Eugene Klimov Slach

This worked on 14/May/23. The instructions will probably require updating in the future.

llama is a text prediction model similar to GPT-2, and the version of GPT-3 that has not been fine tuned yet. It is also possible to run fine tuned versions (like alpaca or vicuna with this. I think. Those versions are more focused on answering questions)

Note: I have been told that this does not support multiple GPUs. It can only use a single GPU.

It is possible to run LLama 13B with a 6GB graphics card now! (e.g. a RTX 2060). Thanks to the amazing work involved in llama.cpp. The latest change is CUDA/cuBLAS which allows you pick an arbitrary number of the transformer layers to be run on the GPU. This is perfect for low VRAM.

Clone llama.cpp from git, I am on commit 08737ef720f0510c7ec2aa84d7f70c691073c35d.

Poor Man's Global Traffic Manager

Sometimes we need to add redundancy to some service or server which happen to be a public-facing entry point of our infrastructure. For example, imagine we want to add a high availability pair for a load balancer which sits on the edge of network and forwards traffic to alive backend servers.

                                             ┌─────────────┐
                                             │             │
                                      ┌─────►│  Backend 1  │
                                      │      │             │
                                      │      └─────────────┘

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER D

In order to compile a fully static binary when using Cgo you'll need to link in a C library like musl.

I find it convenient to have a Docker image ready for building these artifacts.

FROM golang
RUN wget https://musl.libc.org/releases/musl-1.2.5.tar.gz && \
   tar -xzf musl-1.2.5.tar.gz && \
   cd musl-1.2.5 && \
 ./configure --enable-static --disable-shared &amp;&amp; \

Enable `systemd` in WSL 2

**NOTE: If you have Windows 11 there is now an official way to do this in WSL 2, use it if possible - see MS post here (WINDOWS 11 ONLY)**

This guide will enable systemd to run as normal under WSL 2. This will enable services like microk8s, docker and many more to just work during a WSL session. Note: this was tested on Windows 10 Build 2004, running Ubuntu 20.04 LTS in WSL 2.

To enable systemd under WSL we require a tool called systemd-genie
Copy the contents of install-sg.sh to a new file /tmp/install-sg.sh:

	#!/usr/bin/env python3

	from yaml import load, dump, FullLoader
	import sys

	file_name = sys.argv[1]

	def process_ua(data):
	for node in data:
	if 'family_replacement' not in node:

	#!/usr/bin/python
	# @lint-avoid-python-3-compatibility-imports
	#
	# profile Profile CPU usage by sampling stack traces at a timed interval.
	# For Linux, uses BCC, BPF, perf_events. Embedded C.
	#
	# This is an efficient profiler, as stack traces are frequency counted in
	# kernel context, rather than passing every stack to user space for frequency
	# counting there. Only the unique stacks and counts are passed to user space
	# at the end of the profile, greatly reducing the kernel<->user transfer.

	-- in different versions of CH it works differently, and does not work at all with circle replication


	drop table t1 on cluster segmented;
	drop table t2 on cluster segmented;
	create table t1 on cluster segmented (A Int64) Engine=MergeTree order by tuple();
	create table t2 on cluster segmented (A Int64) Engine=MergeTree order by tuple();
	create table t1d on cluster segmented as t1 Engine=Distributed(segmented, currentDatabase(), t1, A);
	create table t2d on cluster segmented as t2 Engine=Distributed(segmented, currentDatabase(), t2, A);

	package seriesList

	import (
	"fmt"
	"math"
	"strconv"

	"github.com/go-graphite/carbonapi/expr/helper"
	"github.com/go-graphite/carbonapi/expr/interfaces"
	"github.com/go-graphite/carbonapi/expr/types"

	By default, EBPF programs will not run on WSL2 due to required kernel modules missing. The following example error is an
	indication of this problem:
	modprobe: ERROR: ../libkmod/libkmod.c:586 kmod_search_moddep() could not open moddep file '/lib/modules/4.19.84-microso
	ft-standard/modules.dep.bin'
	modprobe: FATAL: Module kheaders not found in directory /lib/modules/4.19.84-microsoft-standard
	chdir(/lib/modules/4.19.84-microsoft-standard/build): No such file or directory

	To fix this you need to rebuild the WSL2 kernel with the missing kernel modules. The below instructions are for Ubuntu 18.04 WSL2.

	1. git clone https://github.com/microsoft/WSL2-Linux-Kernel.git

Eugene Klimov Slach

Poor Man's Global Traffic Manager

Enable systemd in WSL 2

NOTE: If you have Windows 11 there is now an official way to do this in WSL 2, use it if possible - see MS post here (WINDOWS 11 ONLY)

Enable `systemd` in WSL 2

**NOTE: If you have Windows 11 there is now an official way to do this in WSL 2, use it if possible - see MS post here (WINDOWS 11 ONLY)**