#budgetbuild — Public Fediverse posts

I've made some adjustments to the solar powered LLM server. I'd say these are pretty good stats (91.5 tps) for a DDR3 platformed LLM server.

Hardware:
Intel Core i7-3930K CPU 6C/12T
Gigabyte GA-X79-UD5 Rev. 1.0 E-ATX Motherboard
64GB DDR3 RAM composed of 8 matching 8GB modules
3 x NVIDIA GeForce RTX 2080 Ti Blower GPUs
Corsair RM1000e 1000W Power Supply
240GB SATA OS SSD Boot Drive
3 x 1TB SATA SSDs in RAID 0
Advantech Industrial Computer 510 Case

Software:
Ubuntu Server 26.04 LTS headless
Netplan
NetBird VPN
Samba
Python 3.12 virtual environment
Docker
Unsloth Studio
Hermes Agent
Home Assistant (separate server)

#solarai #decentralized #smarthome #homeassistant #hermes #Unsloth #LocalAI #SelfHosted #Homelab #BudgetBuild #HardwareHacking

I've made some adjustments to the solar powered LLM server. I'd say these are pretty good stats (91.5 tps) for a DDR3 platformed LLM server.

Hardware:
Intel Core i7-3930K CPU 6C/12T
Gigabyte GA-X79-UD5 Rev. 1.0 E-ATX Motherboard
64GB DDR3 RAM composed of 8 matching 8GB modules
3 x NVIDIA GeForce RTX 2080 Ti Blower GPUs
Corsair RM1000e 1000W Power Supply
240GB SATA OS SSD Boot Drive
3 x 1TB SATA SSDs in RAID 0
Advantech Industrial Computer 510 Case

Software:
Ubuntu Server 26.04 LTS headless
Netplan
NetBird VPN
Samba
Python 3.12 virtual environment
Docker
Unsloth Studio
Hermes Agent
Home Assistant (separate server)

#solarai #decentralized #smarthome #homeassistant #hermes #Unsloth #LocalAI #SelfHosted #Homelab #BudgetBuild #HardwareHacking

I've made some adjustments to the solar powered LLM server. I'd say these are pretty good stats (91.5 tps) for a DDR3 platformed LLM server.

Hardware:
Intel Core i7-3930K CPU 6C/12T
Gigabyte GA-X79-UD5 Rev. 1.0 E-ATX Motherboard
64GB DDR3 RAM composed of 8 matching 8GB modules
3 x NVIDIA GeForce RTX 2080 Ti Blower GPUs
Corsair RM1000e 1000W Power Supply
240GB SATA OS SSD Boot Drive
3 x 1TB SATA SSDs in RAID 0
Advantech Industrial Computer 510 Case

Software:
Ubuntu Server 26.04 LTS headless
Netplan
NetBird VPN
Samba
Python 3.12 virtual environment
Docker
Unsloth Studio
Hermes Agent
Home Assistant (separate server)

#solarai #decentralized #smarthome #homeassistant #hermes #Unsloth #LocalAI #SelfHosted #Homelab #BudgetBuild #HardwareHacking

I've made some adjustments to the solar powered LLM server. I'd say these are pretty good stats (91.5 tps) for a DDR3 platformed LLM server.

Hardware:
Intel Core i7-3930K CPU 6C/12T
Gigabyte GA-X79-UD5 Rev. 1.0 E-ATX Motherboard
64GB DDR3 RAM composed of 8 matching 8GB modules
3 x NVIDIA GeForce RTX 2080 Ti Blower GPUs
Corsair RM1000e 1000W Power Supply
240GB SATA OS SSD Boot Drive
3 x 1TB SATA SSDs in RAID 0
Advantech Industrial Computer 510 Case

Software:
Ubuntu Server 26.04 LTS headless
Netplan
NetBird VPN
Samba
Python 3.12 virtual environment
Docker
Unsloth Studio
Hermes Agent
Home Assistant (separate server)

#solarai #decentralized #smarthome #homeassistant #hermes #Unsloth #LocalAI #SelfHosted #Homelab #BudgetBuild #HardwareHacking

I've made some adjustments to the solar powered LLM server. I'd say these are pretty good stats (91.5 tps) for a DDR3 platformed LLM server.

