HFM.NET - Monitoring Application for Folding@Home v7

[harlam357]

No worries! You have the dev work, I can only add my experiences.
You might be right, that HFM should be able to handle such large log files, but IMHO the log of the FAH client should not report these error exceptions in this way. Isn't it possible to log nothing for the GPU slot when FAH don't found a GPU at start? But I think this question should better be asked in the FAH Client thread.

However, thank you for your reply.

I agree as well. I did notice that the logging verbosity for your client was at 5. I wonder if these error messages would have printed so liberally if the verbosity were left at the default level of 3. If you bring this up to the FAHClient developers that may be their answer if the error messages only print once, or much less liberally, with the verbosity at default 3.

Thanks again for your help!
-H

[bruce]

(FYI, I recommend against Verbosity 5. I personally find it a lot easier to understand what's going on in the native log with Verbosity 3 than with 5 .. . . and so far, nobody has given me an example where there was a benefit to having 4 or 5.)

[Fahrenheit451]

I set it to 5 because v7 was a beta when I installed the client initially. Also I used the highest available verbosity level because it was recommended for the v6 client, IIRC. I thouhgt it gives more useful information for the developers. But I never compared the information on log level 3 vs. 5.

[Joe_H]

The only time I have seen a benefit to verbosity 5 with V7 was to see what the other default settings were. I set my client to that temporarily in the past to see the settings for log retention, and then modified my settings to retain 64 instead of the default of 16.

[harlam357]

New Version: 0.9.1 - Revision 595 - Download

Hello FAH community. Your friendly neighborhood HFM guy here. Here's the good news and a couple important items.

- HFM now requires the .NET Framework 4.0 Client Profile on Windows or Mono 2.8+ (2.10 recommended) on Linux.
- A 32-bit only package is now being offered. This build has "x86" in the file name. The only difference in this build vs. the existing distribution, now with "Any CPU" in the file name, is that the "x86" build will run as a 32-bit process on a 64-bit OS. The builds are mutually exclusive and cannot be installed side-by-side with windows installer. If you have a 32-bit OS then either the "Any CPU" or "x86" build will offer the same experience.
- There was a rather large, pun intended, memory issue with v0.9.0.548. This largely pertained to how I was handling the log file coming from the v7 client (growing pains with v7). I'm happy to report that the memory consumption is much, much improved in this version. I certainly hope this gets all you guys experienced any memory related issue back in action, especially those who contacted me and help me. The move to .NET 4.0 was largely prompted by this issue as that move in and of itself made a positive impact on the situation.

Otherwise... not a lot has changed. This is a well overdue bug fix release for v0.9.0. I hope to bring some goodies with the next release. Primarily an overhaul of the WU History Database. This should include some extra data added as well as much enhanced querying (wildcards, ANDs, ORs, etc).


Check the HFM Google Code page for downloads and details on the changes.

---

Need help with HFM? Have a feature request? Please join the discussions. Post in the HFM Google Group.

http://groups.google.com/group/hfm-net

[7im]

Excellent work as always!

[bollix47]

FYI

I've tried the new version on a couple of computers, one windows vista 32-bit and one windows 7 64-bit.
I get the invalid wire-type on both and on both I see an exessive load on the cpu ... as much as an entire core at times. The vista 32-bit computer became very laggy and not very useable so I reverted to 9.0 and that's still working fine. I did try both installers on vista 32-bit and both displayed the same behaviour when running.

The program appears to continually check the remotes (7) every minute or two even though I have it set to update every 15 minutes. I'm not sure whether that's a problem or just the way it's supposed to work.

Code: Select all

[28/07/2012-3:10:56 PM] - Starting - HFM.NET v0.9.1.595
[28/07/2012-3:10:56 PM] - 
[28/07/2012-3:11:00 PM] X Invalid wire-type; this usually means you have over-written a file without truncating or setting the length; see http://stackoverflow.com/q/2152978/23354
[28/07/2012-3:11:00 PM] X ProtoBuf.ProtoException: Invalid wire-type; this usually means you have over-written a file without truncating or setting the length; see http://stackoverflow.com/q/2152978/23354
   at ProtoBuf.ProtoReader.ReadSingle()
   at proto_8(Object , ProtoReader )
   at ProtoBuf.Meta.TypeModel.TryDeserializeAuxiliaryType(ProtoReader reader, DataFormat format, Int32 tag, Type type, Object& value, Boolean skipOtherFields, Boolean asListItem, Boolean autoCreate, Boolean insideList)
   at ProtoBuf.Meta.TypeModel.TryDeserializeList(ProtoReader reader, DataFormat format, Int32 tag, Type listType, Type itemType, Object& value)
   at ProtoBuf.Meta.TypeModel.TryDeserializeAuxiliaryType(ProtoReader reader, DataFormat format, Int32 tag, Type type, Object& value, Boolean skipOtherFields, Boolean asListItem, Boolean autoCreate, Boolean insideList)
   at ProtoBuf.Meta.TypeModel.Deserialize(Stream source, Object value, Type type, SerializationContext context)
   at ProtoBuf.Serializer.Deserialize[T](Stream source)
   at HFM.Core.Serializers.ProtoBufFileSerializer`1.Deserialize(String fileName)
   at HFM.Core.DataContainer`1.Read()
[28/07/2012-3:11:01 PM] - SQLite DLL Path: C:\Program Files (x86)\HFM.NET\SQLite\AMD64\System.Data.SQLite.dll
[28/07/2012-3:11:04 PM] - Starting Retrieval Timer Loop: 15 Minutes
[28/07/2012-3:11:04 PM] - Stopping retrieval timer loop.
[28/07/2012-3:11:05 PM] - Starting Retrieval Timer Loop: 15 Minutes
[28/07/2012-3:11:05 PM] - Total Retrieval Execution Time: 577 ms
[28/07/2012-3:11:09 PM] - (Enterprise Slot 00) Slot Status: Running
[28/07/2012-3:11:09 PM] - (Gemini Slot 00) Slot Status: Running
[28/07/2012-3:11:09 PM] - (Gemini) Retrieval finished in 1,123 ms
[28/07/2012-3:11:09 PM] - (Enterprise) Retrieval finished in 1,154 ms
[28/07/2012-3:11:09 PM] - (Gemini Slot 00) Slot Status: Running
[28/07/2012-3:11:09 PM] - (Gemini) Retrieval finished in 78 ms
[28/07/2012-3:11:10 PM] - (Enterprise Slot 00) Slot Status: Running
[28/07/2012-3:11:10 PM] - (Enterprise) Retrieval finished in 343 ms
[28/07/2012-3:11:10 PM] - (Endeavour Slot 00) Slot Status: Running
[28/07/2012-3:11:10 PM] - (Endeavour) Retrieval finished in 2,059 ms
[28/07/2012-3:11:11 PM] - (Atlantis Slot 00) Slot Status: Running
[28/07/2012-3:11:11 PM] - (Atlantis) Retrieval finished in 374 ms
[28/07/2012-3:11:12 PM] - (Endeavour Slot 00) Slot Status: Running
[28/07/2012-3:11:12 PM] - (Endeavour) Retrieval finished in 2,137 ms
[28/07/2012-3:11:13 PM] - (Apollo13 Slot 00) Slot Status: Running
[28/07/2012-3:11:13 PM] - (Apollo13) Retrieval finished in 1,810 ms
[28/07/2012-3:11:13 PM] - (Atlantis Slot 00) Slot Status: Running
[28/07/2012-3:11:13 PM] - (Atlantis) Retrieval finished in 78 ms
[28/07/2012-3:11:13 PM] ! (Discovery Slot 00) Could not find log section for slot 0. Cannot update frame data for this slot.
[28/07/2012-3:11:13 PM] - (Discovery Slot 00) Slot Status: RunningNoFrameTimes
[28/07/2012-3:11:13 PM] - (Discovery) Retrieval finished in 4,789 ms
[28/07/2012-3:11:14 PM] - (Apollo13 Slot 00) Slot Status: Running
[28/07/2012-3:11:14 PM] - (Apollo13) Retrieval finished in 1,154 ms
[28/07/2012-3:11:17 PM] ! (Discovery Slot 00) Could not find log section for slot 0. Cannot update frame data for this slot.
[28/07/2012-3:11:17 PM] - (Discovery Slot 00) Slot Status: RunningNoFrameTimes
[28/07/2012-3:11:17 PM] - (Discovery) Retrieval finished in 3,619 ms
[28/07/2012-3:11:27 PM] - (Challenger Slot 01) Slot Status: Running
[28/07/2012-3:11:27 PM] - (Challenger) Retrieval finished in 281 ms
[28/07/2012-3:11:27 PM] - (Challenger Slot 01) Slot Status: Running
[28/07/2012-3:11:27 PM] - (Challenger) Retrieval finished in 312 ms
[28/07/2012-3:12:32 PM] - (Mother Slot 01) Slot Status: Running
[28/07/2012-3:12:32 PM] - (Mother) Retrieval finished in 1,841 ms
[28/07/2012-3:12:34 PM] - (Mother Slot 01) Slot Status: Running
[28/07/2012-3:12:34 PM] - (Mother) Retrieval finished in 1,841 ms
[28/07/2012-3:13:34 PM] - (Atlantis Slot 00) Slot Status: Running
[28/07/2012-3:13:34 PM] - (Atlantis) Retrieval finished in 320 ms
[28/07/2012-3:13:37 PM] - (Apollo13 Slot 00) Slot Status: Running
[28/07/2012-3:13:37 PM] - (Apollo13) Retrieval finished in 2,150 ms
[28/07/2012-3:14:04 PM] - Stopping retrieval timer loop.
[28/07/2012-3:14:06 PM] - (Apollo13 Slot 00) Slot Status: Running
[28/07/2012-3:14:06 PM] - (Apollo13) Retrieval finished in 1,100 ms
[28/07/2012-3:14:06 PM] - (Atlantis Slot 00) Slot Status: Running
[28/07/2012-3:14:06 PM] - (Atlantis) Retrieval finished in 70 ms
[28/07/2012-3:14:06 PM] - (Challenger Slot 01) Slot Status: Running
[28/07/2012-3:14:06 PM] - (Challenger) Retrieval finished in 290 ms
[28/07/2012-3:14:08 PM] - (Endeavour Slot 00) Slot Status: Running
[28/07/2012-3:14:08 PM] - (Endeavour) Retrieval finished in 1,790 ms
[28/07/2012-3:14:08 PM] ! (Discovery Slot 00) Could not find log section for slot 0. Cannot update frame data for this slot.
[28/07/2012-3:14:08 PM] - (Discovery Slot 00) Slot Status: RunningNoFrameTimes

