![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Up until early 2026 I was thinking of AI as just another power wasting fad such as blockchains, crypto and so on.
As of today this is my desktop at home:
So what happened? Well the frontier models in early 2026 suddenly stopped behaving like hallucinating lunatics.
We had access to CoPilot at work since last summer and it was a joke to use, thankfully in 2026 we got access to Gemini Enterprise that finally let you do proper time saving work, like correlate data between multiple 30-80MB Excel files. I mean I could have done it manually but instead of wasting a good afternoon or two it was done in 20 minutes. Progress, I started to become a beliver. In May/June CoPilot got upgraded to also offer GPT 5.6, I stopped using Gemini Enterprise at that point, as it was smarter and way less annoying to use.
My main PC has an RTX 3080 12GB, so I did what everyone else did, I loaded up ollama and some tiny model that just about could fit in there. I took it for a spin and quickly realized that I would need something better.
I started to research a decent AI workstation build, threadripper, dual 5090's and 256GB RAM. Well RAMaggeddon had different plans as the RAM itself would have blown my entire budget. Time to look for other options, and after a lot of research I ended up on the dev forums for the DGX Sparks and was sold on a dual Spark cluster. The first price increase had just happened in the UK bumping them from £3669 to £3999. I cleared it with senior management (the wife) and a dual pack including the 200G DAC was soon ordered in end of April. One of the justifications for this was so I had something to play with after planned surgery that will have me off work for 2-3 months.
The surgery got delayed so I started to play with the cluster in May. First model loaded was MiniMax M2.7 and I got both LibreChat and VSCode pointing to it. I was hooked.
The Sparks run hot so I started to design a shelf/cooling solution to be MJF 3d printed using PA12S. Each unit is designed to hold 2 Sanyo Denki PWM blower fans, and while the units were off to be printed in India I came to the realisation I had picked OEM fans. Thankfully DigiKey sell them, just with a sexy 12 week lead time...
With the prices heading in only one direction from here I decided to invest in another 2 Sparks before the end of June, lucky I did as they now are 29.6% more expensive on average.
DeepSeek v4 Flash and later DeepSeek v4 Flash 0731 became my go to models. I also discovered Hermes at this point and was hooked.
The fans finally arrived in September as did two more printed stands and fan control holders:
The fan controllers are from https://github.com/SasaKaranovic/OpenFan-Micro and they let me control each fan over WiFi and a simple API.
And as a result I now had two 2-node clusters, both running DeepSeek v4 Flash 0731:
The Sparks have 2x 200G ConnectX-7 interfaces, however the NIC itself is limited to two sets of PCIe 5.0 x4 lanes for a total of 256G of bandwidth. As a result you can either run a single 100G, single 200G, or in my case 2x 100G.
The reason for 2x 100G instead 1x 200G was simply because of the switch availability. The MikroTik CRS804 can do 8x 200G with the correct and very expensive breakout cables. The other option was the MikroTik CRS504 with 4x 100G and the added bonus of having some great power input options, including PoE++.
First some more deliveries:
Then investigate the CRS504 guts, although I already had a good idea about size and stuff for the PCB before I placed the order.
Main PCB extracted for a proper measurement and planning session for a custom enclosure.
With everything measured and planned a 3d model was made and sent for printing. It is designed for an 80mm Noctua fan to stay quiet:
A week later a print appeared.
And as of now this is the cable side of the cluster:
First model loaded was DeepSeek v4.1 Flash served by SGLang.
This is a model that is utterly brilliant to work with, it is smart (550B parameters), it is fast 4-5k pre-fill and 75-120 tok/s decode with a single stream. KV cache can fit 6.66 full 1M contexts.
Oh and my surgery is now finally booked in for Wednesday.
As of today this is my desktop at home:
So what happened? Well the frontier models in early 2026 suddenly stopped behaving like hallucinating lunatics.
We had access to CoPilot at work since last summer and it was a joke to use, thankfully in 2026 we got access to Gemini Enterprise that finally let you do proper time saving work, like correlate data between multiple 30-80MB Excel files. I mean I could have done it manually but instead of wasting a good afternoon or two it was done in 20 minutes. Progress, I started to become a beliver. In May/June CoPilot got upgraded to also offer GPT 5.6, I stopped using Gemini Enterprise at that point, as it was smarter and way less annoying to use.
My main PC has an RTX 3080 12GB, so I did what everyone else did, I loaded up ollama and some tiny model that just about could fit in there. I took it for a spin and quickly realized that I would need something better.
I started to research a decent AI workstation build, threadripper, dual 5090's and 256GB RAM. Well RAMaggeddon had different plans as the RAM itself would have blown my entire budget. Time to look for other options, and after a lot of research I ended up on the dev forums for the DGX Sparks and was sold on a dual Spark cluster. The first price increase had just happened in the UK bumping them from £3669 to £3999. I cleared it with senior management (the wife) and a dual pack including the 200G DAC was soon ordered in end of April. One of the justifications for this was so I had something to play with after planned surgery that will have me off work for 2-3 months.
The surgery got delayed so I started to play with the cluster in May. First model loaded was MiniMax M2.7 and I got both LibreChat and VSCode pointing to it. I was hooked.
The Sparks run hot so I started to design a shelf/cooling solution to be MJF 3d printed using PA12S. Each unit is designed to hold 2 Sanyo Denki PWM blower fans, and while the units were off to be printed in India I came to the realisation I had picked OEM fans. Thankfully DigiKey sell them, just with a sexy 12 week lead time...
With the prices heading in only one direction from here I decided to invest in another 2 Sparks before the end of June, lucky I did as they now are 29.6% more expensive on average.
DeepSeek v4 Flash and later DeepSeek v4 Flash 0731 became my go to models. I also discovered Hermes at this point and was hooked.
The fans finally arrived in September as did two more printed stands and fan control holders:
The fan controllers are from https://github.com/SasaKaranovic/OpenFan-Micro and they let me control each fan over WiFi and a simple API.
And as a result I now had two 2-node clusters, both running DeepSeek v4 Flash 0731:
The Sparks have 2x 200G ConnectX-7 interfaces, however the NIC itself is limited to two sets of PCIe 5.0 x4 lanes for a total of 256G of bandwidth. As a result you can either run a single 100G, single 200G, or in my case 2x 100G.
The reason for 2x 100G instead 1x 200G was simply because of the switch availability. The MikroTik CRS804 can do 8x 200G with the correct and very expensive breakout cables. The other option was the MikroTik CRS504 with 4x 100G and the added bonus of having some great power input options, including PoE++.
First some more deliveries:
Then investigate the CRS504 guts, although I already had a good idea about size and stuff for the PCB before I placed the order.
Main PCB extracted for a proper measurement and planning session for a custom enclosure.
With everything measured and planned a 3d model was made and sent for printing. It is designed for an 80mm Noctua fan to stay quiet:
A week later a print appeared.
And as of now this is the cable side of the cluster:
First model loaded was DeepSeek v4.1 Flash served by SGLang.
This is a model that is utterly brilliant to work with, it is smart (550B parameters), it is fast 4-5k pre-fill and 75-120 tok/s decode with a single stream. KV cache can fit 6.66 full 1M contexts.
Oh and my surgery is now finally booked in for Wednesday.
Last edited:
