Xulin Zhou

llvm.sdiv results in ld.lld error: undefined symbol __divdi3

1

There is a conv_2d_nchw_fchw im2col implementation(conv2d-nchw-fchw-im2col.mlir) containing `llvm.sdiv` after lowering, which will cause __divdi3 -related errors as follows: ``` bash-5.2$ time mill -i 'tests.run[conv2d-nchw-fchw-im2col.mlir].run' [7/438] tests.cases.buddy[conv2d-nchw-fchw-im2col.mlir].elf ld.lld: error: undefined symbol:...

[buddy] The last step to achieve deep learning model e2e inference in vector repo

1

Recently buddy-benchmark has implemented the e2e compilation of "EfficientNet-Quant" (see https://github.com/buddy-compiler/buddy-benchmark/tree/main/benchmarks/DeepLearning/Models/EfficientNet-Quant), a typical deep learning quantized model without float point operations involved in the lowering process. Executing it on the...

Will setting different `lmul` result in obvious performance difference?

5

In RVV, `lmul` stands for "Vector Register Group Multiplier" to specify the number of registers to create a group. Different `lmul` settings have different number of instructions generated when executing...

Errors occur when modify axpy-masked.mlir in some ways

1

In the buddy case `axpy-masked.mlir`(AX plus Y cases with mask operations), errors will occur if we modify it in the following way: 1. Changing `%mask64 = arith.constant dense : vector`...