Post Reply 
(RWIN) = 32768
Jun. 19, 2024, 02:05 PM
Post: #16
RE: (RWIN) = 32768
Update: So I used a good backup folder I had of Proxomitron that was working perfectly when I saved it, updated it to the latest ProxomitronReborn_4701R.zip, etc, and tested it WITHOUT RcvBuf = 0 in the default.cfg (with both my TcpWindowSize and GlobalMaxTcpWindowSize are set to RWIN 1027840 in the OS registry) and the results were:

Default TCP Receive Window (RWIN) = 65536
RWIN Scaling (RFC1323) = 1 bits (scale factor: 2^1=2)
Unscaled TCP Receive Window = 32768

I then added RcvBuf = 0 to the default.cfg and the results were:

Default TCP Receive Window (RWIN) = 131072
RWIN Scaling (RFC1323) = 9 bits (scale factor: 2^9=512)
Unscaled TCP Receive Window = 256

By the way, I'm not trying to burden anyone with these results, lol. Just experimenting as I thought RcvBuf = 0 causes Proxomitron to use the default of the OS, which in my case is 1027840. Part of the problem too is that I don't know what these analyzer results (131072 and 256) mean. I would like to test my RWIN at another website, but don't know of one. Thank you.
Add Thank You Quote this message in a reply
Jun. 19, 2024, 02:57 PM
Post: #17
RE: (RWIN) = 32768
(Jun. 19, 2024 12:52 PM)Anno Domini Wrote:  Are you referring to proxo.exe in ProxomitronReborn_4701R.zip ?

Yes, https://prxbx.com/forums/showthread.php?tid=2331&pid=19290#pid19290 .
Add Thank You Quote this message in a reply
Jun. 19, 2024, 02:59 PM
Post: #18
RE: (RWIN) = 32768
(Jun. 19, 2024 02:05 PM)Anno Domini Wrote:  Update: So I used a good backup folder I had of Proxomitron that was working perfectly when I saved it, updated it to the latest ProxomitronReborn_4701R.zip, etc, and tested it WITHOUT RcvBuf = 0 in the default.cfg (with both my TcpWindowSize and GlobalMaxTcpWindowSize are set to RWIN 1027840 in the OS registry) and the results were:

Default TCP Receive Window (RWIN) = 65536
RWIN Scaling (RFC1323) = 1 bits (scale factor: 2^1=2)
Unscaled TCP Receive Window = 32768

I then added RcvBuf = 0 to the default.cfg and the results were:

Default TCP Receive Window (RWIN) = 131072
RWIN Scaling (RFC1323) = 9 bits (scale factor: 2^9=512)
Unscaled TCP Receive Window = 256

Ah, "WITHOUT RcvBuf = 0" "Default TCP Receive Window (RWIN) = 65536"...

Something, other than the Proxomitron, is doubling the RWIN Scaling factor, since your Firefox seems to show doubling when used without the Proxomitron.

Have you tested with another browser?
Have you 'tweaked' your operating system? Which OS are you using?
Add Thank You Quote this message in a reply
Jun. 20, 2024, 02:00 PM
Post: #19
RE: (RWIN) = 32768
Yes, without RcvBuf = 0" "Default TCP Receive Window (RWIN) = 65536. At the risk of opening myself up to criticism, I'm still using Windows 7 Professional (64-bit). I had/have it setup exactly as I want with freeware and a rules based firewall where I can control outbound traffic, and never made the switch to another operating system as I just use it to read the news, visit forums like this, and check email. Honestly, I don't want to switch to another OS until I absolutely have to. I only have Firefox installed, but I have several different Firefox versions installed. All of them show the same result (RWIN = 65536) without RcvBuf = 0, and I tested Firefox without using Proxomitron ('no proxy' in Firefox's connection settings) and it showed 'Default TCP Receive Window (RWIN) = 131072,' as seen below. I may have tweaked my operating system registry somewhere along the way, but I don't know WHERE to look to determine WHAT, other than the Proxomitron, would be doubling the RWIN Scaling. I did recently use TCPOptimizer_v308 though, and followed the directions in the video below from the 4:20 minute mark to the 10 minute mark to set TCPOptimizer and my adapter up. Does the TCP Receive Window (RWIN) = 131072 give any clues ?