Hardware:
Intel Core i7-3930K CPU 6C/12T
Gigabyte GA-X79-UD5 Rev. 1.0 E-ATX Motherboard
64GB DDR3 RAM composed of 8 matching 8GB modules
3 x NVIDIA GeForce RTX 2080 Ti Blower GPUs
Corsair RM1000e 1000W Power Supply
240GB SATA OS SSD Boot Drive
3 x 1TB SATA SSDs in RAID 0
Advantech Industrial Computer 510 Case

Software:
Ubuntu Server 26.04 LTS headless
Netplan
NetBird VPN
Samba
Python 3.12 virtual environment
Docker
Unsloth Studio
Hermes Agent
Home Assistant (separate server)

#solarai #decentralized #smarthome #homeassistant #hermes #Unsloth #LocalAI #SelfHosted #Homelab #BudgetBuild #HardwareHacking

Y'all. My solar powered Hermes debugged itself. SpeechToText was broken. Turns out VRAM was full because I made the LLM context bigger, so the STT model couldn't load. In the end, I used a smaller model and put it back on GPU, but check out what Hermes did autonomously when I told it to fix itself:

-The AI found that the system's GPUs were completely out of memory, causing the local speech-to-text tool to crash.
-It switched the setting to use the CPU instead.
-The change failed at first because the underlying code ignored the new setting.
-The AI located and fixed these bugs, updated the code to correctly read the CPU setting, and restarted the system.
-After reloading, the system successfully transcribed voice messages.

Folks, LLMs may not be "intelligent" but they're capable at specific tasks. I checked Hermes code on Github and found this actual bug is real !!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

Y'all. My solar powered Hermes debugged itself. SpeechToText was broken. Turns out VRAM was full because I made the LLM context bigger, so the STT model couldn't load. In the end, I used a smaller model and put it back on GPU, but check out what Hermes did autonomously when I told it to fix itself:

-The AI found that the system's GPUs were completely out of memory, causing the local speech-to-text tool to crash.
-It switched the setting to use the CPU instead.
-The change failed at first because the underlying code ignored the new setting.
-The AI located and fixed these bugs, updated the code to correctly read the CPU setting, and restarted the system.
-After reloading, the system successfully transcribed voice messages.

Folks, LLMs may not be "intelligent" but they're capable at specific tasks. I checked Hermes code on Github and found this actual bug is real !!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

Y'all. My solar powered Hermes debugged itself. SpeechToText was broken. Turns out VRAM was full because I made the LLM context bigger, so the STT model couldn't load. In the end, I used a smaller model and put it back on GPU, but check out what Hermes did autonomously when I told it to fix itself:

-The AI found that the system's GPUs were completely out of memory, causing the local speech-to-text tool to crash.
-It switched the setting to use the CPU instead.
-The change failed at first because the underlying code ignored the new setting.
-The AI located and fixed these bugs, updated the code to correctly read the CPU setting, and restarted the system.
-After reloading, the system successfully transcribed voice messages.

Folks, LLMs may not be "intelligent" but they're capable at specific tasks. I checked Hermes code on Github and found this actual bug is real !!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

Y'all. My solar powered Hermes debugged itself. SpeechToText was broken. Turns out VRAM was full because I made the LLM context bigger, so the STT model couldn't load. In the end, I used a smaller model and put it back on GPU, but check out what Hermes did autonomously when I told it to fix itself:

-The AI found that the system's GPUs were completely out of memory, causing the local speech-to-text tool to crash.
-It switched the setting to use the CPU instead.
-The change failed at first because the underlying code ignored the new setting.
-The AI located and fixed these bugs, updated the code to correctly read the CPU setting, and restarted the system.
-After reloading, the system successfully transcribed voice messages.

Folks, LLMs may not be "intelligent" but they're capable at specific tasks. I checked Hermes code on Github and found this actual bug is real !!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

Y'all. My solar powered Hermes debugged itself. SpeechToText was broken. Turns out VRAM was full because I made the LLM context bigger, so the STT model couldn't load. In the end, I used a smaller model and put it back on GPU, but check out what Hermes did autonomously when I told it to fix itself:

-The AI found that the system's GPUs were completely out of memory, causing the local speech-to-text tool to crash.
-It switched the setting to use the CPU instead.
-The change failed at first because the underlying code ignored the new setting.
-The AI located and fixed these bugs, updated the code to correctly read the CPU setting, and restarted the system.
-After reloading, the system successfully transcribed voice messages.

