The Unhappy Me thread

[Batalov]

Originally Posted by firejuggler

My smartphone got stolen.
Nothing really important on it, but still annoying.

Consider bricking it with the remote tools. (Once you make certain that it was not just lost.)
Let them have just a dead mass of silicon and wires.

[Uncwilly]

Originally Posted by Batalov

Consider bricking it with the remote tools. (Once you make certain that it was not just lost.)
Let them have just a dead mass of silicon and wires.

Years ago I had a mobile phone (pre smartphones) stolen. It did not use a SIM card. Once I was aware of it, I called the carrier. They set it up so that any attempted calls made from it would go to the carrier's security office.

[ixfd64]

George mentioned indirectly in the other thread that he tested negative for COVID-19.

[firejuggler]

State enforced curfew in France, starting saturday, at 6 PM for at least 2 week. Untill now, it was restricted to the eastern region of France.

[petrw1]

Our AC stopped working on the hottest week of the summer so far.
I had to shut down the 5 extra heaters I have running here.

[masser]

Originally Posted by petrw1

Our AC stopped working on the hottest week of the summer so far.
I had to shut down the 5 extra heaters I have running here.

Oh no! I am having power supply issues with one of my boxen. It stinks to see it just sitting there.

[chalsall]

Originally Posted by petrw1

I had to shut down the 5 extra heaters I have running here.

I'd highly advise you to look into "the cloud". AWS Spot instances can be viable (depending on one's power costs; Capex out).

Also, Google GCE offers a one-time $300 USD free trial. This doesn't offer GPUs to play with, but their 8-vcore instances make excellent P-1'ers...

[kriesel]

Originally Posted by chalsall

I'd highly advise you to look into "the cloud". AWS Spot instances can be viable (depending on one's power costs; Capex out).

Also, Google GCE offers a one-time $300 USD free trial. This doesn't offer GPUs to play with, but their 8-vcore instances make excellent P-1'ers...

Care to post a how-to for that GCE?

[MooMoo2]

Originally Posted by chalsall

Also, Google GCE offers a one-time $300 USD free trial. This doesn't offer GPUs to play with, but their 8-vcore instances make excellent P-1'ers...

Originally Posted by kriesel

Care to post a how-to for that GCE?

In 2018, I looked into breaking one of these records: http://www.numberworld.org/y-cruncher/ with GCE. I looked through my files and found the following notes. I don't know how much (if any) of it is useful, but here it goes:

Sign into Google Compute (requires Gmail or other Google account)
Click “Try it free”, fill out form. You need a credit card, but it gives you $300 of free credit. You won’t be billed unless you upgrade to a paid account, but this gives you a limited quota (not more than 8 cores and 2TB of hard drive memory/”persistent disk”, for example).
It’s preferable to use Linux/Ubuntu instead of Windows. One CPU using Linux is $24.67/month, while one CPU using Windows server is $54.75/month. Turning pre-emptibility on (meaning your computing instance can be interrupted at any time) using Linux can cut it down to $7.70/month, but you’ll need to restart and possibly re-install the program you’re trying to run at least once a day. This may be OK for prime-hunting programs but is a no-go for computing large constants to trillions of decimal places.
But first, type in “quotas” on the search bar to go to your quotas page. Increase your “persistent disk standard” for your region (us-east1, for example) to allow for the increased storage. Calculating square roots and the Golden ratio to 12 billion (12*10^9) decimal digits requires 50.1 GiB, so let’s be safe and use 15000 GB (~15 TB) for our 2.5 trillion (2.5*10^12) decimal digit attempt. Click on “edit quotas” to request an increase. It will require your phone and a reason for the increase, but it’s usually granted within a few minutes, no questions asked.
Create the instance. Select the CPUs (number of cores) and use the Ubuntu 18.04 LTS (or any other Linux or Ubuntu version) as a boot disk. Under “firewall”, allow both HTTP and HTTPS traffic. Under “management, disks, networking, SSH keys”, make sure pre-emptibility is off.
The instance will take a minute or so to create. You can connect to it via SSH by clicking on the tab to open in browser window. You may have to do it twice if it gets blocked by a pop-up blocker. A wall of text should appear (starting with “Welcome to Ubuntu”) with no GUI.
Install some packages to view the instance via a GUI/desktop environment. Use these commands and click “y” for “yes” when prompted.
$ sudo apt-get update
$ sudo apt-get install aptitude tasksel
$ sudo tasksel install gnome-desktop --new-install
$ sudo apt-get install tightvncserver
$ sudo apt-get upgrade
$ sudo apt-get install gnome-shell
$ sudo apt-get install ubuntu-gnome-desktop
$ sudo apt-get install autocutsel
$ sudo apt-get install gnome-core
$ sudo apt-get install vnc4server
Install a VNC server:
$ vncserver
Choose a password and type it in when prompted. Download a VNC Client, such as VNC Viewer.
Open the firewall. Go to the search toolbar on Google Compute, type in firewall, choose firewall rules for the VPC Network. Create a new firewall rule for “ingress”, allow all IP ranges (0.0.0.0/0), type in “tcp:5901” under “protocols and ports”, and type in a description (such as “one”) under “target tags”.
Go back to the compute engine under “VM Instances”, click on your instance, and then click “edit”. On “network tags”, type in your earlier firewall description (“one” in our previous example). Click “save”.
Open VNC Viewer, go to “new connection”, and type in the VNC Server. This should be the external IP identifier, followed by :5901 (xx.xxx.xx.xx:5901 for the machine I got). Type in anything (“one”?) for the name and leave everything else as the defaults. Double click on it (in VNC Viewer) to connect. It will ask you for your password, which is what you entered earlier after the command $ vncserver
The connection should work, but there’ll be a blank screen. To fix it, close VNC Viewer. Then, open and edit the startup file on the SSH window:
$ vim .vnc/xstartup
Change the file to state the following:
#!/bin/sh