Default TCP Receive Window (RWIN) = 131072
RWIN Scaling (RFC1323) = 9 bits (scale factor: 2^9=512)
Unscaled TCP Receive Window = 256

How to Optimize Internet Adapter Settings to Lower Ping and Increase Internet Speeds For Gaming
https://www.youtube.com/watch?v=r4tMBS3Wy8A
Add Thank You Quote this message in a reply
Jun. 21, 2024, 04:03 AM
Post: #20
RE: (RWIN) = 32768
(Jun. 20, 2024 02:00 PM)Anno Domini Wrote:  ...At the risk of opening myself up to criticism,...

The typical Proxomitron user doesn't think like that...
If you are ok with your computer's current performance, don't spend much time or worry on this.
I'm just curious.

(Jun. 20, 2024 02:00 PM)Anno Domini Wrote:  ... I did recently use TCPOptimizer_v308 though, and followed the directions in the video below from the 4:20 minute mark to the 10 minute mark to set TCPOptimizer and my adapter up. Does the TCP Receive Window (RWIN) = 131072 give any clues ?

I didn't spot anything in that video that would cause this.

(Jun. 20, 2024 02:00 PM)Anno Domini Wrote:  I may have tweaked my operating system registry somewhere along the way, but I don't know WHERE to look to determine WHAT, other than the Proxomitron, would be doubling the RWIN Scaling. ...

I have a theory. At some point a value was added to the registry that is larger than the value that Win7 expects. So Win7 doubles the scaling to provide for the larger number.

Like https://learn.microsoft.com/en-us/troubleshoot/windows-server/networking/description-tcp-features#windows-scaling

microsoft Wrote:For example:

If the window size in the registry is entered as 269000000 (269M) in decimal, the scaling factor during the three-way handshake is 13. A scaling factor of 12 only allows a window size up to 268,431,360 bytes (268M).

The initial window size in this example would be calculated as follows:
65,535 bytes with a window scale factor of 13.
True window size = 65535*2^13
True window size = 536,862,720

When the value for window size is added to the registry, and its size is larger than the default value, Windows attempts to use a scale value that accommodates the new window size.
Add Thank You Quote this message in a reply
Jun. 21, 2024, 06:54 AM (This post was last modified: Jun. 21, 2024 06:58 AM by ProxRocks.)
Post: #21
RE: (RWIN) = 32768
Personally, I've used this "snake oil" app call "TCPOptimizer" in the past (WinXP) and it was an utter and complete WASTE OF TIME with ZERO speed difference.
A sugar pill offers the same PLACEBO EFFECT.
Add Thank You Quote this message in a reply
Jun. 21, 2024, 09:12 PM (This post was last modified: Jun. 21, 2024 09:12 PM by Anno Domini.)
Post: #22
RE: (RWIN) = 32768
(Jun. 21, 2024 06:54 AM)ProxRocks Wrote:  Personally, I've used this "snake oil" app call "TCPOptimizer" in the past (WinXP) and it was an utter and complete WASTE OF TIME with ZERO speed difference.
A sugar pill offers the same PLACEBO EFFECT.

You could be right, ProxRocks. It's hard to tell if it actually helps increase speed, at least for me it is. Although, today while testing different RcvBuf values in Proxomitron's default.cfg, I was switching the 'TCP Window Auto-Tuning Level' from normal (enabled) to disabled in TCPOptimizer (see attached pic), and I noticed that when it was enabled my download speed at https://testmy.net/download was consistently faster than when it was disabled. That said, I would have to test it MORE to be dogmatic about any speed increase. Btw, I'm not defending TCPOptimizer, just an observation. :-) Also, for whatever reason, enabling and disabling 'TCP Window Auto-Tuning Level' changed the RWIN test results at https://www.speedguide.net/analyzer.php. More on that after the jump.


Attached File(s)
.jpg  TCP Window Auto-tuning .JPG (Size: 75.08 KB / Downloads: 26)
Add Thank You Quote this message in a reply
Jun. 21, 2024, 10:15 PM (This post was last modified: Jun. 22, 2024 02:47 PM by Anno Domini.)
Post: #23
RE: (RWIN) = 32768
(Jun. 21, 2024 04:03 AM)JJoe Wrote:  
(Jun. 20, 2024 02:00 PM)Anno Domini Wrote:  ...At the risk of opening myself up to criticism,...

The typical Proxomitron user doesn't think like that... If you are ok with your computer's current performance, don't spend much time or worry on this. I'm just curious.