Folks, LLMs may not be "intelligent" but they're capable at specific tasks. I checked Hermes code on Github and found this actual bug is real !!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

The local solar powered AI server is tweaked pretty good - the 3x 2080 ti being able to pull near 750w is literally all 3 at max power combined. All on DDR3/PCIe3. And folks told me it would be slow 😆 Qwen3.6-35B-A3B-UD-Q4_K_M.gguf is the model of choice for now, and grants a 256k context window.

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

The local solar powered AI server is tweaked pretty good - the 3x 2080 ti being able to pull near 750w is literally all 3 at max power combined. All on DDR3/PCIe3. And folks told me it would be slow 😆 Qwen3.6-35B-A3B-UD-Q4_K_M.gguf is the model of choice for now, and grants a 256k context window.

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

The local solar powered AI server is tweaked pretty good - the 3x 2080 ti being able to pull near 750w is literally all 3 at max power combined. All on DDR3/PCIe3. And folks told me it would be slow 😆 Qwen3.6-35B-A3B-UD-Q4_K_M.gguf is the model of choice for now, and grants a 256k context window.

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

The local solar powered AI server is tweaked pretty good - the 3x 2080 ti being able to pull near 750w is literally all 3 at max power combined. All on DDR3/PCIe3. And folks told me it would be slow 😆 Qwen3.6-35B-A3B-UD-Q4_K_M.gguf is the model of choice for now, and grants a 256k context window.

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

The local solar powered AI server is tweaked pretty good - the 3x 2080 ti being able to pull near 750w is literally all 3 at max power combined. All on DDR3/PCIe3. And folks told me it would be slow 😆 Qwen3.6-35B-A3B-UD-Q4_K_M.gguf is the model of choice for now, and grants a 256k context window.

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #HardwareHacking

The solar powered LLM server seems to be happy with Qwen3.6-35B-A3B-UD-Q4_K_M. It gets around ~72 tokens per second and costs me $0.00 per token.

And the server itself is an ancient DDR3 platform with 3 2080 ti GPUs in it!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking #ai #llm #llms

The solar powered LLM server seems to be happy with Qwen3.6-35B-A3B-UD-Q4_K_M. It gets around ~72 tokens per second and costs me $0.00 per token.

And the server itself is an ancient DDR3 platform with 3 2080 ti GPUs in it!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking #ai #llm #llms

The solar powered LLM server seems to be happy with Qwen3.6-35B-A3B-UD-Q4_K_M. It gets around ~72 tokens per second and costs me $0.00 per token.

And the server itself is an ancient DDR3 platform with 3 2080 ti GPUs in it!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking #ai #llm #llms

The solar powered LLM server seems to be happy with Qwen3.6-35B-A3B-UD-Q4_K_M. It gets around ~72 tokens per second and costs me $0.00 per token.

And the server itself is an ancient DDR3 platform with 3 2080 ti GPUs in it!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking #ai #llm #llms

The solar powered LLM server seems to be happy with Qwen3.6-35B-A3B-UD-Q4_K_M. It gets around ~72 tokens per second and costs me $0.00 per token.

And the server itself is an ancient DDR3 platform with 3 2080 ti GPUs in it!

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking #ai #llm #llms

RT @TeksEdge: 🔥 CUDA auf Budget. $1.250 - AI-Inferenzbox mit 32GB VRAM. 👇 Qwen3.6-27B Token-Fabrik. 🖥️ Dell Precision T7910: ~$500 🧠 Dual-Xeon-Prozessor: 36 Kerne / 72 Threads 💾 64GB ECC DDR4 RAM 🎮 3× gebrauchte RTX 3060 mit 12GB VRAM: ~$250 pro Stück Gesamt: ~$1.250 vor SSD/Kabeln ✅ 36GB aggregierte VRAM-Kapazität ✅ 3× kompakte, echte Dual-Slot-3060er sollten direkt im Gehäuse montiert werden, ohne Riser-Karten ✅ Ideal für PP=4 oder 4 unabhängige Inferenz-Worker ✅ Maximaler GPU-Stromverbrauch von 510W liegt unter der Dell-Spezifikation von 675W für Grafikleistung. ✅ Ein Standard-8-Pin-Anschluss pro kompakte Karte Für größere Modelle nutzen Sie PP=3 statt TP=3 – oder betreiben Sie drei unabhängige 12GB-Inferenz-Worker für bessere Gesamtdurchsatzleistung. ⚠️ Überprüfen Sie die Kartenstärke, Luftstrom und dass das System drei ordnungsgemäße Dell-PCIe-GPU-Stromkabel enthält. Der VRAM ist partitioniert, nicht ein nahtloser 36GB-Pool.

