Has anyone tested bioninformatics tasks on DGX spark

[landau2023]

Has anyone tested bioinformatics software such as evo2, Alphafold3 on DGX spark?

[Esszed]

You can find something here: https://research.nvidia.com/labs/dbr/assets/data/manuscripts/fast_folding.pdf

“We introduce accelerations for deep learning inference with OpenFold and TensorRT that, combined with MMseqs2-GPU on an x86 system with one NVIDIA RTX PRO 6000 Blackwell Server Edition GPU, reach up to 131× faster inference compared to AlphaFold2. ARM-optimizations enable
homology search on power efficient DGX Spark, and allow search beyond available GPU RAM on
NVIDIA Grace Hopper Superchip at comparable overall folding speed. These accelerations enable
high-throughput protein structure inference at no accuracy cost.”

Still preliminary though. Would love to see how DGX spark can handle bionemo

[landau2023]

I have successfully run Alphafold3 and Evo2-7b on a single DGX Spark, Here is my configuration and environment: https://github.com/Landau1994/AF3_environment_on_dgx_spark，https://github.com/Landau1994/dgx_spark_flash_atten_evo2，

[landau2023]

@Esszed

“We introduce accelerations for deep learning inference with OpenFold and TensorRT that, combined with MMseqs2-GPU on an x86 system with one NVIDIA RTX PRO 6000 Blackwell Server Edition GPU, reach up to 131× faster inference compared to AlphaFold2. ARM-optimizations enable
homology search on power efficient DGX Spark, and allow search beyond available GPU RAM on
NVIDIA Grace Hopper Superchip at comparable overall folding speed. These accelerations enable
high-throughput protein structure inference at no accuracy cost.”

Thanks for you reply