I'm sorry I don't know how to create ELF headers either. If it were me, I'd look at FreePascal and the Go language toolchains as they are independent from the standard GNU tools and likely simpler.@trejan,
Thank you for taking a look. ELF format is an arcane dark corner I never seem to understand. I am basically using the ELF generation code, verbatim, from the open source AMaCC compiler on github (my project is forked from that one), and I have never understood it. I have tried looking at ELF documentation, but everything I have found seems to assume you have read the entire ELF standard and worked with it on every architecture, in every possible configuration, aka all-or-nothing knowledge assumption.
Do you happen to know of a terse hands-on ELF tutorial that shows you step-by-step how to properly create an ELF header?
I found it interesting on Github where
My experience writing parallel programs in Go, Chapel, C/Cilk and C/OpenMP is that there is not enough control where to allocate memory and run threads on NUMA architectures. To me it seems that a complex arrangement of multiple zones became mainstream with the first generation of Epyc processors and are likely to become even more common with the commoditization of CXL memory and network-based memory clustering. The heterogenous combination of performance and efficiency cores on desktop processors is also not well managed.By the end of 2030, I expect to have automatic vectorization and/or parallelization working in my HPC compiler. The HPC extensions/restrictions make it "natural" to manage parallel partitions, unlike the mess created by C standard semantics.
Does your notion of parallel partitions provide a way to manage NUMA and heterogeneous CPU architectures?
What mess do C standard semantics make?
Statistics: Posted by ejolson — Tue Apr 22, 2025 7:15 pm