Re: R on riscv64 (Re: building fewer packages or reduced functionality on i386)
Aurelien Jarno <[email protected]>
| Newsgroups | gmane.linux.debian.riscv,gmane.linux.debian.devel.general |
|---|---|
| Message-ID | <[email protected]> |
Hi Charles, On 2026-08-15 21:33, Charles Plessy wrote: > Le Sat, Aug 15, 2026 at 02:20:25PM +0300, Adrian Bunk a écrit : > > > > riscv64 is not yet used in HPC, but work on that is ongoing.[1] > > > > [1] https://www.researchgate.net/publication/405221610_Monte_Cimone_v3_Where_RISC-V_Stands_in_High-Performance_Computing > > Hi Adrian and riscv64 porters, > > The R language has some imporance in scientific computing performed on > HPC clusters, and I think that our autopktests have uncovered a serious > issue between the language and the architecture, which I worry might be > release-critical. > > I have summarised the issue on https://wiki.debian.org/R/riscv64 (thanks > to the original input from Aurélien Jarno) and I want to contact the > core R developers (who do not yet support that architecture formally), > but before that I wanted to be totally confident in what I wrote, and I > asked for proofreading on the riscv porters mailing list, three times. > Unfortunately, I had no answer… > > I do not want to open now a full discussion on what to do in Debian for > Forky, but I would be relieved and grateful if I could have another pair > of eyes checking https://wiki.debian.org/R/riscv64 before it reaches > outside Debian circles. Thanks for writing this document. I have proofread it, and it looks clear to me. I haven't spotted any problem in the "Summary" and "Debian context" sections, and I feel less competent for reviewing the "Suggestions for R packages developers", but it also looks clear. I have made a few minor changes directly on the wiki, I hope it helps. Regards Aurelien -- Aurelien Jarno GPG: 4096R/1DDD8C9B [email protected] http://aurel32.net