ydb/build/scripts/fatbinary_wrapper.py, branch CLI_2.30.0Mirror of YDB github reposhttps://code.mastervirt.ru/ydb/atom?h=CLI_2.30.02026-02-02T17:54:05Z[build] cuda: Add .module_id sanity check to CUDA_SRCS2026-02-02T17:54:05Zdeshevoydeshevoy@yandex-team.com2026-02-02T17:17:29Zurn:sha1:527205698ee1e45c26021bea3ea3a94b9bfaca5e
IDs generated by different cicc invocations should match
ISSUE:
commit_hash:7cd593cee44b31875e7166709d7614dcfa3f1f14
[build] cuda: Introduce CUDA_SRCS macro utilizing parallelized device code compilation2026-02-01T17:51:53Zdeshevoydeshevoy@yandex-team.com2026-02-01T17:25:47Zurn:sha1:2726a4c04b0e9ddc111f2d1ea56390624cb67228
Instead of a single graph node launching NVCC to compile .cu for both host and all device architectures
CUDA_SRCS generates multiple nodes:
- node per each device architecture producing PTX and CUBIN
- node merging all PTX and CUBIN files into a single FATBIN blob
- node producing .cpp with host code
- node compiling host .cpp with embedded FATBIN blob
CUDA_ARCHITECTURES variable is used to determine the list of architectures to compile device code for.
ISSUE:
commit_hash:0a4c2797dd238ae062482af30694df6978301278