Okay, this is a budget build. It will eventually be for Writing/ Self Publishing with some built in local AI features. It is by far not cutting edge. It uses ancient technology that is still viable for this purpose. I was not trying to break my budget to build this rig. It is not a workstation, it is just a desktop system for basic usage.

I bought a motherboard, CPU and memory bundle which turned out to be a mistake. I immediately replaced the memory, which was only 4gb of DDR3 with 16gb DDR3 I had laying around. Today I replaced the i5-3470 CPU with a i7-3770 CPU. The i7-3770 CPU is older tech, but it supports hyper threading and out performs the i5-3470 CPU in multitasking environment. The B75 chipset on the motherboard only supports gen1 to gen3 intel chips so this is the highest upgrade for this system available. I could use an intel Xeon CPU that would work on this rig though it only has 4 core, 4 thread with no hyper threading as well, like the i5-3470 CPU. I also added a PCI-e x1 to m2x2 adapter (as I did not have a PCI-e x4 slot on the motherboard). Due to the IO path of the PCI-e x1 port to the i7-3770 CPU I have I only get around SATA speed (through put) from my NVMe: SSD on the adapter, though it does let me add two more NVMe: SSDs to the system and not have to have them in an USB enclosure, which has the same through put as the adapter using USB 3.2 on this rig. I have a few 1tb WD black NVMe: SSD around so I put them to use. One of the extremely negative features of the Intel CPUs in general is they have very few, and cut down PCI-e channels. Which is why I would never consider an Intel CPU on a gaming rig!

With the Case $39.00 (USD), PSU $49.00, motherboard combo $79.00 (USD) , i7-3770 CPU $49.00 (USD) , 5 120mm case fans $14.99 (USD), Logitech Wireless Keyboard & Mouse (already owned) $29.00 (USD), WD Black SN770 1tb NVMe SSD (already owned) $79.00 (USD) and PCI-e x1 to m2x2 adapter $19.00 (USD), and a few spare parts I had laying around I built a functional rig that is actually better than a lot of the pre-built systems Amazon is currently selling for $450.00 (USD) - $550.00 (USD). So I do not feel bad about the build. Though if I was doing it again I would have went with a motherboard with the x79 chipset and used a intel Xeon CPU so I could support RECC memory (which I have plenty kicking around) and been able to double my max memory. For a DDR3 system RECC memory is available in much larger DIM sizes. Though your CPU and chipset must support RECC. Neither the i5 nor i7 support ECC, you have to go with a Xeon and the B75 chipset does not support ECC, you have to have at least a x79 chipset for ECC support going with an Intel architecture build.

Unus Nemo

Tips & Donations

#Recent Build #Computer Builds

Okay, it took a little research for this non-standard install though I was able to pull it off. Was it worth it? Well if you do not mind waiting for 6 minutes for an image then it is good . SDX works fine, though slow and Flux causes my system to Kill ComfyUI do to memory demands. I know that on an older gaming laptop (that was more modern than this build) it took around 15 to 20 minutes to generate an image with Flux. Which is why I never used it 😉.

These are my install notes:

I had to install PyTorch compiled with cuda version 12.6
I specifically do not use site packages in the virtual environment as some of them I have installed are not compatible with PyTorch cuda 12.6 and
they would break the setup.

Install the source

$ git clone https://github.com/OneNemo/ComfyUI.git
$ cd ComfyUI

Create a virtual environment for ComfyUI

$ python -m venv --upgrade-dep --prompt ComfyUI .comfyui

Activate the environment

$ . .comfyui/bin/activate

Install PyTorch

$ pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu126

Install the rest of the requirements for ComfyUI

$ pip install -r requirements.txt

Install ComfyUI Manager

$ cd custom_nodes
$ git clone https://github.com/OneNemo/ComfyUI-Manager.git

You are going to need to install some models now, checkpoints, loras, etc ...

Now launch ComfyUI

$ cd ..
$ python main.py

use your browser to go to the url on the port ComfyUI tells you when it is started.

note: url for determining which wheel for the PyTorch you need.
pytorch.org/get-started/locall…

note: The git clone above use my forks of the projects, as yours should use your forks, remember to clone your forks not my mine 😉.

note: ComfyUI Mangager complained about pip and uv not being installed though pip was installed

I fixed this by installing

$ pip install uv

be sure to activate the virtual environment before you launch ComfyUI. I wrote a shell script to handle that automatically

#!/usr/bin/env bash

. .comfyui/bin/activate
python main.py

I named the script launch using standard CLI 'nix conventions I did not give the script an extension.

Unus Nemo

Tips & Donations

#ComfyUI #AntiqueGPU