export XKL_XMODMAP_DISABLE=1
unset SESSION_MANAGER
unset DBUS_SESSION_BUS_ADDRESS

[ -x /etc/vnc/xstartup ] && exec /etc/vnc/xstartup
[ -r $HOME/.Xresources ] && xrdb $HOME/.Xresources
xsetroot -solid grey
vncconfig -iconic &

gnome-panel &
gnome-settings-daemon &
metacity &
nautilus &
gnome-terminal &
Press the escape key once and then type in:
:wq
Install the gnome panel:
$ sudo apt-get install gnome-panel
Adjust the geometry:
$ vncserver -geometry 1024x640
Kill and restart the VNC Session:
$ vncserver -kill :1
$ vncserver
and log back into the VNC Viewer. 
And voila! The desktop should come up. On the remote desktop, click the upper left button to access the Internet. The web browser will be Firefox. Download the appropriate programs (y-cruncher if you want to calculate the constants, LLR if you want to hunt for primes). LLR might not work, so enable 32-bit architecture on the SSH terminal:
$ sudo dpkg --add-architecture i386 
Add the repository:
wget -nc https://dl.winehq.org/wine-builds/Release.key
sudo apt-key add Release.key
sudo apt-add-repository https://dl.winehq.org/wine-builds/ubuntu/
and download wine:
$ sudo apt-get update
$ sudo apt-get install --install-recommends winehq-stable
You can now download the 64-bit windows version of LLR on the remote desktop. To run it, go to the appropriate directory. Use cd .. to go up a directory, if necessary, and type in the following commands:
sudo apt-get install wine (on the SSH terminal)
sudo apt-get install winehq-stable
wine llr64.exe (on the remote desktop, for LLR) or
wine prime95.exe (for Prime95)
Click install on the popup screens, and the Windows LLR screen should come up!
Now back to the constants. Download the static linux file of y-cruncher (using Firefox):
http://www.numberworld.org/y-cruncher/
If it needs a library called “libcilkrts.so.5”, get it from: https://rpmfind.net/linux/rpm2html/search.php?query=libcilkrts.so.5()(64bit)
under “Fedora 27 updates for x86_64” . Find the path with:
$ sudo find / -name libcilkrts.so.5
and set the library path to the appropriate directory:
$ LD_LIBRARY_PATH=/usr/local/lib (for example)
$ export LD_LIBRARY_PATH
Use
$ ./y-cruncher
to run the Y-cruncher program.
To go for the record, we’ll need a bigger hard drive. 
Create, format, and mount a persistent disk. Google does a pretty good job of explaining this, so I’ll mostly copy and paste the following from: https://cloud.google.com/compute/docs/disks/add-persistent-disk
with some slight modifications:
Go to the VM instances page.
Click the name of the instance where you want to add a disk.
At the top of the instance details page, click Edit.
Under Additional disks, click Add item.
In the Name column for additional lists, click the Select a disk drop down menu and select Create disk.
Configure the properties for your new disk. The source image should be the ubuntu version you’re using. Specify a name for the disk and select the Blank disk option. If you want the disk to be deleted when the instance is deleted, choose the appropriate option (keep in mind that there is a significant cost for large disk storage)
Click Create to create the disk.
Unfortunately, please note that there may not always be enough resources in the region to create one, so you may have to do this during a different date and time. You may not have to wait long, though – trying to create another one again in less than half an hour sometimes succeeds.
At the bottom of the instance details page, click Save to apply your changes to the instance and attach the new disk.
On the SSH terminal that has the instance, use the lsblk command to list the disks that are attached to your instance and find the disk that you want to format and mount.
$ sudo lsblk

NAME   MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
sda      8:0    0   10G  0 disk
└─sda1   8:1    0   10G  0 part /
sdb      8:16   0  250G  0 disk

In this example, sdb is the device ID for the new persistent disk.
Format the disk using the mkfs tool. This command deletes all data from the specified disk, so make sure that you specify the disk device correctly. To maximize disk performance, use the recommended formatting options in the -E flag. It is not necessary to reserve space for root on this secondary disk, so specify -m 0 to use all of the available disk space.
$ sudo mkfs.ext4 -m 0 -F -E lazy_itable_init=0,lazy_journal_init=0,discard /dev/[DEVICE_ID]
where [DEVICE_ID] is the device ID of the persistent disk that you are formatting. For this example, specify sdb to format the entire disk with no partition table. This step may take some time (15 minutes or so) for large (>10 TB) disks.
Note: RHEL 6 and CentOS 6 do not accept the lazy_journal_init=0 argument. Remove this parameter when formatting persistent disks on instances with these images.
Create a directory that serves as the mount point for the new disk. You can use any directory that you like, but this example creates a new directory under /mnt/disks/.
$ sudo mkdir -p /mnt/disks/[MNT_DIR]
where: [MNT_DIR] is the directory where you will mount your persistent disk. (I called it the “Z” Drive, so [MNT_DIR]=Z)
Use the mount tool to mount the disk to the instance with the discard option enabled:
$ sudo mount -o discard,defaults /dev/[DEVICE_ID] /mnt/disks/[MNT_DIR]
where:
[DEVICE_ID] is the device ID of the persistent disk that you are mounting.
[MNT_DIR] is the directory where you will mount your persistent disk. 
Configure read and write permissions on the device. For this example, grant write access to the device for all users.
$ sudo chmod a+w /mnt/disks/[MNT_DIR]
where: [MNT_DIR] is the directory where you mounted your persistent disk.
Your disk name (“Z” in this example) will appear on the virtual machine. To get to it, click on “Other Locations” on the virtual machine, and then navigate to the computer/MNT/disks/Z directory. Copy the y-cruncher files onto that directory, then navigate to the MNT/disks/Z directory using the virtual machine terminal. The command “cd ..” may need to be used more than once to get to the home directory before entering cd/MNT/disks/Z.
Navigate to the y-cruncher folder, and then use ./y-cruncher to run the program.
Some notes and benchmarks:
Square roots and the Golden Ratio take approximately the same amount of time to compute and require similar amounts of memory. Calculating pi needs a bit more memory and considerably more time. Other constants such as log (2) take even more time to compute than pi.
1 core, Intel Xeon 2.3 GHz: (Standard one-core instance on Google Compute)
500 million decimal digits of the Golden Ratio – 2.94 minutes needed
2 billion (2*10^9) decimal digits of the square root of 125 – 8.31 GiB Peak disk usage, 20.67 minutes needed
2 billion (2*10^9) decimal digits of the Golden Ratio – 20.88 minutes needed
2 billion (2*10^9) decimal digits of Pi – 9.42 GiB Peak disk usage, 80.06 minutes needed
4 billion (4*10^9) decimal digits of the Golden Ratio – 16.6 GiB Peak disk usage, 49.67 minutes needed
12 billion (12*10^9) decimal digits of the square root of 2 – 50.0 GiB Peak disk usage, 215.58 minutes needed (3.593 hours)
Now let’s delete the old instance, and ramp things up by creating a new one. On the machine type, select the 64 vCPUs, high memory (416 GB) option and run some benchmarks (the default version only has 240 GB for the 64 vCPUs). This CPU itself (excluding the costs of the persistent disks) costs $1,937/month, so let’s hope it won’t take that long to calculate those numbers.
…
The next run will calculate the square root of 5 to 100 billion (100 x 10^9) decimal digits. Unfortunately, this run takes quite a while (almost 5 hours!) because the CPU utilization is too low. This is mostly due to low disk read speeds, so let’s see what they are:
$ sudo hdparm -tT /dev/sdb
Cached disk reads: 8348 MB/sec
Buffered disk reads: 163 MB/sec
To get around this, install 8 local SSDs (3 TB total):
sudo apt-get update && sudo apt-get install mdadm --no-install-recommends
and follow the rest of the steps on: https://cloud.google.com/compute/docs/disks/local-ssd#formatmultiple
Use the lsblk command to identify all of the local SSDs that you want to mount together. For this example, the instance has eight local SSD devices in NVMe mode:
$ lsblk

