Overclocking an Asus RT-AC88U

[Denna]

In the post below, @ryzhov_al showed there are pairings for CPU and RAM frequencies.

https://www.snbforums.com/threads/o...-56_2-rt-ac68u-and-rt-n66u.28043/#post-216213
​

Does anyone know what the pairings are for an RT-AC88U ?

 

[ryzhov_al]

https://github.com/RMerl/asuswrt-me...e/src-rt-7.14.114.x/src/shared/hndmips.c#L826

 

[Denna]

@ryzhov_al,

I don't see an entry for the BCM4709.

This router defaults to 1400 MHz CPU / 800 MHz RAM.

 

[john9527]

Save yourself the trouble....if you do a search you will find that ASUS disabled the ability to overclock on the AC88U. If you try, the router will cripple itself to a severely underclocked state.

 

[Denna]

@john9527,

Search where ?

There's no way to circumvent that ?​

 

[john9527]

Here.....snbforums
https://www.snbforums.com/threads/ac88u-overclock-strange-behavior.33401/#post-268788
https://www.snbforums.com/threads/ac-88u-openvpn-overclocking-issue.34836/#post-281995

 

[john9527]

There's no way to circumvent that ?

Not that I've seen. It appears to be tied to the CFE bootloader, which is now encrypted on the later routers.

 

[Denna]

@john9527,

In your post below, you mentioned valid pairings.

https://www.snbforums.com/threads/ac-88u-openvpn-overclocking-issue.34836/#post-281976
​

Where did you find this documented ?

I was using the 1600/800 pair and getting the result above.

After commenting out the nvram overclock commands and powering off / on, the same slow result happens.

Do I have to reload the firmware again ?​

 

[john9527]

The valid values were pulled from the source code....but as I said, they are no longer honored for the AC88.

To get things back to normal, you need to manually reset the value to the defaults
nvram set clkfreq=1400,800 (or nvram unset clkfreq)
nvram commit

then reboot.

 

[Denna]

@RMerlin,

You displayed the results of an openssl speedtest on an RT-AC88U in the post below.

https://www.snbforums.com/threads/req-need-some-openssl-numbers-info.31444/#post-249448
​

When I ran the same test on the same router, the following results are less than half of what you showed. Why ?

admin@RT-AC88U-0000:/tmp/home/root# openssl speed aes-128-cbc aes-256-cbc bf-cbc
Doing aes-128 cbc for 3s on 16 size blocks: 3547016 aes-128 cbc's in 3.00s
Doing aes-128 cbc for 3s on 64 size blocks: 949965 aes-128 cbc's in 3.00s
Doing aes-128 cbc for 3s on 256 size blocks: 244608 aes-128 cbc's in 3.00s
Doing aes-128 cbc for 3s on 1024 size blocks: 61500 aes-128 cbc's in 3.00s
Doing aes-128 cbc for 3s on 8192 size blocks: 7699 aes-128 cbc's in 3.00s
Doing aes-256 cbc for 3s on 16 size blocks: 2784445 aes-256 cbc's in 3.00s
Doing aes-256 cbc for 3s on 64 size blocks: 735159 aes-256 cbc's in 3.00s
Doing aes-256 cbc for 3s on 256 size blocks: 188170 aes-256 cbc's in 3.00s
Doing aes-256 cbc for 3s on 1024 size blocks: 47199 aes-256 cbc's in 3.00s
Doing aes-256 cbc for 3s on 8192 size blocks: 5909 aes-256 cbc's in 2.99s
Doing blowfish cbc for 3s on 16 size blocks: 4150520 blowfish cbc's in 3.00s
Doing blowfish cbc for 3s on 64 size blocks: 1143728 blowfish cbc's in 3.00s
Doing blowfish cbc for 3s on 256 size blocks: 288782 blowfish cbc's in 3.00s
Doing blowfish cbc for 3s on 1024 size blocks: 72597 blowfish cbc's in 3.00s
Doing blowfish cbc for 3s on 8192 size blocks: 9110 blowfish cbc's in 3.00s
OpenSSL 1.0.2k 26 Jan 2017
built on: reproducible build, date unspecified
options:bn(64,32) rc4(ptr,char) des(idx,cisc,2,long) aes(partial) blowfish(ptr)
compiler: arm-openwrt-linux-gnueabi-gcc -I. -I.. -I../include -fPIC -DOPENSSL_PIC -DZLIB_SHARED -DZLIB -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -I/media/ware4/Entware-ng.2017.02/staging_dir/target-arm_cortex-a9_glibc-2.23_eabi/opt/include -I/media/ware4/Entware-ng.2017.02/staging_dir/target-arm_cortex-a9_glibc-2.23_eabi/include -I/media/ware4/Entware-ng.2017.02/staging_dir/toolchain-arm_cortex-a9_gcc-6.3.0_glibc-2.23_eabi/usr/include -I/media/ware4/Entware-ng.2017.02/staging_dir/toolchain-arm_cortex-a9_gcc-6.3.0_glibc-2.23_eabi/include -DOPENSSL_SMALL_FOOTPRINT -DOPENSSL_NO_ERR -DTERMIOS -O2 -pipe -march=armv7-a -mtune=cortex-a9 -fno-caller-saves -fhonour-copts -Wno-error=unused-but-set-variable -Wno-error=unused-result -mfloat-abi=soft -fpic -I/media/ware4/Entware-ng.2017.02/package/libs/openssl/include -fomit-frame-pointer -Wall
The 'numbers' are in 1000s of bytes per second processed.
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
blowfish cbc 22136.11k 24399.53k 24642.73k 24779.78k 24876.37k
aes-128 cbc 18917.42k 20265.92k 20873.22k 20992.00k 21023.40k
aes-256 cbc 14850.37k 15683.39k 16057.17k 16110.59k 16189.47k​

 

[Denna]

@john9527,

Running "nvram unset clkfreq" has the cpu and memory running at normal speeds. Thank you.​

 

[john9527]

Are you sure things are back to normal? Is the bogomips value back to around 2800?

 

[RMerlin]

When I ran the same test on the same router, the following results are less than half of what you showed. Why ?

Probably means you are not running at a valid clock rate. Check your bogomips. Incorrect overclocking attempts with downgrade the CPU clock.

 

[martinr]

Probably means you are not running at a valid clock rate. Check your bogomips. Incorrect overclocking attempts with downgrade the CPU clock.

Well, he was warned https://www.snbforums.com/threads/overclocking-an-asus-rt-ac88u.37809/#post-311101

 

[Denna]

@john9527 and @RMerlin ,

This is with no overclocking.

 

[john9527]

Yes, that's about 1/2 of expected. They may have added some other change besides just clkfreq. About the only other advice I can give right now is to do a factory reset.

 

[Denna]

@john9527,

After a factory reset.

​

Maybe if I keep factory resetting, the speed will keep doubling.

​

 

[john9527]

Much better

 

[bryantjopplin]

@john9527,

After a factory reset.

View attachment 8677
​

Maybe if I keep factory resetting, the speed will keep doubling.

​

I like the way you think, brb defaulting now. [emoji12]

Sent from my Nexus 6 with Pixel XL build.prop

 

[Denna]

Given that Asus has allowed / supported overclocking recent routers, it seems like this is a bug.

How do you report a bug like this ?