[Feb 28 Session] Next-level CUDA-to-SYCL optimizations

Further fine-tune cross-platform code to target GPUs and AI accelerators.

[Time zone converter] [Available On-Demand on Thursday, February 29] When migrating code from CUDA to C++ with SYCL using the Intel® DPC++ Compatibility Tool or SYCLomatic, additional CUDA-specific routines may be needed to ensure optimal performance on your target GPU or AI accelerator. This session unpacks what they are: CPU offload kernel problem-sizing CPU-GPU data caching, data transfer, and memory use optimization Unified Shared Memory vs Buffered Memory SYCL interoperability Resolving use of custom PTX instructions Coexistence with MPI and distributed computing in HPC Note: It is recommended you have working knowledge of C/C++. Skill level: All Register now Featured software This session features the following tools, which you can get standalone or as part of the Intel® oneAPI Base Toolkit (Note the Intel DPC++ Compatibility Tool is not available standalone): SYCLomatic [GitHub] Intel® oneAPI DPC++ Compiler Intel® Advisor Intel® VTune™Profiler Intel® oneAPI Math Kernel Library Download code samples Get Started: Migrate CUDA to SYCL See all code samples View in browser.

                                                           

This was sent to ivwinds.steeds@blogger.com. If you forward this email, your contact information will appear in any auto-populated form connected to links in this email.  To view and manage your marketing-related email preferences with Intel, please click here. © 2024 Intel Corporation Intel Corporation, 2200 Mission College Blvd., M/S RNB4-145, Santa Clara, CA 95054 USA. www.intel.com Privacy | Cookies | *Trademarks | Unsubscribe | Manage Preferences