Windows TCP Settings - Follow up - Help with server communication


log in

Advanced search

Message boards : Number crunching : Windows TCP Settings - Follow up - Help with server communication

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 14 · Next
Author Message
Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8551
Credit: 50,424,923
RAC: 51,160
United Kingdom
Message 1344094 - Posted: 8 Mar 2013, 17:58:43 UTC - in response to Message 1344088.
Last modified: 8 Mar 2013, 18:01:47 UTC

This topic with all of its changes and such started 2 days ago and many have applied the changes to their machines, so the question is -- Has anyone noticed any problem with Synergy, georgem, and vader handling the extra load? The extra load being no drops, retries, etc.

Well, at risk of invoking the curse of the yellow fluffy thing, I've had a scout round Cricket and setistats.

The only bad thing on either set of graphs (that I can see) is the falling-off in the number of AP tasks ready to send over the last 9 hours or so - but it's OK, it's just that we've finished splitting all the current tapes on the SSP. And people have obviously been collecting them, so that's actually a good thing.

Downloads were obviously in extended backoff during this week's maintenance (just before this thread started), because we never dipped throughout the whole outage. It will be interesting to see next week whether we've had enough readers to make a difference next time round.

That's why I included Matt in my 'heads up' email, in case he saw something bad on the server and wanted to call us off. But I'd say, so far, it's all looking good.

Profile Cliff HardingProject donor
Volunteer tester
Avatar
Send message
Joined: 18 Aug 99
Posts: 992
Credit: 52,094,535
RAC: 43,768
United States
Message 1344096 - Posted: 8 Mar 2013, 18:03:56 UTC
Last modified: 8 Mar 2013, 18:06:50 UTC

Just modified my other machine, which does MBs only and it d/l'ed 81 tasks in 10 minutes. Went to get a drink and when I came back it was all gone!!!.

[edit] You guys must realize that we just may have created a MONSTER, that we may not be able to control once everybody gets on board.
____________


I don't buy computers, I build them!!

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8551
Credit: 50,424,923
RAC: 51,160
United Kingdom
Message 1344100 - Posted: 8 Mar 2013, 18:07:58 UTC - in response to Message 1344096.

Just modified my other machine, which does MBs only and it d/l'ed 81 tasks in 10 minutes. Went to get a drink and when I came back it was all gone!!!.

The landlord of my local pub commented that my drink did that last night, too. Cheers!

Profile Bill GProject donor
Avatar
Send message
Joined: 1 Jun 01
Posts: 347
Credit: 42,011,674
RAC: 74,013
United States
Message 1344104 - Posted: 8 Mar 2013, 18:15:20 UTC - in response to Message 1344101.

I just wanted to point out that the download site does not mention Win8.....I tried it on my Win8 computer and it does not work there.
____________

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8551
Credit: 50,424,923
RAC: 51,160
United Kingdom
Message 1344105 - Posted: 8 Mar 2013, 18:16:02 UTC - in response to Message 1344104.

I just wanted to point out that the download site does not mention Win8.....I tried it on my Win8 computer and it does not work there.

Which download site, which "it"?

Profile HAL9000
Volunteer tester
Avatar
Send message
Joined: 11 Sep 99
Posts: 4244
Credit: 116,005,824
RAC: 140,284
United States
Message 1344106 - Posted: 8 Mar 2013, 18:16:11 UTC - in response to Message 1344087.

OK, now we've had a good laugh, can we keep this thread clear for serious, hard, technical appraisal and constructive feedback, please?

[edit: sorry, not referring to you, Cliff - unfortunate timing there]

Well, I am sure that the option shall not be pushed in the next MS update...LOL.

But, one has to be curious why it was chosen not to enable it by default.
Does anybody know of a downside?
For example, does it limit transmission speeds when one is NOT dealing with a congested network connection?

The TimeStamp information adds two 4 byte tags for 8 bytes total in each 64k packet. So technically yes it actually does reduce the amount of data sent per packet.
However, if you are having data transmission issues it is advantageous to have a slight reduction in speed for data accuracy.

It is sort of like enabling Jumbo Packets. If you need to enable it then it is a gain. Otherwise it just slows things down.
____________
SETI@home classic workunits: 93,865 CPU time: 863,447 hours

Join the BP6/VP6 User Group today!

Profile TRuEQ & TuVaLu
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 473
Credit: 18,696,043
RAC: 34,496
Sweden
Message 1344110 - Posted: 8 Mar 2013, 18:33:06 UTC
Last modified: 8 Mar 2013, 18:35:07 UTC

I've done some dl testing without knowing of this thread.
And I've noticed increased speed and less stalled transfers the last couple of days.

I have not ran the optimizer.

I thought someone changed some setting on the servers....

ClaggyProject donor
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4101
Credit: 33,140,948
RAC: 7,717
United Kingdom
Message 1344111 - Posted: 8 Mar 2013, 18:33:27 UTC - in response to Message 1344069.

I don't think there's likely to be any way in which SETI, as a project, can access the client registry on Windows machines and make changes at that low level - and it would probably be roundly criticised as a security risk or a hacking exploit if it attempted any such thing. The one possibility might be to ask BOINC to include it into the BOINC installer, when the user is present and is explicitly giving permission for system changes to be made.

Could updating the Libcurl version supplied with Boinc do the same thing?

Libcurl Changelog

Claggy

Profile CLYDEProject donor
Volunteer tester
Avatar
Send message
Joined: 9 Aug 99
Posts: 2163
Credit: 23,352,718
RAC: 35,278
United States
Message 1344114 - Posted: 8 Mar 2013, 18:41:01 UTC - in response to Message 1344104.
Last modified: 8 Mar 2013, 18:42:28 UTC

I just wanted to point out that the download site does not mention Win8.....I tried it on my Win8 computer and it does not work there.

According to Speedguide.net Forums (TCP Optimizer): It doesn't yet work for Win8 :(
____________

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8551
Credit: 50,424,923
RAC: 51,160
United Kingdom
Message 1344115 - Posted: 8 Mar 2013, 18:43:36 UTC - in response to Message 1344111.

I don't think there's likely to be any way in which SETI, as a project, can access the client registry on Windows machines and make changes at that low level - and it would probably be roundly criticised as a security risk or a hacking exploit if it attempted any such thing. The one possibility might be to ask BOINC to include it into the BOINC installer, when the user is present and is explicitly giving permission for system changes to be made.

Could updating the Libcurl version supplied with Boinc do the same thing?

Libcurl Changelog

Claggy

Hmmm. I doubt it - I think RFC 1323 operates at a deeper level in the network protocol stack than libcurl, but there's no harm in asking. I'll run them past Rom sometime - I think these are more his thing that David's.

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8551
Credit: 50,424,923
RAC: 51,160
United Kingdom
Message 1344118 - Posted: 8 Mar 2013, 18:46:20 UTC - in response to Message 1344114.

I just wanted to point out that the download site does not mention Win8.....I tried it on my Win8 computer and it does not work there.

According to Speedguide.net Forums (TCP Optimizer): It doesn't yet work for Win8 :(

Does anyone have Win8, and a spirit of adventure enough to try one of the other deployment mechanisms? Does Win8 have regedit? Or a command prompt?

Josef W. SegurProject donor
Volunteer developer
Volunteer tester
Send message
Joined: 30 Oct 99
Posts: 4252
Credit: 1,050,582
RAC: 248
United States
Message 1344120 - Posted: 8 Mar 2013, 18:52:51 UTC

Joseph Davies TechNet article on TCP Receive Window Auto-Tuning as implemented in Win Vista and later also describes how XP (and Win2k) handled RWIN. That's useful for deciding whether the window scaling option is needed. I'm on dial up so I'll never need RWIN larger than 65536 bytes, but I did enable the timestamps feature on the theory that the sooner the download servers know I have a slow connection the better. None of my few downloads since have broken, though even before the change the majority completed in one attempt.

The additional timestamps header is 10 bytes, but headers are always padded to a mutiple of 4 bytes so the cost is 12 bytes. The tradeoff is that when a download breaks for an HTTP error, BOINC backs off 5k bytes when resuming the transfer to wipe any HTML error information which would otherwise corrupt the download. So doing what we can to ensure no breaks is well worthwhile IMO.

Joe

Profile ivan
Volunteer tester
Avatar
Send message
Joined: 5 Mar 01
Posts: 612
Credit: 140,124,713
RAC: 152,278
United Kingdom
Message 1344126 - Posted: 8 Mar 2013, 19:02:54 UTC - in response to Message 1344039.

It didn't seem to make much difference at work, on a W7 desktop and laptop. However, when I brought the laptop home to my own wireless network, it slid smoothly through its download backlog. But then, our work network/firewall is highly micro-managed by an obsessive manager who has "tuned" it to near unusability. When we were getting lots of "ghost" WUs it was almost always my work machines, and they also got lots of download failure errors.

On my home XP machine a stalled download completed smoothly but to soon to say if there's any effect. Now to go see if this can be tweaked on Linux...
____________

Profile TRuEQ & TuVaLu
Volunteer tester
Avatar
Send message
Joined: 4 Oct 99
Posts: 473
Credit: 18,696,043
RAC: 34,496
Sweden
Message 1344127 - Posted: 8 Mar 2013, 19:04:47 UTC - in response to Message 1344120.

Joseph Davies TechNet article on TCP Receive Window Auto-Tuning as implemented in Win Vista and later also describes how XP (and Win2k) handled RWIN. That's useful for deciding whether the window scaling option is needed. I'm on dial up so I'll never need RWIN larger than 65536 bytes, but I did enable the timestamps feature on the theory that the sooner the download servers know I have a slow connection the better. None of my few downloads since have broken, though even before the change the majority completed in one attempt.

The additional timestamps header is 10 bytes, but headers are always padded to a mutiple of 4 bytes so the cost is 12 bytes. The tradeoff is that when a download breaks for an HTTP error, BOINC backs off 5k bytes when resuming the transfer to wipe any HTML error information which would otherwise corrupt the download. So doing what we can to ensure no breaks is well worthwhile IMO.
Joe



And data is encapsulated in a package within each layer according to the settings for that layer... OSI.model....
The articel forgot that....

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8551
Credit: 50,424,923
RAC: 51,160
United Kingdom
Message 1344128 - Posted: 8 Mar 2013, 19:06:18 UTC - in response to Message 1344126.

Now to go see if this can be tweaked on Linux...

We think Linux has it enabled by default, but an actual observation would be interesting.

Your obsessive network micro-manager at work could probably push RFC 1323 to all Windows workstations via Group Policy. That would be fun.

Cruncher v.2
Send message
Joined: 8 Feb 04
Posts: 4
Credit: 3,569,734
RAC: 0
United States
Message 1344132 - Posted: 8 Mar 2013, 19:17:10 UTC

Guys sounds like a lot of people have dedicated systms for running seti tasks. However I use my everyday desktop to run tasks, and do day to day stuff. After I made this change to TCP1323 my downloads from all other website seem to be affected. My downloads begin at a normal pace compared to what I am use to with my ISP but after 2 or three minutes my download drops from 1.2MB per second to 23k/per second. Has anyone seen this as a side affect in making the change to fix the SETI downlaod problem?

Richard HaselgroveProject donor
Volunteer tester
Send message
Joined: 4 Jul 99
Posts: 8551
Credit: 50,424,923
RAC: 51,160
United Kingdom
Message 1344135 - Posted: 8 Mar 2013, 19:21:13 UTC - in response to Message 1344132.

Guys sounds like a lot of people have dedicated systms for running seti tasks. However I use my everyday desktop to run tasks, and do day to day stuff. After I made this change to TCP1323 my downloads from all other website seem to be affected. My downloads begin at a normal pace compared to what I am use to with my ISP but after 2 or three minutes my download drops from 1.2MB per second to 23k/per second. Has anyone seen this as a side affect in making the change to fix the SETI downlaod problem?

I haven't, but that's exactly the sort of side effect that we ought to be watching out for.

Can you give any examples of the sort of download (and some source websites) that we can use to try and replicate that observation?

Previous · 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 . . . 14 · Next

Message boards : Number crunching : Windows TCP Settings - Follow up - Help with server communication

Copyright © 2014 University of California