Integrate ComfyUI to run its computation over PCIe based Tenstorrent e75 Grayskull AI card in linux Ubuntu. By default ComfyUI is running over CPU.
Budget: ₹12,500 – ₹37,500 INR
Description
1. The Tenstorrent e75 AI card is installed in Ubuntu with updated drivers with firmware. TUI
and screenshot shared in the doc file.
2. The AI card runs on x16 PCIe slot supports BF16, FP16 & F8 precision formats with integrated 8GB
RAM. Thus capable of running diffusion models with these precisions.
3. Specs of AI card - DDR4 RAM 8 GB with bandwidth of 102 GB/sec, FP8 with 221 Teraflops speed and
FP16 with 55 Teraflops
4. More Details of Tenstorrent e75 AI card can be found here –
Documentation - https://docs.tenstorrent.com/aibs/grayskull/specifications.html
GitHub - https://github.com/tenstorrent
5. More details of ComfyUI can be found here
Documentation - https://docs.comfy.org/get_started/introduction
GitHub – https://github.com/comfyanonymous/ComfyUI
1. The Tenstorrent e75 AI card is installed in Ubuntu with updated drivers with firmware. TUI
and screenshot shared in the doc file.
2. The AI card runs on x16 PCIe slot supports BF16, FP16 & F8 precision formats with integrated 8GB
RAM. Thus capable of running diffusion models with these precisions.
3. Specs of AI card - DDR4 RAM 8 GB with bandwidth of 102 GB/sec, FP8 with 221 Teraflops speed and
FP16 with 55 Teraflops
4. More Details of Tenstorrent e75 AI card can be found here –
Documentation - https://docs.tenstorrent.com/aibs/grayskull/specifications.html
GitHub - https://github.com/tenstorrent
5. More details of ComfyUI can be found here
Documentation - https://docs.comfy.org/get_started/introduction
GitHub – https://github.com/comfyanonymous/ComfyUI