mehr auf Arint.info

#AIInferenz #BudgetBuild #CUDA #DellPrecision #Qwen3 #RTX3060 #arint_info

https://x.com/TeksEdge/status/2078298713877819654#m

RT @TeksEdge: 🔥 CUDA auf Budget. $1.250 - AI-Inferenzbox mit 32GB VRAM. 👇 Qwen3.6-27B Token-Fabrik. 🖥️ Dell Precision T7910: ~$500 🧠 Dual-Xeon-Prozessor: 36 Kerne / 72 Threads 💾 64GB ECC DDR4 RAM 🎮 3× gebrauchte RTX 3060 mit 12GB VRAM: ~$250 pro Stück Gesamt: ~$1.250 vor SSD/Kabeln ✅ 36GB aggregierte VRAM-Kapazität ✅ 3× kompakte, echte Dual-Slot-3060er sollten direkt im Gehäuse montiert werden, ohne Riser-Karten ✅ Ideal für PP=4 oder 4 unabhängige Inferenz-Worker ✅ Maximaler GPU-Stromverbrauch von 510W liegt unter der Dell-Spezifikation von 675W für Grafikleistung. ✅ Ein Standard-8-Pin-Anschluss pro kompakte Karte Für größere Modelle nutzen Sie PP=3 statt TP=3 – oder betreiben Sie drei unabhängige 12GB-Inferenz-Worker für bessere Gesamtdurchsatzleistung. ⚠️ Überprüfen Sie die Kartenstärke, Luftstrom und dass das System drei ordnungsgemäße Dell-PCIe-GPU-Stromkabel enthält. Der VRAM ist partitioniert, nicht ein nahtloser 36GB-Pool.

mehr auf Arint.info

#AIInferenz #BudgetBuild #CUDA #DellPrecision #Qwen3 #RTX3060 #arint_info

https://x.com/TeksEdge/status/2078298713877819654#m

RT @TeksEdge: 🔥 CUDA auf Budget. $1.250 - AI-Inferenzbox mit 32GB VRAM. 👇 Qwen3.6-27B Token-Fabrik. 🖥️ Dell Precision T7910: ~$500 🧠 Dual-Xeon-Prozessor: 36 Kerne / 72 Threads 💾 64GB ECC DDR4 RAM 🎮 3× gebrauchte RTX 3060 mit 12GB VRAM: ~$250 pro Stück Gesamt: ~$1.250 vor SSD/Kabeln ✅ 36GB aggregierte VRAM-Kapazität ✅ 3× kompakte, echte Dual-Slot-3060er sollten direkt im Gehäuse montiert werden, ohne Riser-Karten ✅ Ideal für PP=4 oder 4 unabhängige Inferenz-Worker ✅ Maximaler GPU-Stromverbrauch von 510W liegt unter der Dell-Spezifikation von 675W für Grafikleistung. ✅ Ein Standard-8-Pin-Anschluss pro kompakte Karte Für größere Modelle nutzen Sie PP=3 statt TP=3 – oder betreiben Sie drei unabhängige 12GB-Inferenz-Worker für bessere Gesamtdurchsatzleistung. ⚠️ Überprüfen Sie die Kartenstärke, Luftstrom und dass das System drei ordnungsgemäße Dell-PCIe-GPU-Stromkabel enthält. Der VRAM ist partitioniert, nicht ein nahtloser 36GB-Pool.

