Google Diet Colab Notebook

[Fan Ming]

Originally Posted by moebius

Thanks I missed that, I will try that later. I let the DC go through before colab throws me out.


2020-06-29 19:35:51 Tesla T4-0 53537243 OK 2000000 (jacobi == -1)
/

Tesla T4 is mainly used for trial factorting(really fast). It's a little bit waste to use it for LL/PRP -- it's not very faster than K80/P4. The fastest GPU for LL/PRP is P100, thus mainly used for LL/PRP. In fact, The speed of K80/P4 is also okay.

BTW, try -nospin option, otherwise the output would be excessive on colab.

[moebius]

Originally Posted by Fan Ming

Tesla T4 is mainly used for trial factorting(really fast). It's a little bit waste to use it for LL/PRP -- it's not very faster than K80/P4. The fastest GPU for LL/PRP is P100, thus mainly used for LL/PRP. In fact, The speed of K80/P4 is also okay.

Colab ended my runtime at 5:00 AM and deleted my gpuowl savefiles. However, I had saved this on my PC shortly beforehand so that colab couldn't get me. We continue with Tesla K80.

2020-06-30 06:16:48 Tesla K80-0 OpenCL compilation in 0.01 s
2020-06-30 06:16:48 Tesla K80-0 53537243 LL 10500000 loaded: daf6d21a9cf17d2b
2020-06-30 06:20:14 Tesla K80-0 53537243 LL 10600000 19.80%; 2064 us/it; ETA 1d 00:37; 8d6c9cba23eac3a2
2020-06-30 06:23:42 Tesla K80-0 53537243 LL 10700000 19.99%; 2074 us/it; ETA 1d 00:41; 6b9075e74a24a356
2020-06-30 06:27:09 Tesla K80-0 53537243 LL 10800000 20.17%; 2075 us/it; ETA 1d 00:38; 335defcdc4c59d52
2020-06-30 06:30:37 Tesla K80-0 53537243 LL 10900000 20.36%; 2075 us/it; ETA 1d 00:34; ee962f85f121fae0
2020-06-30 06:34:04 Tesla K80-0 53537243 LL 11000000 20.55%; 2075 us/it; ETA 1d 00:31; b6a6e068b4a5fbb9
2020-06-30 06:37:32 Tesla K80-0 53537243 LL 11100000 20.73%; 2075 us/it; ETA 1d 00:28; dfbd758a5964c750
2020-06-30 06:37:32 Tesla K80-0 53537243 OK 11000000 (jacobi == -1)

[axn]

Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
CPU(s):              2
On-line CPU(s) list: 0,1
Thread(s) per core:  2
Core(s) per socket:  1
Socket(s):           1
NUMA node(s):        1
Vendor ID:           AuthenticAMD
CPU family:          23
Model:               49
Model name:          AMD EPYC 7B12
Stepping:            0
CPU MHz:             2250.000
BogoMIPS:            4500.00
Hypervisor vendor:   KVM
Virtualization type: full
L1d cache:           32K
L1i cache:           32K
L2 cache:            512K
L3 cache:            16384K
NUMA node0 CPU(s):   0,1
Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid extd_apicid tsc_known_freq pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw topoext ssbd ibrs ibpb stibp vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 clzero xsaveerptr arat npt nrip_save umip rdpid

Freshly caught EPYC.

Appears to be about half as fast as Xeons :-(

[PhilF]

Originally Posted by axn

Freshly caught EPYC.

Appears to be about half as fast as Xeons :-(

So that one you throw back if you catch one. Thanks for the heads up.

[chalsall]

Originally Posted by PhilF

So that one you throw back if you catch one. Thanks for the heads up.

LOL... This really is a bit like fishing and is a function of the type of work being done. (As opposed to phishing, which is an entirely different space...)

[kriesel]

Originally Posted by moebius

Colab ended my runtime at 5:00 AM and deleted my gpuowl savefiles. However, I had saved this on my PC shortly beforehand so that colab couldn't get me.

Have you considered using google drive as your working directory? Automatic save from one session to the next.

[moebius]

Originally Posted by kriesel

Have you considered using google drive as your working directory? Automatic save from one session to the next.

Sure, I run gpu owl in /usr/local/bin. The savefiles are saved apparently in a subdirectory of the /content directory. You first have to find out which directory Google colab defines as working directory, apparently /content/drive . It appears as if colab occasionally breaks the connection at runtime so that the file browser cannot be used, which can be negated simply by calling up the page in a new tab.
In any case, the program is still running, even if Google Inc. thinks it could attract paying customers because of its restrictions to its 2nd class commercial hardware.


2020-06-30 23:16:25 Tesla K80-0 53537243 OK 38500000 (jacobi == -1)
2020-06-30 23:19:45 Tesla K80-0 53537243 LL 38700000 72.29%; 1999 us/it; ETA 0d 08:14; 3ae8c4a0dce1f3c9
2020-06-30 23:23:05 Tesla K80-0 53537243 LL 38800000 72.47%; 2000 us/it; ETA 0d 08:11; a2843afc77614d8e
2020-06-30 23:26:25 Tesla K80-0 53537243 LL 38900000 72.66%; 1999 us/it; ETA 0d 08:08; f08fc566607b0c59
/

[moebius]

Originally Posted by kriesel

Have you considered using google drive as your working directory? Automatic save from one session to the next.

Sorry for the foul language in my last post, I tried to modify the Phyton program so that it runs from google drive. However, the savefile folders are still created in the /content/ directory and deleted even after the runtime. I am not very familiar with Phyton and colab, only a bit with C / C ++; Linux. Do you have a hint for me?


import os.path
from google.colab import drive
if not os.path.exists('/content/drive/My Drive'):
drive.mount('/content/drive')
%cd '/content/drive/My Drive/gpuowl-master/'
!chmod 755 '/content/drive/My Drive/gpuowl-master/gpuowl.exe'
!cd '.' && /content/drive/My\ Drive/gpuowl-master/gpuowl.exe -prp 333XXXXXX

Oh yes, I was still assigned a Tesla P100.

2020-07-01 07:35:07 Tesla P100-PCIE-16GB-0 OpenCL compilation in 0.01 s
2020-07-01 07:35:09 Tesla P100-PCIE-16GB-0 333XXXXXX OK 52300000 loaded: blockSize 400, 160953a7a67244b6
2020-07-01 07:35:14 Tesla P100-PCIE-16GB-0 333XXXXXX OK 52300800 15.67%; 3430 us/it; ETA 11d 04:06; 7b9b2b946e50140d (check 1.79s)
2020-07-01 07:40:56 Tesla P100-PCIE-16GB-0 333XXXXXX OK 52400000 15.70%; 3432 us/it; ETA 11d 04:09; e77842f74a72a9a5 (check 1.79s)
2020-07-01 07:46:41 Tesla P100-PCIE-16GB-0 333XXXXXX OK 52500000 15.73%; 3431 us/it; ETA 11d 04:01; 0c2480f5c0259489 (check 1.79s)

[bayanne]

Completed exponents are not being reported from Colab ...

Can you advise?

[moebius]

I have now logged in with a different Google account and lo and behold, the save, log and result files are created where they belong, in /content/drive/My Drive/gpuowl-master/. Hope they stay there persistent.
Maybe something has been changed in the configuration of colab.

[bayanne]

Originally Posted by bayanne

Completed exponents are not being reported from Colab ...

Can you advise?

The p-1 factoring exponents are being accepted, but not the TF ones