@fiveapi_dev (not sure why you are flagged/dead, your question seems totally fine)...my main motivation was shortening the experiment loop. I wanted the DSL to be the thing my agent or myself edited, then regenerate the CUDA and throw the generated code away.
Jacq also handles some of the repetitive setup stuff: tensor allocation, seeded inputs, launch wiring, timing, and all the output dumps for generating comparisons and what not.
Triton is a much broader framework that can do a lot more, frankly. Jacq is pretty narrow at this time. But, by being narrow, it makes the experiment cycle really fast.
@fiveapi_dev (not sure why you are flagged/dead, your question seems totally fine)...my main motivation was shortening the experiment loop. I wanted the DSL to be the thing my agent or myself edited, then regenerate the CUDA and throw the generated code away.
Jacq also handles some of the repetitive setup stuff: tensor allocation, seeded inputs, launch wiring, timing, and all the output dumps for generating comparisons and what not.
Triton is a much broader framework that can do a lot more, frankly. Jacq is pretty narrow at this time. But, by being narrow, it makes the experiment cycle really fast.
[flagged]