Cannot upload results.

[JohnthePilot]

Hi,
In the absence of the main forum, I hope someone here can answer my
query.
I have been folding for some time but have had all sorts of problems.
I won't go into them as they are all on the main forum. However, on
the 26 November I completed my first WU (out of over a thousand
attempts) but F@H crashed when the WU finished. Since then I keep
getting this message in my log file.


[00:15:52] + Attempting to send results
[00:16:24] Couldn't send HTTP request to server (wininet)
[00:16:24] + Could not connect to Work Server (results)
[00:16:24] (171.64.122.78:8080)
[00:16:24] - Error: Could not transmit unit 07 (completed November 26)
to work server.

Has anyone any idea what the problem is?

Many thanks,
Pilot.

[John_Weatherman]

Hi!
First, make sure use Internet Explorer setting is NO, and use proxy
server setting is correct. The server seems to be ok, so check
firewall(s) and connections your end. Try those to begin with.

[Jmarks]

What type of wu was it and your OS?

On Nov 28, 7:28 pm, JohnthePilot <johnthepi...@blueyonder.co.uk>
wrote:

[Jmarks]

Bad WU.
Stop the client (Ctrl+C),
Deleted the Work folder, and the queue.dat file.
Restart client.

On Nov 28, 7:28 pm, JohnthePilot <johnthepi...@blueyonder.co.uk>
wrote:

[alancabler]

Hello JohnthePilot,
John_Weatherman is correct- the (wininet) error message is a classic
indicator that your config's "Use IE connection settings" should be
changed to NO.

Regarding the "over 1000 attempts" issue, with more info, someone here
can help you solve that problem, too.

[JohnthePilot]

Thanks for all the replies. I've unchecked IE and will see what
happens next time F@H attempts to send the results.
With regard to the 1000 WUs, this is not an exageration.

[URL=http://imageshack.us][IMG]http://img529.imageshack.us/img529/6223/
fahqg2.jpg[/IMG][/URL]

A considerable amount of time has been spent in the main forum trying
to resolve my problem, so I don't really want to go over old ground.
To put you in the picture, I keep getting various error messages
ranging from "Gromacs cannot proceed further through "NaN/Inf
detected" to "Simulation instability has been encountered". The
Prime95 torture test failed after 3 hours so I systematically checked
my hardware. I've replaced my memory, bought a new video card,
upgraded my PSU and flashed my BIOS. I've run a stresscpu test and
memtest86 on both RAM sticks for 8 hours each with no problems, but
Prime95 is still crashing. I can't think what else might be wrong.

[JohnthePilot]

I don't know what went wrong with the image in the last post but
basically, it showed my certificate from F@H for 1000 WUs :grin:

[noorman]

http://img529.imageshack.us/img529/6223/fahqg2.jpg

In fact, it is a certificate for 900 WU's under your name ...

Have you ever checked your CPU die temperature whilst running Prime
and F@H ?
Are you Overclocking ?
Check those temps (if you can) whith system under load.
In F@H, a CPU die can get hot locally; I experienced that with my
A64X2 setup at the start. I had it Overclocked to 2.6 GHz, running 2
single-core clients, one on each CPU-core.
One core did well and didn't crash, the other CPU-core did crash
sometimes.
After I got the CPU clock to 2.5 GHz it has been rock solid for over
2.5 years now (24/7) !
My conclusion on that was that one die got a bit hotter than the other
and crashed the process, the other held at 2.6 GHz.


NOTE: I get an OK message in my browser if I try to connect to
171.64.122.78 port 8080 ;-)


On Nov 29, 11:59 am, JohnthePilot <johnthepi...@blueyonder.co.uk>
wrote:

[John_Weatherman]

There are free benchmarking programs (like SiSoftware Sandra) to check
all your settings, temps etc. and I see that Prime95 can be customized
to stresstest single components - which might answer some questions.
According to Wiki Prime95 also tests the hard drive - most
manufacturers have a hard drive test program on their websites, if you
want to check yours.

[JohnthePilot]

@ Noorman. The forum seems to have lost my last reply although it said
"post successful". I said that that must have been an old image I
uploaded. This is my present status.

http://vspx27.stanford.edu/cgi-bin/main.py?qtype=userpage&username=JohnthePilot

As for temps, these are all normal; 38° for mobo and 53° for CPU. No
overclocking. FSB at 133MHz and CPU at 1666MHz.

@ John_Weatherman. I'll give that a go and see what it throws up.
Unchecking IE seems to have worked. Thanks.
[13:04:36] + Attempting to send results
[13:04:36] + Results successfully sent
[13:04:36] Thank you for your contribution to Folding@Home.
[13:04:40] - Preparing to get new work unit...

[JohnthePilot]

Further to my last post I have stresstested the CPU and memory and
have run the manufacturer's diagnostics on my hard drives. All OK. The
only thing I haven't checked is the mobo itself.

[noorman]

I can connect to 171.64.122.78 port 8080 (get an OK message on an
otherwise blank webpage and that 's 100% OK)
You can this check on any server address by copying the server address
followed by :8080
in to the URL bar of your browser and hitting Enter.

To communicate with the Stanford servers you need to be sure your port
8080 is opened (for FaH anyway) !
The log indicated that it couldn't connect to that specific server.
You could try connecting to some other servers from Stanford (List is
here: http://fah-web.stanford.edu/serverstat.html)
When you don't get an OK message for any of them, you have a
communication problem and the problem then isn't at Stanford's
side ...

Testing the path from you to the destination can be done with
'tracert' in Windows [Command box](there is a similar way in Linux,
but I can't remember the correct command.
Behind the 'tracert' (without '') you need to type a space and then
the IP address (or a website URL), then hit Enter.


On Nov 29, 1:28 am, JohnthePilot <johnthepi...@blueyonder.co.uk>
wrote:

[JohnthePilot]

Thanks noorman. I can get in now. It was the IE setting that was
causing the problem. Now I just need to sort out my instability
problem.