Presenter: Zhi Chen, Cody Yu Amazon SageMaker Neo, Deep Engine Science Bring Your Own Codegen to TVM AWS AI© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Considering You. NMS is supported by TVM!© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Let TVM Be the Compiler of Your Chip Your chip can run any models Your compiler (TVM) supports multiple reserved. Example showcase: Intel MKL-DNN (DNNL) library 1. Import packages import numpy as np from tvm import relay 2. Load a pretrained network mod, params = relay.testing.mobilenet.get_workload(batch_size=1)

Services, Inc. or its Affiliates. All rights reserved. Amazon Trademark Deploying GluonCV models using TVM© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Trademark Deploy GluonCV Models GluonCV Models MXNet Computational Graph Json Acyclic Graph Export As-is Optimize with TVM© 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon Trademark Deploy GluonCV Affiliates. All rights reserved. Amazon Trademark Effects of Vision-specific Optimizations using TVM Speedup 0 1 2 3 SSD_MobileNet1.0 SSD_ResNet50 Yolov3 w/o w/ Amazon DeepLens Speedup 0 1

Python3 Step 4.3: Copy cma.ko to home directory Step 5: Cross compile TVM with USE_VTA_FPGA flag ON Step 6: Copy the compiled TVM to the SDCard Step 7: Install kernel module cma.ko and run apps/vta_

Poligeist SYCL MLIR Bisheng C++ Celerity SYCL MPI Vulkan Sylkan Nvidia HPC SDK MotorSYCL TVM/AKG Ascend AI Chipsets DPC++ fork PIM SDK Samsung PIM-HBM2 PIM emulator DPC++ fork SYCLops NewSYCL

이에 대해서는 다음 절에 설명한다. 110 10. RV32/64 특권 구조 XLEN-1 XLEN-2 23 22 21 20 19 18 17 SD Reserved TSR TW TVM MXR SUM MPRV 1 XLEN-24 1 1 1 1 1 1 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 XS FS MPP Res. SPP MPIE