MIOPEN_ENABLE_LOGGING_ELAPSED_TIME: Adds a timestamp to each log line that indicates the time elapsed (in milliseconds) since the previous log message.

Tip

If you require technical support, include the console log that is produced from:

export MIOPEN_ENABLE_LOGGING=1
export MIOPEN_ENABLE_LOGGING_CMD=1
export MIOPEN_LOG_LEVEL=6

Layer filtering#

The following sections contain environment variables that you can use to enable or disable various kinds of kernels and algorithms. These are helpful for debugging MIOpen and framework integrations.

For these environment variables, you can use the following values:

  • To enable kernel/algorithm: 1, yes, true, enable, enabled

  • To disable kernel/algorithm: 0, no, false, disable, disabled

Warning

When you use the library with layer filtering, the results of *Find() calls become narrower than during normal operation. This means that relevant FindDb entries won’t include all the solutions that are normally there. Therefore, the subsequent immediate mode *Get() calls may return incomplete information or run into fallback path.

In order to rehabilitate immediate mode, you can:

  • Re-enable all solvers and re-run the same *Find() calls you previously ran

  • Completely remove the User FindDb

If no variable is set, MIOpen behaves as if the variable is set to enabled. This means that kernels and algorithms are enabled by default.

Filtering by algorithm#

These variables control the sets (families) of convolution solutions. For example, the direct algorithm is implemented in several solutions that use OpenCL and GCN assembly. The corresponding variable can disable them all.

  • MIOPEN_DEBUG_CONV_FFT: FFT convolution algorithm.

  • MIOPEN_DEBUG_CONV_DIRECT: Direct convolution algorithm.

  • MIOPEN_DEBUG_CONV_GEMM: GEMM convolution algorithm.

  • MIOPEN_DEBUG_CONV_WINOGRAD: Winograd convolution algorithm.

  • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM: Implicit GEMM convolution algorithm.

Filtering by build method#

  • MIOPEN_DEBUG_GCN_ASM_KERNELS: Kernels written in assembly language. These are used in many convolutions (some direct solvers, Winograd kernels, and fused convolutions) and batch normalization.

  • MIOPEN_DEBUG_HIP_KERNELS: Convolution kernels written in HIP. These implement the ImplicitGemm algorithm.

  • MIOPEN_DEBUG_OPENCL_CONVOLUTIONS: Convolution kernels written in OpenCL; this only affects convolutions.

Filtering out all but one solution#

  • MIOPEN_DEBUG_FIND_ONLY_SOLVER=solution_id: Directly affects only *Find() calls. However, there is an indirect connection to immediate mode (per the previous warning).

    • solution_id must be a numeric or a string identifier of some solution.

    • If solution_id denotes some applicable solution, then only that solution is found (in addition to GEMM and FFT, if applicable–refer to the following note).

    • If solution_id is valid, but not applicable, then *Find() fails with all algorithms (except for GEMM and FFT,–refer to the following note).

    • Otherwise, the solution_id is invalid (i.e., it doesn’t match any existing solution) and the *Find() call fails.

Note

This environmental variable doesn’t affect the GEMM and FFT solutions. For now, GEMM and FFT can only be disabled at the algorithm level.

Filtering the solutions on an individual basis#

Some of the solutions have individual controls, which affect both find and immediate modes.

  • Direct solutions:

    • MIOPEN_DEBUG_CONV_DIRECT_ASM_3X3UConvAsm3x3U

    • MIOPEN_DEBUG_CONV_DIRECT_ASM_1X1UConvAsm1x1U

    • MIOPEN_DEBUG_CONV_DIRECT_ASM_1X1UV2ConvAsm1x1UV2

    • MIOPEN_DEBUG_CONV_DIRECT_ASM_5X10U2V2ConvAsm5x10u2v2f1`, `ConvAsm5x10u2v2b1

    • MIOPEN_DEBUG_CONV_DIRECT_ASM_7X7C3H224W224ConvAsm7x7c3h224w224k64u2v2p3q3f1

    • MIOPEN_DEBUG_CONV_DIRECT_ASM_WRW3X3ConvAsmBwdWrW3x3

    • MIOPEN_DEBUG_CONV_DIRECT_ASM_WRW1X1ConvAsmBwdWrW1x1

    • MIOPEN_DEBUG_CONV_DIRECT_OCL_FWD11X11ConvOclDirectFwd11x11

    • MIOPEN_DEBUG_CONV_DIRECT_OCL_FWDGENConvOclDirectFwdGen

    • MIOPEN_DEBUG_CONV_DIRECT_OCL_FWDConvOclDirectFwd

    • MIOPEN_DEBUG_CONV_DIRECT_OCL_FWD1X1ConvOclDirectFwd1x1

    • MIOPEN_DEBUG_CONV_DIRECT_OCL_WRW2ConvOclBwdWrW2<n> (where n = {1,2,4,8,16}) and ConvOclBwdWrW2NonTunable

    • MIOPEN_DEBUG_CONV_DIRECT_OCL_WRW53ConvOclBwdWrW53

    • MIOPEN_DEBUG_CONV_DIRECT_OCL_WRW1X1ConvOclBwdWrW1x1

  • Winograd solutions:

    • MIOPEN_DEBUG_AMD_WINOGRAD_3X3ConvBinWinograd3x3U, FP32 Winograd Fwd/Bwd, filter size fixed to 3x3

    • MIOPEN_DEBUG_AMD_WINOGRAD_RXSConvBinWinogradRxS, FP32/FP16 F(3,3) Fwd/Bwd and FP32 F(3,2) WrW Winograd. Subsets:

      • MIOPEN_DEBUG_AMD_WINOGRAD_RXS_WRW – FP32 F(3,2) WrW convolutions only

      • MIOPEN_DEBUG_AMD_WINOGRAD_RXS_FWD_BWD – FP32/FP16 F(3,3) Fwd/Bwd

    • MIOPEN_DEBUG_AMD_WINOGRAD_RXS_F3X2ConvBinWinogradRxSf3x2, FP32/FP16 Fwd/Bwd F(3,2) Winograd

    • MIOPEN_DEBUG_AMD_WINOGRAD_RXS_F2X3ConvBinWinogradRxSf2x3, FP32/FP16 Fwd/Bwd F(2,3) Winograd, serves group convolutions only

    • MIOPEN_DEBUG_AMD_WINOGRAD_RXS_F2X3_G1ConvBinWinogradRxSf2x3g1, FP32/FP16 Fwd/Bwd F(2,3) Winograd, for non-group convolutions

  • Multi-pass Winograd:

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F3X2ConvWinograd3x3MultipassWrW<3-2>, WrW F(3,2), stride 2 only

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F3X3ConvWinograd3x3MultipassWrW<3-3>, WrW F(3,3), stride 2 only

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F3X4ConvWinograd3x3MultipassWrW<3-4>, WrW F(3,4)

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F3X5ConvWinograd3x3MultipassWrW<3-5>, WrW F(3,5)

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F3X6ConvWinograd3x3MultipassWrW<3-6>, WrW F(3,6)

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F5X3ConvWinograd3x3MultipassWrW<5-3>, WrW F(5,3)

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F5X4ConvWinograd3x3MultipassWrW<5-4>, WrW F(5,4)

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F7X2:

      • ConvWinograd3x3MultipassWrW<7-2>, WrW F(7,2)

      • ConvWinograd3x3MultipassWrW<7-2-1-1>, WrW F(7x1,2x1)

      • ConvWinograd3x3MultipassWrW<1-1-7-2>, WrW F(1x7,1x2)

    • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_F7X3:

      • ConvWinograd3x3MultipassWrW<7-3>, WrW F(7,3)

      • ConvWinograd3x3MultipassWrW<7-3-1-1>, WrW F(7x1,3x1)

      • ConvWinograd3x3MultipassWrW<1-1-7-3>, WrW F(1x7,1x3)

    • MIOPEN_DEBUG_AMD_MP_BD_WINOGRAD_F2X3ConvMPBidirectWinograd<2-3>, FWD/BWD F(2,3)

    • MIOPEN_DEBUG_AMD_MP_BD_WINOGRAD_F3X3ConvMPBidirectWinograd<3-3>, FWD/BWD F(3,3)

    • MIOPEN_DEBUG_AMD_MP_BD_WINOGRAD_F4X3ConvMPBidirectWinograd<4-3>, FWD/BWD F(4,3)

    • MIOPEN_DEBUG_AMD_MP_BD_WINOGRAD_F5X3ConvMPBidirectWinograd<5-3>, FWD/BWD F(5,3)

    • MIOPEN_DEBUG_AMD_MP_BD_WINOGRAD_F6X3ConvMPBidirectWinograd<6-3>, FWD/BWD F(6,3)

