What's new

[Release] FreshJR Adaptive QOS (Improvements / Custom Rules / and Inner workings)

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

Status
Not open for further replies.
I understand now thanks, I wonder if HSBC packet scheduler would help in anyway?

I have not read about that one but queuing disciplines have a lot more impact than schedulers.

Htb is a solid allocation method. No ideas what more you can want from it
 
@FreshJR I hope no discontinuing happens to the fast version. I noticed with 86u it is slightly faster than both alternative and compatible versions at least this is my experience.
 
@FreshJR I hope no discontinuing happens to the fast version. I noticed with 86u it is slightly faster than both alternative and compatible versions at least this is my experience.

It being faster than the “compatible” version is a placebo, as the end configurations are exactally the same.

It is possible to be faster than the “alternate version” as that end configuration is significantly different, but the alternative version works with all connections types which is a good trade off, even if it ends up being less cpu efficient.
 
@FreshJR I hope no discontinuing happens to the fast version. I noticed with 86u it is slightly faster than both alternative and compatible versions at least this is my experience.
There is no difference between the compatible and fast version in regards to performance.
 
To clarify, those of us with working configurations using the "fast version" can leave it in place? However, I suppose, we'll be on our own to debug issues that may come from future firmware updates?
 
To clarify, those of us with working configurations using the "fast version" can leave it in place? However, I suppose, we'll be on our own to debug issues that may come from future firmware updates?

Yup, you can leave it in place.

Correct, future firmware updates might break it at some point down the line and I am not looking to maintain it anymore as having multiple versions leads to confusion.
 
please include your alternative script with pppoe, vdsl rules

That is the only version available and is what installed via the one-line install with no configuration required for RMerlin users.

I did want to upload it to the first post for users wishing to make modifications, but I was having issues.

Either uploading text scripts got disabled, or the forum was having temporary issues.

If it’s disabled, a github link will be provided.
 
Last edited:
Can someone help me understand what I should expect to see when using FreshJR QOS and DSL speed test.

I am on cable internet with published speeds of 250/25. My speeds vary between 175-260. I set my WebUI speed limits to 225.23. When I turn off all QOS and run DSL speed test I get 175-260 speed an A for quality and a C for buffer bloat however when I turn on FreshJR QOS and run speed test I get 23-25 for speed C for quality and A+ for bufferbloat. I am confused why - quality and speed decease so drastically. I have tried increasing WebUI upload speed, re-arranging groups etc. Could someone please help me first understand what I am seeing (what does Quality mean, Why is bandwidth limited to 25, etc.) and understand if I can do anything to increase my quality or verify this is expected.
 
what does Quality mean

Why is bandwidth limited to 25, etc.

Can someone help me understand what I should expect to see when using FreshJR QOS and DSL speed test.

To perform limiting, QOS throws away a few initial packets so TCP can negotiate at a lower transfer rate.

If you throw away too many packets (happens when speed is set too low from your actual connection) you get poor quality. Quality describes how many packets were thrown away, with its full description available on dslreports.

Bandwidth is limited to 25mbps since you most likely input the numbers backwards.

The correlation between bufferbloat and quality was already mentioned as part of the install instructions!

Shoot for A’s, but YMMV
 
To perform limiting, QOS throws away a few initial packets so TCP can negotiate at a lower transfer rate.

If you throw away too many packets (happens when speed is set too low from your actual connection) you get poor quality. Quality describes how many packets were thrown away, with its full description available on dslreports.

Bandwidth is limited to 25mbps since you most likely input the numbers backwards.

The correlation between bufferbloat and quality was already mentioned as part of the install instructions!

Shoot for A’s, but YMMV

First let me say thank you for your patients and responding without making me feel as foolish as I am. The key that you mentioned is that I input the numbers backwards. Once you pointed this out everything fell into place, and more, in regards to understanding how it all fits together. A's across the board now. Thank you so much for your help.
 
To perform limiting, QOS throws away a few initial packets so TCP can negotiate at a lower transfer rate.

If you throw away too many packets (happens when speed is set too low from your actual connection) you get poor quality. Quality describes how many packets were thrown away, with its full description available on dslreports.

Bandwidth is limited to 25mbps since you most likely input the numbers backwards.

The correlation between bufferbloat and quality was already mentioned as part of the install instructions!

Shoot for A’s, but YMMV

Since your doing away with the fast version does that mean merlin can remove the modified part of the FW that allowed it and possibly get rid of some of those errors everyone complains about? Or was that just the 2nd file needed for the fast version?

Like the TC intercept mod or whatever it was :)
 
Since your doing away with the fast version does that mean merlin can remove the modified part of the FW that allowed it and possibly get rid of some of those errors everyone complains about? Or was that just the 2nd file needed for the fast version?

No. That file is required.

Fast version had a modified version of RMerlins file present but now RMerlins default version will be in use
 
Last edited:
ahh kk gotcha.. was excited there for a sec :)
... it offers better performance so everyone should hope RMerlin maintains it as long as he can

It is what offers you your additional queuing disciplines and also the framework to do much more.
 
Last edited:
I haven't decided what im doing yet.. I have the latest compatible modified and im happy with it. I also haven't tried 384.6. holding out for something truly great lol
 
@Freshj i think its a placebo effect, but it seems i dont lag at all when using the fast version compared to the new alt & compatible version. Maybe im gng krazy haha
 
@Freshj i think its a placebo effect, but it seems i dont lag at all when using the fast version compared to the new alt & compatible version. Maybe im gng krazy haha

I've tweaked various things over the years, believing they did something useful, only to have people explain how it had nothing to do with anything important.
 
@Freshj i think its a placebo effect, but it seems i dont lag at all when using the fast version compared to the new alt & compatible version. Maybe im gng krazy haha

The only diff between the two versions is how the script is applied.. with the fast version theres that intercept file that runs inplace of whats standard in the firmware allowing for imediate changes etc to qos right after u hit apply. The compatable has a 5 min delay due to having to follow that standard process plus aditional time was added to the delay for compatibility betwwen slower routers and people running 15 addons that need to load 1st. Otherwise theyre exactly the same. When the script is actually executed the cpus hit close to 100% for 1-2secs either imediately with the fast version or 5min after you hit apply with the compatable version =)
 
@FreshJR i think its a placebo effect

There is no placebo effect, only your mind self-deceives you by the name Fast (speed, without limit, extreme, etc) and when you read compatible you think: it is not complete, lag, slow, limited, etc.

My advice is trust the creator of the script and do not be fooled by the names.

@Sinner is right
The only diff between the two versions is how the script is applied...

and the fast version has a huge bug that does not exist in the compatible version.
Compatible > Fast (now it is no longer called compatible, it is called FreshJR_QOS)

@FreshJR but it seems i dont lag at all when using the fast version compared to the new alt & compatible version. Maybe im gng krazy haha
if you have lag I am sure that the culprit is "Snapchat", because it is in "others" and that category is used for games or you do not have the category "Others" higher than the others, but below the category "Gaming" and it's not because of the compatible version.
 
Last edited:
Status
Not open for further replies.

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top