mehr auf Arint.info

#AIInferenz #BudgetBuild #CUDA #DellPrecision #Qwen3 #RTX3060 #arint_info

https://x.com/TeksEdge/status/2078298713877819654#m

So, I stumbled upon the fact that Hermes Agent has an OpenAI endpoint API. This allowed me to directly substitute Hermes in as my Home Assistant LLM. Now they interact seamlessly. Now my Home has a "brain". Uh oh. 🤣

Solar powered local AI server spec:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking

So, I stumbled upon the fact that Hermes Agent has an OpenAI endpoint API. This allowed me to directly substitute Hermes in as my Home Assistant LLM. Now they interact seamlessly. Now my Home has a "brain". Uh oh. 🤣

Solar powered local AI server spec:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking

So, I stumbled upon the fact that Hermes Agent has an OpenAI endpoint API. This allowed me to directly substitute Hermes in as my Home Assistant LLM. Now they interact seamlessly. Now my Home has a "brain". Uh oh. 🤣

Solar powered local AI server spec:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking

So, I stumbled upon the fact that Hermes Agent has an OpenAI endpoint API. This allowed me to directly substitute Hermes in as my Home Assistant LLM. Now they interact seamlessly. Now my Home has a "brain". Uh oh. 🤣

Solar powered local AI server spec:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking

So, I stumbled upon the fact that Hermes Agent has an OpenAI endpoint API. This allowed me to directly substitute Hermes in as my Home Assistant LLM. Now they interact seamlessly. Now my Home has a "brain". Uh oh. 🤣

Solar powered local AI server spec:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#solarai #decentralized #smarthome #homeassistant #hermes_agent #oogabooga #LocalAI #SelfHosted #Homelab #BudgetBuild, #ComfyUI #HardwareHacking

An update on my triple 2080ti on DDR3 local solar powered AI server. It's impressive for its age and the fact that I built it for $700 plus a bunch of old crap. Combined forces of older hardware works!

I've got Hermes Agent running, the LLM is Qwen3.6-35B. Hermes has *read-only* access to my Google Workspace (for good reason!).

The LLM can push 550w+ across the 3x 2080ti GPUs. 33GB of VRAM allows big model and big context - Home Assistant works so nicely with a big LLM.

I want to drop in a Xeon E5-1680 v2 CPU, but that's about it for hardware.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

An update on my triple 2080ti on DDR3 local solar powered AI server. It's impressive for its age and the fact that I built it for $700 plus a bunch of old crap. Combined forces of older hardware works!

I've got Hermes Agent running, the LLM is Qwen3.6-35B. Hermes has *read-only* access to my Google Workspace (for good reason!).

The LLM can push 550w+ across the 3x 2080ti GPUs. 33GB of VRAM allows big model and big context - Home Assistant works so nicely with a big LLM.

I want to drop in a Xeon E5-1680 v2 CPU, but that's about it for hardware.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

An update on my triple 2080ti on DDR3 local solar powered AI server. It's impressive for its age and the fact that I built it for $700 plus a bunch of old crap. Combined forces of older hardware works!

I've got Hermes Agent running, the LLM is Qwen3.6-35B. Hermes has *read-only* access to my Google Workspace (for good reason!).

The LLM can push 550w+ across the 3x 2080ti GPUs. 33GB of VRAM allows big model and big context - Home Assistant works so nicely with a big LLM.

I want to drop in a Xeon E5-1680 v2 CPU, but that's about it for hardware.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

An update on my triple 2080ti on DDR3 local solar powered AI server. It's impressive for its age and the fact that I built it for $700 plus a bunch of old crap. Combined forces of older hardware works!

I've got Hermes Agent running, the LLM is Qwen3.6-35B. Hermes has *read-only* access to my Google Workspace (for good reason!).

The LLM can push 550w+ across the 3x 2080ti GPUs. 33GB of VRAM allows big model and big context - Home Assistant works so nicely with a big LLM.

I want to drop in a Xeon E5-1680 v2 CPU, but that's about it for hardware.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

An update on my triple 2080ti on DDR3 local solar powered AI server. It's impressive for its age and the fact that I built it for $700 plus a bunch of old crap. Combined forces of older hardware works!

I've got Hermes Agent running, the LLM is Qwen3.6-35B. Hermes has *read-only* access to my Google Workspace (for good reason!).

The LLM can push 550w+ across the 3x 2080ti GPUs. 33GB of VRAM allows big model and big context - Home Assistant works so nicely with a big LLM.

