[QUOTE=ATH;530861]Anyone can figure out how to get CUDALucas to work on the Colab - Tesla P100 ?

It compiled for me but is not working: [URL]https://mersenneforum.org/showthread.php?t=12576[/URL][/QUOTE]Does your compiled version work on any other Colab gpu model? It shouldn't be too hard to get a Tesla K80 for a comparison test.
What output do you see when you try to run it on Colab? Error messages etc.

When I get a K80 on Colab and compile CUDALucas it works fine, but only on the K80, not if I get a P100.

When I compile it on the P100, I can run -cufftbench just fine and -threadbench until the very end it gives a:
*** buffer overflow detected ***
and does not write the theadbench file.

If I start it on an exponent it gives the same buffer overflow error.

Couldn't get any work:
[code]
Beginning GPU Trial Factoring Environment Bootstrapping...
Please see https://www.gpu72.com/ for additional details.

20191118_184450: GPU72 TF V0.32 Bootstrap starting...
20191118_184450: Working as "f16349dfe641d6bfae1bb5d6c3c8e2a4 "...

20191118_184450: Installing needed packages (1/3)
20191118_184453: Installing needed packages (2/3)
20191118_184454: Installing needed packages (3/3)
20191118_184455: Fetching initial work...
20191118_184455: Bootstrap finished. Exiting.
[/code]

[QUOTE=kracker;530907]Couldn't get any work:[/QUOTE]

No... What's going on is there's been a change on Colab...
[CODE]
./mfaktc: error while loading shared libraries: libcudart.so.10.0: cannot open shared object file: No such file or directory[/CODE]

Things are working fine under Kaggle...

Working it, although half of Barbados is without power at the moment, so I'm currently also dealing with some other things...

[QUOTE=ATH;530904]When I get a K80 on Colab and compile CUDALucas it works fine, but only on the K80, not if I get a P100.

When I compile it on the P100, I can run -cufftbench just fine and -threadbench until the very end it gives a:
*** buffer overflow detected ***
and does not write the theadbench file.

If I start it on an exponent it gives the same buffer overflow error.[/QUOTE]Please PM me a session capture of the P100 problem. "buffer overflow detected" is not present in the CUDALucas bug and wish list.
I guess you can try using CUDALucas on K80 and gpuowl on P100 for now.
I've taken to using the following at the very front of Colab scripts, so I can decide whether to go with what the session got, or try again.
!lscpu
!nvidia-smi

I have seen CUDALucas run into problems when run locally, if the span of cufftbench or -threadbench is too large; too many fft lengths for the size of the program's arrays. Threadbench can be run in multiple subranges to avoid that issue.

Running !apt-get install cuda-cudart-10-0 seems to fix the issue running mfaktc on colab.

[QUOTE=kracker;530907]Couldn't get any work:
[code]
Beginning GPU Trial Factoring Environment Bootstrapping...
Please see https://www.gpu72.com/ for additional details.

20191118_184450: GPU72 TF V0.32 Bootstrap starting...
20191118_184450: Working as "f16349dfe641d6bfae1bb5d6c3c8e2a4 "...

20191118_184450: Installing needed packages (1/3)
20191118_184453: Installing needed packages (2/3)
20191118_184454: Installing needed packages (3/3)
20191118_184455: Fetching initial work...
20191118_184455: Bootstrap finished. Exiting.
[/code][/QUOTE]

Same here.
Also problems when running mfaktc_colab.exe from a script to get "custom" mfaktc work: complaining about the missing library.

Got a T4, decided to try gpuowl on it... expected results, others as reference.
[code]
gpuowl(PRP) 92M exponent - 5M FFT

Tesla K80:
4.68 ms/iter - 66.8 GHz/days
430 GHz/days (mfaktc)

Tesla T4:
5.96 ms/iter - 52.4 GHz/days
~1700 GHz/days (mfaktc)

Tesla P100:
1.17 ms/iter - 266 GHz/days
~1100 GHz/days (mfaktc)
[/code]

[QUOTE=gg001;530916]Running !apt-get install cuda-cudart-10-0 seems to fix the issue running mfaktc on colab.[/QUOTE]

How do you do that?

[QUOTE=bayanne;530958]How do you do that?[/QUOTE]

Quick short term solution : create a new cel with only one instruction :

!apt-get install cuda-cudart-10-0

and execute it first, before executing the main script.

[QUOTE=De Wandelaar;530962]Quick short term solution : create a new cel with only one instruction :

!apt-get install cuda-cudart-10-0

and execute it first, before executing the main script.[/QUOTE]

:bow::bow::bow::bow: