New custom firmware for Asus RT-N56U!

Qubicfactor

Supremacy Member
Joined
Mar 26, 2003
Messages
7,628
Reaction score
659
ok. am going back to default Asus firmware. this firmware has 0.5 to 1ms better ping, but when i torrent the upload stays stuck around 380-450KB/sec.

changing back to default firmware i get my usual 800KB/sec - 1MB/sec seeding speed on my fibre.

hmm, I noted the same pattern too, with the 400KBps upload ceiling for torrents. then again, this 1.0.1.8ep1 is the first to use a different kernel, so maybe there are some bugs brought over.

will try to tweak to see if any padavan default parameters are causing it, otherwise it is a bug worth reporting.

ps: if you can, test the previous 1.0.1.8dp3 and see if it occurs; I dont torrent that often, and its not often i get to see people leeching off me enough to see a ceiling like now.
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
hmm, I noted the same pattern too, with the 400KBps upload ceiling for torrents. then again, this 1.0.1.8ep1 is the first to use a different kernel, so maybe there are some bugs brought over.

will try to tweak to see if any padavan default parameters are causing it, otherwise it is a bug worth reporting.

thanks for the verification bro! look forward to your findings
 

Qubicfactor

Supremacy Member
Joined
Mar 26, 2003
Messages
7,628
Reaction score
659
thanks for the verification bro! look forward to your findings

I left a file seeding, and it seems the 400kBps pattern is indeed there, but there are times it can be exceeded, probably with more than 1 client leeching (I hit 600kBps to 1MBps for a period around 5 mins with no visible ceiling). I am guessing the firmware ends up restricting a single client/upload connection to 400kBps, since I noted that the pattern appears again after that.

Too lazy to upload a screencap, but general upload pattern I got is a sawtooth with a distinctive visible ceiling at 400KBps. Need to monitor more.
 

misato

Supremacy Member
Joined
Jan 1, 2000
Messages
6,148
Reaction score
0
I left a file seeding, and it seems the 400kBps pattern is indeed there, but there are times it can be exceeded, probably with more than 1 client leeching (I hit 600kBps to 1MBps for a period around 5 mins with no visible ceiling). I am guessing the firmware ends up restricting a single client/upload connection to 400kBps, since I noted that the pattern appears again after that.

Too lazy to upload a screencap, but general upload pattern I got is a sawtooth with a distinctive visible ceiling at 400KBps. Need to monitor more.
I had also tested this as well and was able to hit 1MB/s upload consistently (50minutes sustained, 5minutes sample time) when there is enough peers leeching from me. Then again mine isn't a high traffic torrent, so its not easy for me to observe the behavior in shorter interval.
 

Qubicfactor

Supremacy Member
Joined
Mar 26, 2003
Messages
7,628
Reaction score
659
I had also tested this as well and was able to hit 1MB/s upload consistently (50minutes sustained, 5minutes sample time) when there is enough peers leeching from me. Then again mine isn't a high traffic torrent, so its not easy for me to observe the behavior in shorter interval.

I used an old batch torrent of some 9GB total size and left it downloading since morning, and I hit an upload spike of 1.6MBps, averaging somewhere around 800kBps, and the 400kBps ceiling is not visible at all. I guess the overall effects are not as bad as I thought it was, unless you had a very specific seeding regime to do, and in specific swarm conditions. I guess I would still stick to this firmware, since the behaviour is not significant to me overall.
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
I used an old batch torrent of some 9GB total size and left it downloading since morning, and I hit an upload spike of 1.6MBps, averaging somewhere around 800kBps, and the 400kBps ceiling is not visible at all. I guess the overall effects are not as bad as I thought it was, unless you had a very specific seeding regime to do, and in specific swarm conditions. I guess I would still stick to this firmware, since the behaviour is not significant to me overall.

thanks for all your observations! but how do you find this firmware performs consistently better than stock?
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
I used an old batch torrent of some 9GB total size and left it downloading since morning, and I hit an upload spike of 1.6MBps, averaging somewhere around 800kBps, and the 400kBps ceiling is not visible at all. I guess the overall effects are not as bad as I thought it was, unless you had a very specific seeding regime to do, and in specific swarm conditions. I guess I would still stick to this firmware, since the behaviour is not significant to me overall.

using the latest p2 firmware.

i'm trying now with 59 torrents being seeded :p (don't ask).

still has the 400KB/sec ceiling..

the speeds u both saw were after leaving it overnight yes? perhaps i'll do that.

could it be anything to do with the TCP timeout as previously mentioned? or perhaps I had my seeding terminal set as Low Priority under Network Map in the original firmware?
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
did more intense testing.

yes padavan's firmware DOES allow you to breach 400KB/sec on occasion, but not all the time. whereas on stock firmware i maintained 1-1.5MB/sec upload throughout. so i believe there is definitely an issue there.

did the old version of padavan 1018d have the bt upload speed bug?
 

Qubicfactor

Supremacy Member
Joined
Mar 26, 2003
Messages
7,628
Reaction score
659
did more intense testing.

yes padavan's firmware DOES allow you to breach 400KB/sec on occasion, but not all the time. whereas on stock firmware i maintained 1-1.5MB/sec upload throughout. so i believe there is definitely an issue there.

did the old version of padavan 1018d have the bt upload speed bug?

I *think* it didnt, but maybe I just didnt notice since I normally dont seed at your intensity level. If you were to run a test and it's fine, then perhaps the issue lies with the new linux kernel that was used in the e variants of the padavan firmware.
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
I *think* it didnt, but maybe I just didnt notice since I normally dont seed at your intensity level. If you were to run a test and it's fine, then perhaps the issue lies with the new linux kernel that was used in the e variants of the padavan firmware.

cool. i'll try it out and report. i've reduced my seeding torrents to 19 now. can you recommend a version of the firmware to try?
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
omg i realized i was on Software QoS and not Hardware NAT all this time =.=... that affects the router detrimentally in terms of performance right
 

Qubicfactor

Supremacy Member
Joined
Mar 26, 2003
Messages
7,628
Reaction score
659
omg i realized i was on Software QoS and not Hardware NAT all this time =.=... that affects the router detrimentally in terms of performance right

haha obviously it does.

Try running a clean setup (without config restoration) with padavan's latest if you have time, then again with 1.0.1.8dp3 and the stock 1.0.1.8e.

Don't forget to do a hardware reset (hold reset for 5 secs) after every reflash to avoid unpredictable outcomes from left over data.

Btw, it is also said that Asus original firmware is not able to run QoS and Hardware NAT at the same time. Padavan's firmware somehow managed to get both working.
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
haha obviously it does.

Try running a clean setup (without config restoration) with padavan's latest if you have time, then again with 1.0.1.8dp3 and the stock 1.0.1.8e.

Don't forget to do a hardware reset (hold reset for 5 secs) after every reflash to avoid unpredictable outcomes from left over data.

Btw, it is also said that Asus original firmware is not able to run QoS and Hardware NAT at the same time. Padavan's firmware somehow managed to get both working.


Cool stuff. Someone on the padavan google page recommended I use firmware 3.0.0.3.. What's that siol! It's very flashy and stuff but seems super clunky
 

Qubicfactor

Supremacy Member
Joined
Mar 26, 2003
Messages
7,628
Reaction score
659
oh the guy replied. apparently it's beta firmware not released to public yet. but aiya Padavan seems better. :p

have u tried HW NAT on LAN/WIFI? (Experimental)

He must be talking about the new black GUI firmware the N66U has. It's apparently being ported over to the N56U by Asus. Of course surface bling cannot beat core optimisations lol. Padavan's firmware will likely port the GUI over once Asus has it ready anyway so no worries if you want best of both worlds, or simply have N66U envy.

HW NAT on LAN/WIFI seems to make no improvements other than to increase pings back to what they were for wired devices. Wireless devices don't seem to see any notable improvements. I left it at the default HW NAT for LAN. Software NAT I used boost for PPTP/L2TP (though I have yet to see the difference since I dont run any VPNs atm.
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
He must be talking about the new black GUI firmware the N66U has. It's apparently being ported over to the N56U by Asus. Of course surface bling cannot beat core optimisations lol. Padavan's firmware will likely port the GUI over once Asus has it ready anyway so no worries if you want best of both worlds, or simply have N66U envy.

HW NAT on LAN/WIFI seems to make no improvements other than to increase pings back to what they were for wired devices. Wireless devices don't seem to see any notable improvements. I left it at the default HW NAT for LAN. Software NAT I used boost for PPTP/L2TP (though I have yet to see the difference since I dont run any VPNs atm.

increase pings back to what they were meaning your Wired pings become higher? gee that's not good.
 

Qubicfactor

Supremacy Member
Joined
Mar 26, 2003
Messages
7,628
Reaction score
659
increase pings back to what they were meaning your Wired pings become higher? gee that's not good.

Well, which is why they labeled it as experimental.

Anyway, makes sense since the cpu now has to accelerate wifi traffic as well in theory, though unfortunately doing so seemingly brings no benefit to wifi at the moment.
 

wohouboy

Banned
Joined
Aug 19, 2005
Messages
3,824
Reaction score
0
Well, which is why they labeled it as experimental.

Anyway, makes sense since the cpu now has to accelerate wifi traffic as well in theory, though unfortunately doing so seemingly brings no benefit to wifi at the moment.

noted. i'm still seeing a ceiling on my torrent seeding, i'm thinking of trying the 1.0.1.8.d p3 firmware for padavan. did you have any major bugs on that version?
 
Important Forum Advisory Note
This forum is moderated by volunteer moderators who will react only to members' feedback on posts. Moderators are not employees or representatives of HWZ Forums. Forum members and moderators are responsible for their own posts. Please refer to our Community Guidelines and Standards and Terms and Conditions for more information.
Top