RT-AX89X

[K-2SO]

My question was more related to RAM utilization. I would trade some performance for less RAM.

 

[K-2SO]

Yes, but the AES acceleration is hardware.

Available in hardware doesn’t necessarily mean used by software. Speaking of VPN, how many available CPU cores your VPN client uses?

 

[L&LD]

One.

 

[RMerlin]

So basically routers these days have 64-bit processors but run 32-bit software?

Only a few do have a 64-bit capable CPU. In Asus's case, the kernel is 64-bit, however all the userland components are 32-bit.

Unrelated to this thread, but what is your advice in regards to 32-bit vs 64-bit Entware used with Asuswrt-Merlin scripts?

Shouldn't matter much, as nothing you'll run will require the extra memory address space anyway.

 

[K-2SO]

AMTM Entware installer suggests 64-bit on AC86U and AX88U. Perhaps @thelonelycoder can give more details why. I found what to play with tomorrow.

 

[Odkrys]

pi4 (cortex-a72 quadcore) 32bit vs 64bit.
It is true that overall performance is good on 64bit.

https://medium.com/@matteocroce/why-you-should-run-a-64-bit-os-on-your-raspberry-pi4-bd5290d48947

I just think 32bit kernel is one of the cons.

 

[K-2SO]

Good numbers. What part of that performance increase is usable in scripts available with AMTM? We don't have 64-bit OS.

 

[thelonelycoder]

AMTM Entware installer suggests 64-bit on AC86U and AX88U. Perhaps @thelonelycoder can give more details why. I found what to play with tomorrow.

The 64-bit Entware version will be around longer than the 32-bit, that is my thinking for the recommendation.
Also, some packages may no longer be available on 32-bit in the (far) future as development on that platform phases out. The Entware team also recommends to use 64-bit.

 

[K-2SO]

may no longer be available on 32-bit in the (far) future

In this case 32-bit version is OK in order to save some RAM.

 

[thelonelycoder]

In this case 32-bit version is OK in order to save some RAM.

I doubt there's a noticeable difference for the RAM usage between the 32 and 64-bit Entware versions.

 

[RMerlin]

pi4 (cortex-a72 quadcore) 32bit vs 64bit.
It is true that overall performance is good on 64bit.

https://medium.com/@matteocroce/why-you-should-run-a-64-bit-os-on-your-raspberry-pi4-bd5290d48947

I just think 32bit kernel is one of the cons.

Really interesting read. I suspect it's not just about 32-bit vs 64-bit, but possibly other enhancements the armv8 ISA has over armv7, but the end result would still be the same, regardless of the "why".

Something is off with their tests however. Their OpenVPN armv8 test clearly has no AES acceleration, as it's half of the speed of a BCM4908's Cortex B53.

 

[netware5]

.............

Haha @RMerlin, just occasionally noticed that the above is your 32768th (1000000000000000) post

 

[Odkrys]

Something is off with their tests however. Their OpenVPN armv8 test clearly has no AES acceleration, as it's half of the speed of a BCM4908's Cortex B53.

pi series has no aes acceleration.
It need a additional license.
cpuinfo of pi4.

pi@raspberrypi:~ $ cat /proc/cpuinfo
processor : 0
model name : ARMv7 Processor rev 3 (v7l)
BogoMIPS : 270.00
Features : half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm crc32
CPU implementer : 0x41
CPU architecture: 7
CPU variant : 0x0
CPU part : 0xd08
CPU revision : 3

https://static.docs.arm.com/ddi0501/f/DDI0501F_cortex_a53_cryptography_trm.pdf

1.1 About the Cortex-A53 processor Cryptography Extension

The Cortex-A53 processor Cryptography Extension supports the ARMv8 Cryptography
Extensions. The Cryptography Extensions add new A64, A32, and T32 instructions to
Advanced SIMD that accelerate Advanced Encryption Standard (AES) encryption and
decryption, and the Secure Hash Algorithm (SHA) functions SHA-1, SHA-224, and SHA-256.

Note
The optional Cryptography Extension is not included in the base product.
ARM supplies the Cryptography Extension only under an additional licence to the Cortex-A53 processor and Advanced SIMD and Floating

 

[RMerlin]

pi series has no aes acceleration.
It need a additional license.

Was wondering about that. Someone told me that he thought that it was always included by default when your CPU had am armv8 ISA licence, I guess that's not always the case then.

 

[tanoanian]

New firmware up this morning through the update function. No release notes yet.

New version is 3.0.0.4.384_81444

 

[mbze430]

Have ASUS or Merlin release a firmware for the AX89X with WAN Aggregation on 2x 1Gb ports? I cancelled my order on the AX89X because my AX88U is having issue with error message "kernel: protocol 0800 is buggy, dev eth1" and didn't want to get the AX89X if it is going to have the same issue.

 

[RMerlin]

Have ASUS or Merlin release a firmware for the AX89X with WAN Aggregation on 2x 1Gb ports?

I don't support the RT-AX89X.

 

[juched]

Haha @RMerlin, just occasionally noticed that the above is your 32768th (1000000000000000) post

Love this.

How does the old joke go.

There are 10 types of people in this world. Those that get binary and those that don’t.

 

[Starrbuck]

New firmware up this morning through the update function. No release notes yet.

New version is 3.0.0.4.384_81444

Version 3.0.0.4.384.81444
2020/05/04 48.97 MBytes

ASUS RT-AX89X Firmware version 3.0.0.4.384.81444
- Add Game page & Open NAT
- Fix LED issues
- Add smart connect
- Fix AiMesh backhaul disconnect issue
- Fix 160MHz unstable issue
- Update OFDMA GUI

 

[rodexo]

Mine went back. I wasted to much time trying to workout its quirks. Bye for now!


Sent from my iPhone using Tapatalk