    • MIOPEN_DEBUG_AMD_MP_BD_XDLOPS_WINOGRAD_F2X3ConvMPBidirectWinograd_xdlops<2-3>, FWD/BWD F(2,3)

    • MIOPEN_DEBUG_AMD_MP_BD_XDLOPS_WINOGRAD_F3X3ConvMPBidirectWinograd_xdlops<3-3>, FWD/BWD F(3,3)

    • MIOPEN_DEBUG_AMD_MP_BD_XDLOPS_WINOGRAD_F4X3ConvMPBidirectWinograd_xdlops<4-3>, FWD/BWD F(4,3)

    • MIOPEN_DEBUG_AMD_MP_BD_XDLOPS_WINOGRAD_F5X3ConvMPBidirectWinograd_xdlops<5-3>, FWD/BWD F(5,3)

    • MIOPEN_DEBUG_AMD_MP_BD_XDLOPS_WINOGRAD_F6X3ConvMPBidirectWinograd_xdlops<6-3>, FWD/BWD F(6,3)

    • MIOPEN_DEBUG_AMD_MP_BD_WINOGRAD_EXPEREMENTAL_FP16_TRANSFORMConvMPBidirectWinograd*, FWD/BWD FP16 experimental mode (use at your own risk). Disabled by default.

    • MIOPEN_DEBUG_AMD_FUSED_WINOGRAD – Fused FP32 F(3,3) Winograd, variable filter size.

Implicit GEMM solutions:

  • ASM implicit GEMM

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_FWD_V4R1ConvAsmImplicitGemmV4R1DynamicFwd

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_FWD_V4R1_1X1ConvAsmImplicitGemmV4R1DynamicFwd_1x1

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_BWD_V4R1ConvAsmImplicitGemmV4R1DynamicBwd

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_WRW_V4R1ConvAsmImplicitGemmV4R1DynamicWrw

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_FWD_GTC_XDLOPSConvAsmImplicitGemmGTCDynamicFwdXdlops

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_BWD_GTC_XDLOPSConvAsmImplicitGemmGTCDynamicBwdXdlops

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_WRW_GTC_XDLOPSConvAsmImplicitGemmGTCDynamicWrwXdlops

  • HIP implicit GEMM

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_FWD_V4R1ConvHipImplicitGemmV4R1Fwd

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_FWD_V4R4ConvHipImplicitGemmV4R4Fwd

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_BWD_V1R1ConvHipImplicitGemmBwdDataV1R1

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_BWD_V4R1ConvHipImplicitGemmBwdDataV4R1

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_WRW_V4R1ConvHipImplicitGemmV4R1WrW

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_WRW_V4R4ConvHipImplicitGemmV4R4WrW

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_FWD_V4R4_XDLOPSConvHipImplicitGemmForwardV4R4Xdlops

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_FWD_V4R5_XDLOPSConvHipImplicitGemmForwardV4R5Xdlops

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_BWD_V1R1_XDLOPSConvHipImplicitGemmBwdDataV1R1Xdlops

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_BWD_V4R1_XDLOPSConvHipImplicitGemmBwdDataV4R1Xdlops

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_WRW_V4R4_XDLOPSConvHipImplicitGemmWrwV4R4Xdlops

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_FWD_V4R4_PADDED_GEMM_XDLOPSConvHipImplicitGemmForwardV4R4Xdlops_Padded_Gemm

    • MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_HIP_WRW_V4R4_PADDED_GEMM_XDLOPSConvHipImplicitGemmWrwV4R4Xdlops_Padded_Gemm

rocBlas logging and behavior#

The ROCBLAS_LAYER environmental variable can be set to output GEMM information:

  • ROCBLAS_LAYER=: Not set–there is no logging

  • ROCBLAS_LAYER=1: Trace logging

  • ROCBLAS_LAYER=2: Bench logging

  • ROCBLAS_LAYER=3: Trace and bench logging

You can also set the MIOPEN_GEMM_ENFORCE_BACKEND environment variable to override the default GEMM backend (rocBLAS):

  • MIOPEN_GEMM_ENFORCE_BACKEND=1: Use rocBLAS if enabled

  • MIOPEN_GEMM_ENFORCE_BACKEND=2: Reserved

  • MIOPEN_GEMM_ENFORCE_BACKEND=3: No GEMM is called

  • MIOPEN_GEMM_ENFORCE_BACKEND=4: Reserved

  • MIOPEN_GEMM_ENFORCE_BACKEND=<any other value>: Use default behavior

To disable using rocBlas entirely, set the -DMIOPEN_USE_ROCBLAS=Off configuration flag during MIOpen configuration.