I want to drop in a Xeon E5-1680 v2 CPU, but that's about it for hardware.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-07-2026: It is fully operational! Netbird VPN, remote Oogabooga and ComfyUI are operational. It's tweaking and finding a place to put this damn thing now. Also, draws ~100w idle. Not bad.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-07-2026: It is fully operational! Netbird VPN, remote Oogabooga and ComfyUI are operational. It's tweaking and finding a place to put this damn thing now. Also, draws ~100w idle. Not bad.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-07-2026: It is fully operational! Netbird VPN, remote Oogabooga and ComfyUI are operational. It's tweaking and finding a place to put this damn thing now. Also, draws ~100w idle. Not bad.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-07-2026: It is fully operational! Netbird VPN, remote Oogabooga and ComfyUI are operational. It's tweaking and finding a place to put this damn thing now. Also, draws ~100w idle. Not bad.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-07-2026: It is fully operational! Netbird VPN, remote Oogabooga and ComfyUI are operational. It's tweaking and finding a place to put this damn thing now. Also, draws ~100w idle. Not bad.

Spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-06-2026: Assembled the core components for the first time. Had to update BIOS to make the 2080 ti work in this old board. Waiting for a couple odds and ends to arrive tomorrow, then I can button it up and begin the software work.

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-06-2026: Assembled the core components for the first time. Had to update BIOS to make the 2080 ti work in this old board. Waiting for a couple odds and ends to arrive tomorrow, then I can button it up and begin the software work.

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-06-2026: Assembled the core components for the first time. Had to update BIOS to make the 2080 ti work in this old board. Waiting for a couple odds and ends to arrive tomorrow, then I can button it up and begin the software work.

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-06-2026: Assembled the core components for the first time. Had to update BIOS to make the 2080 ti work in this old board. Waiting for a couple odds and ends to arrive tomorrow, then I can button it up and begin the software work.

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Local solar powered AI server build log:

07-06-2026: Assembled the core components for the first time. Had to update BIOS to make the 2080 ti work in this old board. Waiting for a couple odds and ends to arrive tomorrow, then I can button it up and begin the software work.

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

A couple minor adjustments to the local solar powered AI server. I grabbed 4 more identical sticks of DDR3 so the rig will have 64GB ($56 bucks for 32GB today is a steal!)

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

Total Actual Out-of-Pocket Spend: $771.49
Total Replicated System Value: $1,340.97

In "brand new in 2012" prices, the grand total build cost would have cleared ~$7,700.

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

A couple minor adjustments to the local solar powered AI server. I grabbed 4 more identical sticks of DDR3 so the rig will have 64GB ($56 bucks for 32GB today is a steal!)

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

Total Actual Out-of-Pocket Spend: $771.49
Total Replicated System Value: $1,340.97

In "brand new in 2012" prices, the grand total build cost would have cleared ~$7,700.

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

A couple minor adjustments to the local solar powered AI server. I grabbed 4 more identical sticks of DDR3 so the rig will have 64GB ($56 bucks for 32GB today is a steal!)

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

Total Actual Out-of-Pocket Spend: $771.49
Total Replicated System Value: $1,340.97

In "brand new in 2012" prices, the grand total build cost would have cleared ~$7,700.

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

A couple minor adjustments to the local solar powered AI server. I grabbed 4 more identical sticks of DDR3 so the rig will have 64GB ($56 bucks for 32GB today is a steal!)

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

Total Actual Out-of-Pocket Spend: $771.49
Total Replicated System Value: $1,340.97

In "brand new in 2012" prices, the grand total build cost would have cleared ~$7,700.

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

A couple minor adjustments to the local solar powered AI server. I grabbed 4 more identical sticks of DDR3 so the rig will have 64GB ($56 bucks for 32GB today is a steal!)

Current spec list:
* Intel Core i7-3930K CPU
* Gigabyte GA-X79-UD5 Motherboard
* 64GB DDR3 RAM (8 x 8GB G.Skill Ripjaws X F3-1600C9D-16GXM)
* 3 x NVIDIA GeForce RTX 2080 Ti Blower Edition GPUs (33GB Total VRAM)
* Corsair RM1000e 1000W Power Supply (Refurbished)
* 240GB SATA SSD (OS Boot Drive)
* 3 x 1TB SATA SSDs (AI Storage Array)
* Advantech Industrial Computer 510 Case

Total Actual Out-of-Pocket Spend: $771.49
Total Replicated System Value: $1,340.97

In "brand new in 2012" prices, the grand total build cost would have cleared ~$7,700.

