Swatantra-66 commented on issue #20483: URL: https://github.com/apache/nuttx/issues/20483#issuecomment-6252733940
Hi @acassis, I audited the current 1,230 ARM defconfigs to check the workload distribution and ensure the containers are well-balanced as you noted. Currently, `arm-01` through `arm-14` have significant imbalance: some jobs build only ~41 configs (`arm-04`, `arm-10`) while others build ~199 configs (`arm-06`, `arm-13`), leading to idle runners waiting on the slowest container. By grouping chips by vendor/family, we can distribute them across **13 jobs** (1 fewer container than today) where every job compiles between **60 and 120 configs**: - `arm-stm32f4-1` & `arm-stm32f4-2`: ~81 configs each (splitting STM32F4) - `arm-stm32f0-f3`: ~116 configs - `arm-stm32h7`: ~97 configs - `arm-stm32l-u-g`: ~82 configs (L0-L5, U0/U5, G0/G4, WB/WL) - `arm-stm32-other`: ~63 configs (F7, H5, C0, N6) - `arm-rp2040`: ~130 configs (RP2040 & RP2350) - `arm-imxrt`: ~74 configs (i.MX RT & i.MX6/9) - `arm-nxp`: ~100 configs (LPC17xx–54xx, Kinetis, S32K) - `arm-nrf`: ~83 configs (nRF52, nRF53, nRF91) - `arm-sam`: ~88 configs (SAM34, SAMA5, SAMD/E/V) - `arm-ti-cxd56`: ~74 configs (Tiva, RM57, CXD56xx) - `arm-other`: ~60 configs (GD32, AT32, Max326, MPS) This ensures every container handles a meaningful workload without runner starvation, while allowing isolated chip PRs (e.g. on `imxrt` or `nrf52`) to execute in just 1 container in 2–3 minutes. I will go ahead and prepare the branch and PR with these balanced groupings for your and @xiaoxiang781216's review. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