NAME    MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
sda       8:0    0   10G  0 disk
└─sda1    8:1    0   10G  0 part /
nvme0n1 259:0    0  375G  0 disk
nvme0n2 259:1    0  375G  0 disk
nvme0n3 259:2    0  375G  0 disk
nvme0n4 259:3    0  375G  0 disk
nvme0n5 259:4    0  375G  0 disk
nvme0n6 259:5    0  375G  0 disk
nvme0n7 259:6    0  375G  0 disk
nvme0n8 259:7    0  375G  0 disk

Local SSDs in SCSI mode have standard IDs like sdb. Local SSDs in NVMe mode have IDs like nvme0n1.
Use mdadm to combine multiple local SSD devices into a single array named /dev/md0. This example merges eight Local SSD devices in NVMe mode. For Local SSD devices in SCSI mode, specify the IDs that you obtained from the lsblkcommand:
$ sudo mdadm --create /dev/md0 --level=0 --raid-devices=8 /dev/nvme0n1 /dev/nvme0n2 /dev/nvme0n3 /dev/nvme0n4 /dev/nvme0n5 /dev/nvme0n6 /dev/nvme0n7 /dev/nvme0n8

mdadm: Defaulting to version 1.2 metadata
mdadm: array /dev/md0 started.

Format the full /dev/md0 array with an ext4 file system. This command deletes all existing data from the local SSDs.
$ sudo mkfs.ext4 -F /dev/md0
Create a directory to where you can mount /dev/md0. For this example, create the /mnt/disks/ssd-array directory:
$ sudo mkdir -p /mnt/disks/[MNT_DIR] 
where: [MNT_DIR] is the directory where you want to mount your local SSD array. (I used “Z” for [MNT_DIR], so it’ll be in the Z Drive)
Mount the /dev/md0 array to the /mnt/disks/ssd-array directory. Optionally, you can disable write cache flushing to improve write performance while risking reduced durability for up to two seconds of cached data writes.
$ sudo mount /dev/md0 /mnt/disks/[MNT_DIR]
where: [MNT_DIR] is the directory where you want to mount your local SSD array.
Configure read and write access to the device. For this example, grant write access to the device for all users.
$ sudo chmod a+w /mnt/disks/[MNT_DIR]
where: [MNT_DIR] is the directory where you mounted your local SSD array.
Or right click -> open in terminal to change directories

[LaurV]

Originally Posted by petrw1

Our AC stopped working on the hottest week of the summer so far.

You mean like +2॰C that you have in SK maximum in summer?
The fact people use AC in northern Canada is news to me! I thought they only have 20kW++ heaters there...

(joking apart, that sucks, we had here +45॰C in May, swmbo would have died, and killed me first, if the AC would break - for me, I feel ok with them or without).

[tServo]

Originally Posted by LaurV

You mean like +2॰C that you have in SK maximum in summer?
The fact people use AC in northern Canada is news to me! I thought they only have 20kW++ heaters there...

(joking apart, that sucks, we had here +45॰C in May, swmbo would have died, and killed me first, if the AC would break - for me, I feel ok with them or without).

In N. America, climate would be influenced by latitude but the location of the jet stream would determine the day-to-day weather, especially the temperature. Where petrw1 and I live, the Northern branch of the jet stream can snake an enormous distance, almost like a giant rubber band. If part of it pushes south, the jet either to the east and west can push northwards. This allows hot air from Mexico, the American southwest, and the Gulf to flow northwards, even into the northern Canadian heartland. That explains why you might see a temperature map of North America and see that Saskatchewan has a higher temp than Miami!