From: "Typical M$ User" 

X-No-Archive: yes
> Can you run your tests again, this time use http://speedtest.nls.net and
the
> first test on that site? That machine has an RWIN of 64240. So maybe try
> multiples of that?
>
> Geo.

I just ran the tests, and here are the results.

2/7/4

RWIN=64240, 8:20PM Central:
 Test 1 - 1 MB data    = 2804.1 Kbps - 343.6 K bytes/sec
 Test 2 - 100 KB data  = 1854.5 Kbps - 227.3 K bytes/sec
 Test 3 - 18.2 MB file = 362 KB/sec

RWIN=513920, 8:44PM Central:
 Test 1 - 1 MB data    = 8242.4 Kbps - 1010.1 K bytes/sec
 Test 2 - 100 KB data  = 3022.2 Kbps - 370.4 K bytes/sec
 Test 3 - 18.2 MB file = 362 KB/sec

RWIN=256960, 8:50PM Central:
 Test 1 - 1 MB data    = 2914.3 Kbps - 357.1 K bytes/sec
 Test 2 - 100 KB data  = 2914.3 Kbps - 357.1 K bytes/sec
 Test 3 - 18.2 MB file = 361 KB/sec

RWIN=128480, 8:57PM Central:
 Test 1 - 1 MB data    = 2978.1 Kbps - 365 K bytes/sec
 Test 2 - 100 KB data  = 2914.3 Kbps - 357.1 K bytes/sec
 Test 3 - 18.2 MB file = 361 KB/sec

RWIN=17520, 9:03PM Central:
 Test 1 - 1 MB data    = 2153 Kbps - 263.9 K bytes/sec
 Test 2 - 100 KB data  = 2092.3 Kbps - 256.4 K bytes/sec
 Test 3 - 18.2 MB file = 251 KB/sec

==========================================
I was curious about the values in the 1st 2 tests, so I ran those again.

RWIN=64240, 9:10PM Central:
 Test 1 - 1 MB data    = 3022.2 Kbps - 370.4 K bytes/sec
 Test 2 - 100 KB data  = 3022.2 Kbps - 370.4 K bytes/sec
 Test 3 - 18.2 MB file = 360 KB/sec

RWIN=513920, 9:17PM Central:
 Test 1 - 1 MB data    = 2598.7 Kbps - 318.5 K bytes/sec
 Test 2 - 100 KB data  = 3022.2 Kbps - 370.4 K bytes/sec
 Test 3 - 18.2 MB file = 357 KB/sec
==========================================
I didn't like those values for 513920, so I ran the 1st test a few more times.

RWIN=513920, 9:21PM Central:
 Test 1 - 1 MB data    = 2756.8 Kbps - 337.8 K bytes/sec
 Test 1 - 1 MB data    = 2914.3 Kbps - 357.1 K bytes/sec
 Test 1 - 1 MB data    = 2967.3 Kbps - 363.6 K bytes/sec
 Test 1 - 1 MB data    = 2978.1 Kbps - 365 K bytes/sec
==========================================
-Overall, the best results I got were with RWIN=513920 AND RWIN=64240, so it
looks like a really high RWIN value on my side doesn't really matter after all.
-3022.2 Kbps, 370.4 K bytes/sec was the cap for me.
-Of course, improvements were obvious with any RWIN value over the default.
-It's interesting how in a couple of test sessions, I got the same results
in both Test1 and Test2.
-What do you think?

--- BBBS/NT v4.01 Flag-5