EDITED TO DEFINE VALUES.  INCLUDES ADDITIONAL TWEAKS
Do a soft reset after making registry changes.

CORRECTED Value Name to EnablePMTUDiscovery

I made tweaks to my Windows 98 and XP machine to improve my computers internet speed.  Today I run Vista Ultimate and do well without them.  I thought, I wonder if the same can be done with the Mogul...well WM6.  I made the two registry changes using Total Commander (awesome by the way).

A noticable improvement to me was seen.  Below are the tweaks and my speed afterwards.  I will say that tweaks found on this site were done to the registry as well.  BUT, I ran IE afterwards with the same results.  So I thought of this.  Have fun.  Copy and Paste the table below if you were successful and want to post your results.

HKLM\Comm\Tcpip\Parms\
TcpMaxConnectRetransmissions = dword:  6
TcpWindowSize = dword:  262144

I went to Http://text.dslreports.com/mspeed and got a download result:

Speed Test Results

Network	Sprint	WiFi
kbit/sec	820	1219
latency sec	0.405	0.103
d/l time sec	10.25	6.819
File Size	1MB	1MB
System	EVDO R 0	Netgear

After adding more tweaks

HKCU\Software\Microsoft\Windows\CurrentVersion\Internet Settings
MaxConnectionsPer1_0Server = dword: 00000020  (Zero before Server)
MaxConnectionsPerServer = dword: 00000010

HKLM\Comm\Tcpip\Parms\
Tcp1323Opts = dword:  1  (O before pts as in Options)
TcpMaxDupAcks = dword:  2
EnablePMTUDiscovery = dword:  1

My results were:

Speed Test Results

Network	Sprint	WiFi
kbit/sec	1005	1235
latency sec	0.232	.111
d/l time sec	8.376	6.741
File Size	1MB	1MB
System	EVDO R 0	Netgear

Sorry but are you making new dword values?

Yes I am.  I tried those that were found online, that I've used before.

im not even getting nearly the speeds you are getting before your hacks.....i have not done any hacks. i live in an evdo area but the speed test only gives me like 350 kbit/sec is my device defective?

I am getting in the 700-850 range with no tweaks in evdo area.

Those numbers I posted are after the tweaks.  I just wanted to post what I found for my phone in my area after the tweaks.  I figured maybe it may work on the mogul if it worked on a PC.

I'm happy with the results. 

I did the tweaks and noticed a speed increase. (didn't do a test prior to the update)

My (post) tweak speed results are:

 

Speed Test Results

Network	Sprint	WiFi
kbit/sec	1044	1179
latency sec	0.464	.106
d/l time sec	8.098	7.054
File Size	1MB	1MB
System	EVDO R 0	Belkin Router

These tweaks definately made in improvement in my numbers pre tweak. Above are post tweak.

Sweet isn't it?  I wonder why I hadn't seen or thought of this before.  It either works for you or it doesn't. 

I can't wait for Rev A.   When that happens I will do the test post ROM update then post tweak.

You know with all the tweaks, programs, and this Seidio case.  I'm glad I kept my Mogul. 

I would like to get a contract free phone cheep to open up and see if the internal memory could be upgraded.  I modified an old Pro 92 scanner once adding memory, switchable, to allow a different memory groups to load on power up, shortwave, and cell phones.  That was in the day.  It was fun.

A 5mp camera and 512 of memory...woo hoo.  Or at least a virtual memory dll to use some of the memory card.. 

Just thinking is all.  Have fun with the tweak.

cmucooper said :

im not even getting nearly the speeds you are getting before your hacks.....i have not done any hacks. i live in an evdo area but the speed test only gives me like 350 kbit/sec is my device defective?

I doubt your Mogul is defective.  Signal quality is paramount on any radio device.  Particularly when its digital.  If your house has foil backed insulation or your in a mobile home with metal siding impacts performance. 

Additionally, the more programs you have running in the background consuming CPU cycles will also result in a slower bit rate.  If you do the test and your indicator flips to 1X then any of the above is suspect to include distance from the tower.  The QualComm 7500 is a dual processor, but I don't know enough about it to say how much in the background is handled while another function has primary control.

Try the test in another location.