lhutton1 opened a new pull request, #15468: URL: https://github.com/apache/tvm/pull/15468
https://github.com/apache/tvm/pull/12455 slightly altered the behaviour when selecting an int8 conv2d schedule. Previously conditions that decide which schedule to select used `is_aarch64` which checks for the existance of `aarch64` in the target triple. However, the conditions now use `has_asimd` which is true if `aarch64` exists in the target triple OR `+neon` is used in the mattr. Both `conv2d_NHWC_quantized_interleaved.arm_cpu` and `depthwise_conv2d_nhwc.arm_cpu` makes calls to LLVM intrinsics that require both `aarch64` and `+neon`. But in the case of the target `rasp4b`, the updated conditions result in compilation failure since the target has `+neon` but doesn't have `aarch64` in the target triple. The conditions have been updated to fix the compilation failure. Likewise, the previous behaviour of the condition for `conv2d_nhwc_spatial_pack.arm_cpu` has been restored ensure a program with a 32-bit target can still be compiled. Finally, we should only select the `depthwise_conv2d_nhwc_dsp.arm_cpu` schedule when a backend that understands `pragma_import_c` has been selected, i.e. "c". For a more detailed discussion of the issue please see: https://discuss.tvm.apache.org/t/tflite-llvm-llvm-error-when-compiling-tflite-model/15411 cc @Mousius @ashutosh-arm @ekalda -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