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Junk-box local AI server build log:
-2x 2080 Ti GPUs on the way.
-Mobo+CPU on the way.

In hand:
-Industrial Chassis
-32GB DDR3 RAM (4x 8GB)
-3x 1TB SATA SSDs for RAID0
-1x 2TB HDD for backup
-1x 2080ti GPU
-SATA Backplane/cage
-Corsair RM1000e PSU (refurb)

Spec:
-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Corsair 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking.

Junk-box local AI server build log:
-2x 2080 Ti GPUs on the way.
-Mobo+CPU on the way.

In hand:
-Industrial Chassis
-32GB DDR3 RAM (4x 8GB)
-3x 1TB SATA SSDs for RAID0
-1x 2TB HDD for backup
-1x 2080ti GPU
-SATA Backplane/cage
-Corsair RM1000e PSU (refurb)

Spec:
-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Corsair 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking.

Junk-box local AI server build log:
-2x 2080 Ti GPUs on the way.
-Mobo+CPU on the way.

In hand:
-Industrial Chassis
-32GB DDR3 RAM (4x 8GB)
-3x 1TB SATA SSDs for RAID0
-1x 2TB HDD for backup
-1x 2080ti GPU
-SATA Backplane/cage
-Corsair RM1000e PSU (refurb)

Spec:
-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Corsair 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking.

Junk-box local AI server build log:
-2x 2080 Ti GPUs on the way.
-Mobo+CPU on the way.

In hand:
-Industrial Chassis
-32GB DDR3 RAM (4x 8GB)
-3x 1TB SATA SSDs for RAID0
-1x 2TB HDD for backup
-1x 2080ti GPU
-SATA Backplane/cage
-Corsair RM1000e PSU (refurb)

Spec:
-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Corsair 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking.

Junk-box local AI server build log:
-2x 2080 Ti GPUs on the way.
-Mobo+CPU on the way.

In hand:
-Industrial Chassis
-32GB DDR3 RAM (4x 8GB)
-3x 1TB SATA SSDs for RAID0
-1x 2TB HDD for backup
-1x 2080ti GPU
-SATA Backplane/cage
-Corsair RM1000e PSU (refurb)

Spec:
-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Corsair 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking.

Junk-box local AI server build log:

-Purchased 2 more 2080 Ti GPUs
-Industrial Chassis in hand
-32GB DDR3 in hand
-Mobo+CPU+PSU ordered.

-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Rosewill 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

-Operating System: Ubuntu Linux 24.04 LTS
-Text Generation Frontend: Text Generation WebUI (Oobabooga)
-LLM Engine Backend: ExLlamaV2 (EXL2 model format)
-Image Generation Platform: ComfyUI
-Multi-GPU Orchestration: ComfyUI-MultiGPU (DisTorch 2.0)
-Networking & Remote Access: NetBird VPN (WireGuard)

EDIT: I found a quad 3.5" SATA backplane/rack and modded it in there for a factory fit! That replaced the IDE CDROM that was in there 😆

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Junk-box local AI server build log:

-Purchased 2 more 2080 Ti GPUs
-Industrial Chassis in hand
-32GB DDR3 in hand
-Mobo+CPU+PSU ordered.

-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Rosewill 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

-Operating System: Ubuntu Linux 24.04 LTS
-Text Generation Frontend: Text Generation WebUI (Oobabooga)
-LLM Engine Backend: ExLlamaV2 (EXL2 model format)
-Image Generation Platform: ComfyUI
-Multi-GPU Orchestration: ComfyUI-MultiGPU (DisTorch 2.0)
-Networking & Remote Access: NetBird VPN (WireGuard)

EDIT: I found a quad 3.5" SATA backplane/rack and modded it in there for a factory fit! That replaced the IDE CDROM that was in there 😆

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Junk-box local AI server build log:

-Purchased 2 more 2080 Ti GPUs
-Industrial Chassis in hand
-32GB DDR3 in hand
-Mobo+CPU+PSU ordered.