[sortofageek]

Thank you! So very much! I finally installed V7 on my two old Windows boxen yesterday. I've been tearing my hair out trying to figure out why I couldn't get monitoring to work. Your post resolved it for me, helped me get the syntax right on the remote server's FAH client and confirmed the need to restart the client.

Monitoring works perfectly here now in my W2K3 domain from a Win7 box from both the V7 client and from HFM.NET.

127.0.0.1 to 192.168.2.0/24

Where you have the above in both the remote client and the monitoring(local) client try:

Code: Select all

127.0.0.1,192.168.2.100-192.168.2.124

A restart may be necessary.

Also, using the i.p. address in HFM shouldn't be necessary ... you should use the remote Computer Name. That way if the address changes monitoring will still work properly.

[pmsfh2008]

I posted here and created this new topic.

humm,
A look at the task manager shows HFM eating up +/- 25% CPU
On P7200 (17/6/282)
With HFM HFM FAH Control
TPF 14.53 13.06
Without HFM 5.22 - may further decrease based upon how calculated.
--
So HFM is out for me.
FYI

[bollix47]

Reported a similar situation:

viewtopic.php?p=221015#p221015

I was also seeing the 25% cpu usage on 9.1 but it cleared up when I switched back to 9.0.

I did send Harlem an email with a link to the above post but haven't had a reply.

[patonb]

Quick question thats kinda hard to search for.

I reinstalled windows onto my new ssd, leaving my old hfm install on my old hard drive. Now I've lost all my hfm old data.

Can i combine the old data with the new one?

Thanks

[bollix47]

I decided to give the latest version another try after having the excessive CPU usage problem the first time.

This time after using uninstall via the Control Panel I went to the Roaming and Local folders and deleted HFM and Harlam357 so that nothing from the previous version was left.

So far the CPU usage appears to have dropped back to what I consider normal ... i.e. no more jumps to using a complete core. Also, the HFM log looks a lot cleaner than the one I posted earlier.

It's early but unless the usage starts to jump up again over the next couple of days I will consider my problem solved and will begin building the benchmarks and history again.

[^w^ing]

I noticed that ever since HFM started recording my WUs into the WU History Database again (thanks for that by the way), units from my SMP slot are recorded as Uniprocessor units in the database.

[7im]

The hybrid fahcore_a4 work units (UNI/SMP) can only display one WU type in HFM. Since these were originally all UNI WUs, the designation stuck. This was discussed a few pages back if you wanted more detail.

[^w^ing]

hm, I skimmed a few pages but didn't notice, I suspected as much anyway. Thanks for the info. It is rather unfortunate that the column is named "Slot type" then.


Also, I noticed that my HFM sometimes enters one WU into the WU history database multiple times for some reason, I am looking at 31 consecutive entries of the same WU right now. They differ in ppd, credit (both slowly decreasing) and completion date (slowly increasing, in minutes) columns. It looks like the longer they take to upload, the more times they are entered in the database.