You can find more information on logging with rocBLAS in the rocBLAS programmer guide.

Numerical checking#

You can use the MIOPEN_CHECK_NUMERICS environmental variable to debug potential numerical abnormalities. Setting this variable scans all inputs and outputs of each kernel called and attempts to detect infinities (infs), not-a-number (NaN), and all zeros. This environment variable has several settings that help with debugging:

  • MIOPEN_CHECK_NUMERICS=0x01: Fully informative. Prints results from all checks to console.

  • MIOPEN_CHECK_NUMERICS=0x02: Warning information. Prints results only if an abnormality is detected.

  • MIOPEN_CHECK_NUMERICS=0x04: Throw error on detection. MIOpen runs MIOPEN_THROW upon abnormal result.

  • MIOPEN_CHECK_NUMERICS=0x08: Abort upon abnormal result. Allows you to drop into a debugging session.

  • MIOPEN_CHECK_NUMERICS=0x10: Print stats. Computes and prints mean/absmean/min/max (note that this is slow).

Controlling parallel compilation#

MIOpen’s convolution *Find() calls compile and benchmark a set of solvers contained in miopenConvAlgoPerf_t. This is done in parallel per miopenConvAlgorithm_t. Parallelism per algorithm is set to 20 threads. Typically, there are far fewer threads spawned due to the limited number of kernels under any given algorithm.

You can control the level of parallelism using the MIOPEN_COMPILE_PARALLEL_LEVEL environment variable.

To disable multi-threaded compilation, run:

export MIOPEN_COMPILE_PARALLEL_LEVEL=1

Experimental controls#

Using experimental controls may result in:

  • Performance drops

  • Computation inaccuracies

  • Runtime errors

  • Other kinds of unexpected behavior

We strongly recommended only using these controls at the explicit request of the library developers.

Code Object version selection (experimental)#

Different ROCm versions use Code Object (CO) files from different versions (i.e., formats). The library automatically uses the most suitable version. The following variables allow for experimenting and triaging possible problems related to CO version:

  • MIOPEN_DEBUG_AMD_ROCM_METADATA_ENFORCE: Affects kernels written in GCN assembly language.

    • 0 (or unset): Automatically detects the required CO version and assembles to that version. This is the default.

    • 1: Do not auto-detect CO version; always assemble v2 COs.

    • 2: Behave as if both v2 and v3 COs are supported (see MIOPEN_DEBUG_AMD_ROCM_METADATA_PREFER_OLDER).

    • 3: Always assemble v3 COs.

  • MIOPEN_DEBUG_AMD_ROCM_METADATA_PREFER_OLDER: This variable only affects assembly kernels, and only when ROCm supports both v2 and v3 COs (like ROCm 2.10). By default, the newer format is used (v3 CO). When this variable is enabled, the behavior is reversed.

  • MIOPEN_DEBUG_OPENCL_ENFORCE_CODE_OBJECT_VERSION: Enforces CO format for OpenCL kernels. This only works with the HIP backend (cmake ... -DMIOPEN_BACKEND=HIP...).

    • Unset - Automatically detect the required CO version. This is the default.

    • 2: Always build to v2 CO.

    • 3: Always build to v3 CO.

    • 4: Always build to v4 CO.

Winograd multi-pass maximum workspace throttling#

  • MIOPEN_DEBUG_AMD_WINOGRAD_MPASS_WORKSPACE_MAXConvWinograd3x3MultipassWrW, WrW

  • MIOPEN_DEBUG_AMD_MP_BD_WINOGRAD_WORKSPACE_MAXConvMPBidirectWinograd*, FWD BWD

Syntax of value:

  • Decimal or hex (with 0x prefix) value that must fit into a 64-bit unsigned integer

  • If the syntax is violated, then the behavior is unspecified

Semantics:

  • Sets the limit (max allowed workspace size) for multi-pass (MP) Winograd solutions, in bytes.

  • Affects all MP Winograd solutions. If a solution needs more workspace than the limit, it doesn’t apply.

  • If unset, then the default limit is used. The current default is 2000000000 (~1.862 GiB) for gfx900 and gfx906/60 (or less CUs). No default limit is set for other GPUs.

  • Special values:

    • 0: Use the default limit, as if the variable is unset

    • 1: Completely prohibit the use of workspace

    • -1: Remove the default limit