Thank you for making this point, and I have the same impression about most Proxomitron users. :-)

Based on what you said about how possibly at some point a value was added to the registry that is larger than the value that Win7 expects -- and perhaps Win7 doubles the scaling to provide for the larger number -- I did some more reading and testing, and I won't deny it, I don't know what any of these results below mean, but I'll post it and maybe there are some clues here that someone can make sense of it. Having said that, I currently have RcvBuf = 0 in Proxomitron's default.cfg and I think I'll leave it there for a while and see how my download speeds and surfing unfold.

I'll preface the analyzer results below by noting that earlier I posted a pic showing I had set the TcpWindowSize and GlobalMaxTcpWindowSize in Win7's registry to 1027840, but apparently these registry keys have no effect in Win7. You can't control the Tcp Window Size with them in Win7. They worked for XP, but it seems Win 7 has the ability to tune individual connections for best performance. I don't know much more about it, but this post about Vista below explains more. I think it applies to Win7 too.

https://msfn.org/board/topic/87969-heres...-in-vista/

The speedguide results are as follows. Note results 1 and 2 are with TCP Windows Auto-Tuning DISABLED in TCPOptimizer, and test 3 and 4 have TCP Windows Auto-Tuning ENABLED. This can be seen by typing netsh int tcp show global in a command prompt. I also included RcvBuf = 0 in some tests, and excluded it for others. See attached photos.

1.) TCP Windows Auto-Tuning is disabled WITHOUT RcvBuf = 0 in default.cfg

Default TCP Receive Window (RWIN) = 32768
RWIN Scaling (RFC1323) = 0 bits
Unscaled TCP Receive Window = 32768

2.) TCP Windows Auto-Tuning is disabled WITH RcvBuf = 0 in default.cfg

Default TCP Receive Window (RWIN) = 64240
RWIN Scaling (RFC1323) = 0 bits
Unscaled TCP Receive Window = 64240

3.) TCP Windows Auto-Tuning is enabled WITHOUT RcvBuf = 0 in default.cfg

Default TCP Receive Window (RWIN) = 65536
RWIN Scaling (RFC1323) = 1 bits (scale factor: 2^1=2)
Unscaled TCP Receive Window = 32768

Notice in 3rd result that 32768 x 2 = 65536

4.) TCP Windows Auto-Tuning is enabled WITH RcvBuf = 0 in default.cfg

Default TCP Receive Window (RWIN) = 131072
RWIN Scaling (RFC1323) = 9 bits (scale factor: 2^9=512)
Unscaled TCP Receive Window = 256

Notice in 4th result that 65536 in result number 3 x 2 = 131072. I have no idea why the Unscaled TCP Receive Window = 256.

Well, I don't know what to make of these results, but I think I'll leave well enough alone. As mentioned earlier, I currently have RcvBuf = 0 in the default.cfg, and I'm getting about 1 to 2 Mbps faster download speed than advertised by my ISP. So I will monitor it over time, and see how it goes. I just don't understand the doubling RWIN values, lol. Thank you.


Attached File(s)
.jpg  Receive Window Auto-tuning DISABLED .JPG (Size: 47.81 KB / Downloads: 23)
.jpg  Receive Window Auto-tuning ENABLED .JPG (Size: 47.81 KB / Downloads: 22)
.jpg  TCP Window Auto-tuning .JPG (Size: 75.08 KB / Downloads: 21)
Add Thank You Quote this message in a reply
Jun. 22, 2024, 02:42 AM
Post: #24
RE: (RWIN) = 32768
Without RcvBuf = 0, Proxomitron Reborn sets it to 32768 like previous versions of Proxomitron do.

When RcvBuf is set to 0, it leaves it up to the OS to decide.
Add Thank You Quote this message in a reply
Jun. 22, 2024, 02:34 PM
Post: #25
RE: (RWIN) = 32768
(Jun. 22, 2024 02:42 AM)amy Wrote:  Without RcvBuf = 0, Proxomitron Reborn sets it to 32768 like previous versions of Proxomitron do.

When RcvBuf is set to 0, it leaves it up to the OS to decide.

Thank you, Amy. I'm just going to leave RcvBuf = 0 in the default.cfg and see how it affects my Internet speed for the near future. It's currently 1-2 mbps higher than what my ISP advertises, which is good, and I'm content with that. I don't know why my RWIN value doubles on some Speedguide tests, or how to manually set the RWIN on Win 7. Apparently it's dynamic and Win 7 doesn't have registry keys like XP that can be tweaked, but I've decided to leave well enough alone, lol. I'm not done researching on this, but my Internet speed is fine at this point. Thank you ALL for your input !
Add Thank You Quote this message in a reply
Post Reply 


Forum Jump: