Posts by TJM

1) Message boards : Number crunching : Top results history page (Message 5975)
Posted 1 day ago by Profile TJM
Post:
This is a well known text used as a test workunit to check basic functionality of a new app.

Any workunit starting with m3_test, m4_test, g3_test or g4_test[1] most likely runs on a message for which settings are known already. It is the only way to check if the application can actually find the solution, which is very important because we could run broken app for years without knowing that something is wrong. As a side effect test runs that do not fail are used to gather statistical information.
[1] the first character in WU name is always 'm' for batches of CPU workunits and 'g' for anything assigned to GPU. This is used internally by the server so it does know where to look for workunit description. The second character is enigma version, 4 for M4 and 3 currently used for M3 and Heer (which runs as M3 with settings limited to Heer compatible).
2) Message boards : Number crunching : A GPU version of Enigma software (Message 5970)
Posted 2 days ago by Profile TJM
Post:
Please wait a minute, that fix is not final.

I don't think there is a need to re-check anything because each candidate key is tried many times, by doing multiple restarts.


Let me know when the app is ready to update, I've prepared release 1.10 that's just waiting for new executables.
3) Message boards : News : New app: "Enigma Test", OpenCL and Linux tests (Message 5968)
Posted 2 days ago by Profile TJM
Post:
There was a problem with translating --device to -d and parsing it inside the app.
I think it is fixed in 1.02 release because I've just seen a first completed and validated task.
4) Message boards : News : New app: "Enigma Test", OpenCL and Linux tests (Message 5965)
Posted 2 days ago by Profile TJM
Post:
I had to change app weights in the server configuration. With the weight set too low the feeder didn't even touch ATI tasks for some reason. Now it started sending them out as I see numbers raising in the server_status.

For the next few days there will be limited number of opencl test tasks as I'd like to check the app for any possible errors.
5) Message boards : Number crunching : A GPU version of Enigma software (Message 5959)
Posted 2 days ago by Profile TJM
Post:
It's fixed already, I did a stupid mistake patching the work generator to support test app and it placed all the workunits in the wrong queue.
6) Message boards : News : New app: "Enigma Test", OpenCL and Linux tests (Message 5951)
Posted 3 days ago by Profile TJM
Post:
I've just finished adding in on the server but there are no workunits yet.
7) Message boards : Number crunching : No Work - work_fetch blocked by project preferences (Message 5919)
Posted 4 days ago by Profile TJM
Post:
Daily quota was set to 250 tasks, it does not matter that much as long as the client returns only succesfull results. On error it is decreased and eventually the host won't get new tasks. I set it higher.
Weird thing is that the manager did not say anything about this right away.
8) Message boards : Number crunching : Enigma Cuda 1.08 runtime 7 minutes - now 1.09 only around a few seconds ?? (Message 5905)
Posted 4 days ago by Profile TJM
Post:
With all the new batches of work I'll be aiming for workunit length somewhere around the current g4* workunits or slightly longer.
I can't make them too long, as the app performance quickly scales down with GPU speed and the processing time would be painfully slow on mid range GPUs.
BOINC server has a mechanism that allows sending longer workunits to faster hosts and at some point I'll look into this.

In the meantime, I have disabled the g3_alqfi87_2 and g3_alqfi87_3 and replaced them with the same batches but with workunits 10x and 20x longer.
9) Message boards : Number crunching : Enigma Cuda 1.08 runtime 7 minutes - now 1.09 only around a few seconds ?? (Message 5900)
Posted 5 days ago by Profile TJM
Post:
These WUs are real ones just shorter.
Also the workunit length is not constant, it fluctuates a bit and the runtime depends on initial machine settings.
10) Message boards : Number crunching : Enigma Cuda 1.08 runtime 7 minutes - now 1.09 only around a few seconds ?? (Message 5897)
Posted 5 days ago by Profile TJM
Post:
Shorter workunits are from different batch and there is nothing wrong with them, other that the fact they're too short for fast GPUs.
There are a couple of workunits lengths as it was the first release and I was testing various options.
11) Message boards : Number crunching : A GPU version of Enigma software (Message 5876)
Posted 6 days ago by Profile TJM
Post:
Abort all the tasks assigned to app version 1.08 and try with the 1.09 that was just released.
12) Message boards : Number crunching : A GPU version of Enigma software (Message 5873)
Posted 6 days ago by Profile TJM
Post:
I think it ignores -d combined with -R.

this stops with GPU not found:
enigma-cuda -d 1 -M M3 -n 30 -a -f B:423:AA:AAA -t B:423:ZZ:ZZZ -g 023 00trigr.86v1 00bigr.86v1 fhpqx100

but this runs on GPU #0 even if #1 is specified:
enigma-cuda -d 1 -R -o results.txt 00trigr.41v1 00unigr.41v1 fhpqx100
13) Message boards : Number crunching : A GPU version of Enigma software (Message 5871)
Posted 6 days ago by Profile TJM
Post:
According to stderr, wrapper correctly sets -d to 0 or 1.

wrapper [26118 765 +dev#]: starting (GPU #1)
19:49:20 (13592): wrapper: running enigma_cuda.exe (-d 1 -R -o results.txt 04trigr.cur 04unigr.cur 04ciphertext)
20:12:22 (13592): called boinc_finish

wrapper [26118 765 +dev#]: starting (GPU #0)
19:48:32 (10808): wrapper: running enigma_cuda.exe (-d 0 -R -o results.txt 04trigr.cur 04unigr.cur 04ciphertext)
20:11:38 (10808): called boinc_finish

Could you try .bat files from this zip package ?
https://s3.eu-central-1.amazonaws.com/enigma4/test_runs/multigpu.zip

I suspect that the app somehow ignores -d option, as it runs with -d 1 on my single GPU system.
14) Message boards : Number crunching : A GPU version of Enigma software (Message 5869)
Posted 6 days ago by Profile TJM
Post:
It should work already, if it does not I'd like to see stderr output from a BOINC client running the app on two or more GPUs.
Just let it run more than one task for a while, then you can abort them.
15) Message boards : Number crunching : A GPU version of Enigma software (Message 5861)
Posted 7 days ago by Profile TJM
Post:

For sh#%s & giggles I even tried to changing the command to -d 1 in the job_160.xml file


This won't work, as the wrapper adds -d on the fly to the command line that's already in the job file and it always does that, defaults to GPU 0.
Which host id is that ? The wrapper writes command line used to run app to the stderr, so it can be seen online in result details.

You could try it outside of the BOINC, just copy the files from slot folder running GPU app somewhere, and create a test bat file:

wrappergpu_1.08_windows_x86_64__cuda_fermi.exe --device 1
pause

the wrapper will read the job file and should add the number passed after --device to command line.
I've just tested it and it does that on my system, but I do not have second GPU.

16) Message boards : News : Enigma Cuda 1.07 (Message 5857)
Posted 7 days ago by Profile TJM
Post:

PS: Any update on a fix for the multi GPU problem?


http://www.enigmaathome.net/forum_thread.php?id=836&postid=5852#5852 + a couple more following posts.
17) Message boards : Number crunching : A GPU version of Enigma software (Message 5856)
Posted 7 days ago by Profile TJM
Post:
App version updated to 1.08. It ships with new wrapper that translates --device passed by BOINC manager to -d
Seems to be working, at least with -d 0.

Note for app_info users: 1.08 wrapper won't work combined with previous versions of the app. New app will still work with 1.07 and older wrappers, but won't support --device / -d
18) Message boards : Number crunching : A GPU version of Enigma software (Message 5853)
Posted 8 days ago by Profile TJM
Post:

Will the new build of enigma-cuda.exe, be implemented on the enigma server, or is there a cheat sheet for dummies ?


I'll update the apps on the server later today. GPU workunits generator will be stopped for around ~1h right before update.
19) Message boards : News : Enigma Cuda 1.07 (Message 5838)
Posted 9 days ago by Profile TJM
Post:
Thanks TJM, the only v2.0 card that appears on that list is the 560. The 550 ti & 560 ti are both v2.1 cards.


Long time ago I had 560 and AFAIR it was compute capability 2.1
Just did a quick google search and first result seems to confirm that: http://www.geeks3d.com/20100606/gpu-computing-nvidia-cuda-compute-capability-comparative-table/
So most probably the app does not work on compute capability 2.0 cards.


Just out of interest is it possible to add to that list how many of each card is producing those numbers?

In theory that's possible and the project used to do that (for CPUs), but over time the table started getting too large for realtime processing and eventually that feature was removed.

Is that just one 560 producing those numbers or is there a dozens?

There are at least 3 of them that returned results, but as mentioned above I think they're compute capability 2.1 anyway.
20) Message boards : News : New app: "Enigma Test", OpenCL and Linux tests (Message 5837)
Posted 9 days ago by Profile TJM
Post:
The last version I tried didn't work on Intel GPUs - I've encountered lots of driver related issues and barely could run very short workunit without the app or driver crashing. Also the system was lagging so much I could barely move mouse pointer.
At this point it's probably going to be ATI only.


Next 20



Copyright © 2017 TJM