[tmva][sofie] Implement multi-operator fusion for efficient memory usage #22990

Job	Run time
Windows 10 x64 Release	27m 11s
Upload Event Payload	4s
mac13 ARM64 builtin_zlib=ON	25m 45s
Windows 10 x86 Release	40m 52s
mac15 ARM64 CMAKE_CXX_STANDARD=23	3m 11s
mac14 X64 CMAKE_CXX_STANDARD=20	59m 46s
fedora43 CMAKE_CXX_STANDARD=23	26m 32s
mac-beta ARM64 CMAKE_CXX_STANDARD=23	5m 10s
ubuntu2510 CMAKE_CXX_STANDARD=23	37m 53s
alma10	27m 37s
ubuntu22 imt=Off, CMAKE_BUILD_TYPE=Debug	1h 22m 52s
fedora42 CMAKE_CXX_STANDARD=23	29m 45s
ubuntu2404 CMAKE_BUILD_TYPE=Debug	1h 21m 26s
debian125 CMAKE_CXX_STANDARD=20, dev=ON, CMAKE_CXX_FLAGS=-Wsuggest-override	1m 57s
alma9 arm64 CMAKE_BUILD_TYPE=RelWithDebInfo, builtin_zlib=ON, builtin_zstd=ON	40m 29s
alma10 clang Ninja CMAKE_C_COMPILER=clang, CMAKE_CXX_COMPILER=clang++	24m 0s
alma9 CMAKE_BUILD_TYPE=Debug	1h 5m 43s
ubuntu2504 CMAKE_CXX_STANDARD=23	35m 15s
rawhide Fedora pydebug CMAKE_CXX_STANDARD=23	27m 2s
alma9 march_native CMAKE_BUILD_TYPE=RelWithDebInfo, CMAKE_CXX_FLAGS=-march=native, CMAKE_C_FLAGS=-march=native, builtin_zlib=ON, builtin_zstd=ON	27m 40s
alma8	42m 17s
alma9 modules_off runtime_cxxmodules=Off	28m 33s
	12h 21m 0s

Provide feedback