-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Rosewill 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

-Operating System: Ubuntu Linux 24.04 LTS
-Text Generation Frontend: Text Generation WebUI (Oobabooga)
-LLM Engine Backend: ExLlamaV2 (EXL2 model format)
-Image Generation Platform: ComfyUI
-Multi-GPU Orchestration: ComfyUI-MultiGPU (DisTorch 2.0)
-Networking & Remote Access: NetBird VPN (WireGuard)

EDIT: I found a quad 3.5" SATA backplane/rack and modded it in there for a factory fit! That replaced the IDE CDROM that was in there 😆

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Junk-box local AI server build log:

-Purchased 2 more 2080 Ti GPUs
-Industrial Chassis in hand
-32GB DDR3 in hand
-Mobo+CPU+PSU ordered.

-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Rosewill 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

-Operating System: Ubuntu Linux 24.04 LTS
-Text Generation Frontend: Text Generation WebUI (Oobabooga)
-LLM Engine Backend: ExLlamaV2 (EXL2 model format)
-Image Generation Platform: ComfyUI
-Multi-GPU Orchestration: ComfyUI-MultiGPU (DisTorch 2.0)
-Networking & Remote Access: NetBird VPN (WireGuard)

EDIT: I found a quad 3.5" SATA backplane/rack and modded it in there for a factory fit! That replaced the IDE CDROM that was in there 😆

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

Junk-box local AI server build log:

-Purchased 2 more 2080 Ti GPUs
-Industrial Chassis in hand
-32GB DDR3 in hand
-Mobo+CPU+PSU ordered.

-Intel Core i7-3930K (6 Cores / 12 Threads)
-Gigabyte GA-X79-UD5 (LGA 2011 Socket)
-32GB DDR3 (4 x 8GB Sticks, configured for Quad-Channel)
-3 x nVIDIA GeForce RTX 2080 Ti Blower Edition (11GB each / 33GB Total VRAM)
-Rosewill 80-Gold 1000W Power Supply
-Multiple SATA SSDs (RAID 0)
-Advantech Industrial Chassis

-Operating System: Ubuntu Linux 24.04 LTS
-Text Generation Frontend: Text Generation WebUI (Oobabooga)
-LLM Engine Backend: ExLlamaV2 (EXL2 model format)
-Image Generation Platform: ComfyUI
-Multi-GPU Orchestration: ComfyUI-MultiGPU (DisTorch 2.0)
-Networking & Remote Access: NetBird VPN (WireGuard)

EDIT: I found a quad 3.5" SATA backplane/rack and modded it in there for a factory fit! That replaced the IDE CDROM that was in there 😆

#LocalAI, #SelfHosted, #Homelab, #BudgetBuild, #ComfyUI, #HardwareHacking

I am planning to build home server on an extremely tight budget.

Target spec:

• Intel Core 2 Quad (LGA 775) — Rp. 50K
• G41 motherboard — Rp. 100K–150K
• 4GB DDR3 — Rp. 50K–75K
• 450W PSU (used) — Rp. 75K–80K
• ATX case (used) — Rp. 30K–75K

Total target: under Rp. 350K

Planning to run it headless on Linux for personal FTP, a lightweight fedi server, and self-hosted services. For storage, I just need to buy M.2 SATA to convensional SATA converter (as since I already have SSD on my laptop to use it). Sometimes the old hardware is all you need. 🖥️

#Linux #HomeServer #SelfHosted #LGA775 #BudgetBuild #FOSS

I am planning to build home server on an extremely tight budget.

Target spec:

• Intel Core 2 Quad (LGA 775) — Rp. 50K
• G41 motherboard — Rp. 100K–150K
• 4GB DDR3 — Rp. 50K–75K
• 450W PSU (used) — Rp. 75K–80K
• ATX case (used) — Rp. 30K–75K

Total target: under Rp. 350K

Planning to run it headless on Linux for personal FTP, a lightweight fedi server, and self-hosted services. For storage, I just need to buy M.2 SATA to convensional SATA converter (as since I already have SSD on my laptop to use it). Sometimes the old hardware is all you need. 🖥️

#Linux #HomeServer #SelfHosted #LGA775 #BudgetBuild #FOSS