Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

FS#867 - mvebu: Should be split different arches, current (Armada 370, XP and other "legacy") and 385+ #5801

Open
openwrt-bot opened this issue Jun 25, 2017 · 2 comments
Labels

Comments

@openwrt-bot
Copy link

openwrt-bot commented Jun 25, 2017

diizzyy:

The older SoCs such as Armada XP and 370 only supports vfpv3-d16 while 385 and newer does complete vfpv3 and neon. This hampers performance of newer SoCs quite a bit and makes little sense not to split these since we already build vfpv3+neon binaries. While neon-vfpv4 exists, neon-vfpv3 doesn't and for some reason it's just called neon in GCC (which translates to arm_cortex-a9_neon in LEDE for this SoC family). Source: https://lists.linaro.org/pipermail/linaro-toolchain/2016-July/005826.html

@openwrt-bot
Copy link
Author

openwrt-bot commented Jun 25, 2017

anomeome:

Limiting to lowest common [[https://forum.lede-project.org/t/new-package-arch-for-mvebu/4235|denominator]] to reduce target count appears to be a significant penalty for the majority of mvebu targets.

@openwrt-bot
Copy link
Author

openwrt-bot commented May 21, 2020

FrederickZh:

brand model target subtarget packagearchitecture cpu FPU/SIMD
Globalscale Mirabox mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell 88F6707 VFPv3-D16/?
Marvell RD-A370-A1 mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 370 VFPv3-D16/?
SolidRun ClearFog Pro mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 388 88F6828 VFPv3-D32/NEON
SolidRun ClearFog Base mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 388 88F6828 VFPv3-D32/NEON
Linksys WRT1200AC mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 385 88F6820 VFPv3-D32/NEON
Linksys WRT1900AC v2 mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 385 88F6820 VFPv3-D32/NEON
Linksys WRT1900ACS mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 385 88F6820 VFPv3-D32/NEON
Linksys WRT3200ACM mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 385 88F6820 VFPv3-D32/NEON
Turris CZ.NIC Omnia mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 385 88F6820 VFPv3-D32/NEON
Linksys WRT32X mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 385 88F6820 VFPv3-D32/NEON
Marvell Armada A385 DB AP mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada 385 88F6820 VFPv3-D32/NEON
Marvell Armada A370 DB mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada A370 VFPv3-D16/?
Linksys WRT1900AC v1 mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada XP MV78230 VFPv3-D16/?
PlatHome OpenBlocks AX3-4 mvebu cortexa9 arm_cortex-a9_vfpv3 Marvell Armada XP MV78260 VFPv3-D16/?

I just downloaded a copy of the supported devices CSV. There are currently only 5 vfp3-d16 devices, among which RD-A370-A1 and Armada A370 DB are reference boards and WRT1900AC has got a v2 that according to OpenWRT DB was released in the same year (2018). So I reckon it's reasonable to even drop vfp3-d16 if the number of officially built architectures is a concern (sorry PlatHome and Globalscale guys).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant