Google Diet Colab Notebook

[kriesel]

Originally Posted by chalsall

Not any longer.

I just added another, using my sole phone # to receive the 6-digit confirmation code. (Maybe there are regional differences?)

[chalsall]

Originally Posted by kriesel

I just added another, using my sole phone # to receive the 6-digit confirmation code. (Maybe there are regional differences?)

Thanks for running the test. Be aware there will be "signals" based on your usage. I dare you to bring eight (8#) GMail accounts running Colab using the same NPA (I didn't dare try this, but little Bimshire is a "developing country").

No one knows what is actually acceptable. Google's T&C is a "black box". It is worth running experiments, just to see (empirically) what can be done...

Empirically, no one actually knows for sure. In multiple jurisdictions.

That might be unsettling for regulators (assuming, of course, the regulators are actually on the side of the consumers; not always the case).

[slandrum]

Originally Posted by chalsall

Thanks for running the test. Be aware there will be "signals" based on your usage. I dare you to bring eight (8#) GMail accounts running Colab using the same NPA (I didn't dare try this, but little Bimshire is a "developing country").

No one knows what is actually acceptable. Google's T&C is a "black box". It is worth running experiments, just to see (empirically) what can be done...

Empirically, no one actually knows for sure. In multiple jurisdictions.

That might be unsettling for regulators (assuming, of course, the regulators are actually on the side of the consumers; not always the case).

8 of my free google accounts running colab instances all use the same phone number, and same backup email address (which of course is a gmail address). They run from separate tabs on the same browser on the same computer at the same time. I have different names on the accounts, but there's absolutely no verification of that. Of course I'm in the US of A, but there's been no issue. I seem to be able to create as many google accounts as I want, each with their own separate google drive allocation. At some point if I added dozens of accounts maybe Google would notice. When Google randomly decides that one of my accounts can't get a GPU for a week or two, it doesn't affect any of my other accounts.

[LaurV]

During my colab "frenzy" (described somewhere here, in Wayne's thread) I tried to use my phone numbers (yep, I have more phone numbers; I have two sim cards in my phone, and I work with ublox/sara modules at job, writing firmware for devices in which these toys are included, so the company provided two additional sim cards that I use for testing our firmware) for free colab accounts. I mean, the numbers were given for paid (Pro) account, but I tried to create free accounts in parallel for each number, maybe they could get some cores sometime and add on top of what paid accounts were crunching. The free accounts which had the same phone number as the Pro accounts NEVER got any resource. Nor GPU, neither CPU. I gave up after a while.

Edit: also, this part about getting two sessions running for the same account, and getting more GPU time with the trick, never worked for me. I could open how many sessions I wanted, but the performance will degrade, i.e. if one session made a hundred iterations per time unit, then two sessions will make each 45 or 48 iterations for the same time unit (numbers only for example of proportion, not real values).

[pokepoke]

my assignment was struck at
20211106_125823 ( 0:02): Fetching payload...

how to fix it

[ric]

Mostly a memo-to-self: starting from today, Colab (free tier) refused to mount Drive belonging to a different username/ownership.
Solution (from a reddit here -> https://www.reddit.com/r/GoogleColab...or_credential/) was simply to modify in all interested scripts

if not os.path.exists('/content/drive/My Drive'):
  drive.mount('/content/drive', force_remount=True)

into

if not os.path.exists('/content/drive/My Drive'):
  drive._mount('/content/drive', force_remount=True)

(notice the underscore before mount)

[mathgeeks]

hey can you please tell what code to run in the Colab to get started with finding Mersenne prime numbers?

[kriesel]

Originally Posted by mathgeeks

hey can you please tell what code to run in the Colab to get started with finding Mersenne prime numbers?

See https://www.mersenneforum.org/showthread.php?t=24839 for some code examples versus app. Choose according to what kinds of computation you want to do to help the search, and available hardware. Note Colab offers both CPU and GPU capability, free and paid.

[Zhangrc]

I tried to build gpuowl on my colab instance I just created today. I used the script of https://www.mersenneforum.org/showpo...30&postcount=7 .

However, there's an error:

/content/drive/My Drive
/content/drive/My Drive/gpuowl
fatal: destination path 'gpuowl' already exists and is not an empty directory.
/content/drive/My Drive/gpuowl/gpuowl
Reading package lists... Done
Building dependency tree       
Reading state information... Done
libgmp-dev is already the newest version (2:6.1.2+dfsg-2).
0 upgraded, 0 newly installed, 0 to remove and 39 not upgraded.
Reading package lists... Done
Building dependency tree       
Reading state information... Done
g++-8 is already the newest version (8.4.0-1ubuntu1~18.04).
gcc-8 is already the newest version (8.4.0-1ubuntu1~18.04).
0 upgraded, 0 newly installed, 0 to remove and 39 not upgraded.
update-alternatives: using /usr/bin/gcc-8 to provide /usr/bin/gcc (gcc) in auto mode
update-alternatives: using /usr/bin/g++-8 to provide /usr/bin/g++ (g++) in auto mode
There is only one alternative in link group gcc (providing /usr/bin/gcc): /usr/bin/gcc-8
Nothing to configure.
There is only one alternative in link group g++ (providing /usr/bin/g++): /usr/bin/g++-8
Nothing to configure.
g++ (Ubuntu 8.4.0-1ubuntu1~18.04) 8.4.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

echo \"`git describe --tags --long --dirty --always`\" > version.new
diff -q -N version.new version.inc >/dev/null || mv version.new version.inc
echo Version: `cat version.inc`
Version: "v7.2-91-g9c22195-dirty"
g++ -MT Pm1Plan.o -MMD -MP -MF .d/Pm1Plan.Td -Wall -g -O3 -std=gnu++17   -c -o Pm1Plan.o Pm1Plan.cpp
In file included from /usr/include/c++/8/cassert:44,
                 from Pm1Plan.cpp:5:
Pm1Plan.cpp: In function ‘constexpr u32 {anonymous}::firstMissingFactor(u32)’:
Pm1Plan.cpp:65:3: error: call to non-‘constexpr’ function ‘void __assert_fail(const char*, const char*, unsigned int, const char*)’
   assert(false);
   ^~~~~~
Makefile:33: recipe for target 'Pm1Plan.o' failed
make: *** [Pm1Plan.o] Error 1
create config.txt, worktodo.txt before continuing

Judging from it, there was an error in the code.
How can I solve this problem?
(PS: mprime is already crunching happily on my instance, though not as fast as my own computer)

[tdulcet]

Originally Posted by Zhangrc

Judging from it, there was an error in the code.
How can I solve this problem?

This is a known GCC bug. Just run this command before compiling GpuOwl:

sed -i 's/assert(false);/abort();/' Pm1Plan.cpp

[lycorn]

Is it just me or Colab is really grumpy?
All the instances I start (CPU only, as it has been for months now) are terminated a few minutes later (sometimes they last less than a minute). I have been using Colab regularly (5 gmail addresses) without any issue but suddenly....
Could be just a hiccup, or is it not?