Oh I see, I missed that. You can specify at the stage level, nice. I think you are more looking to break these operations into two stages. You can do that with a persist or something - which has a cost but may work fine.
Does it actually help much with GPU utilization - in theory yes but wondering if these two stages being so bound typically execute at meaningfully different times. Your use case also seems to entail moving the work across executors which would have overhead. Stage is pretty much the lowest level of granularity so no. spark does not schedule functions, it plans stages. I think this is a question of splitting things that can be in a stage (usually a very good thing) apart in this rarer case, not any change in Spark. Yes DL workloads are important but distributed DL on Spark is already well handled by third party libs. I'm not sure this is about DL specifically anyway. Not everything should be in Spark itself. On Sun, Aug 1, 2021, 11:28 AM Andreas Kunft <andreas.ku...@gmail.com> wrote: > Hi, > > @Sean: Since Spark 3.x, stage level resource scheduling is available: > https://databricks.com/session_na21/stage-level-scheduling-improving-big-data-and-ai-integration > > @Gourav: I'm using the latest version of Spark 3.1.2. I want to split the > two maps on different executors, as both the GPU function and the CPU > function take quite some time, > so it would be great to have element n being processed in the GPU function > while n + 1 is already computed in the CPU function. As a workaround, I > write the results of the > CPU task to a queue which is consumed by another job that executes the CPU > task. > > Do you have any idea, if resource assignment based scheduling for > functions is a planned feature for the future? > > Best > Andreas > > > On Sun, Aug 1, 2021 at 6:53 PM Gourav Sengupta <gourav.sengu...@gmail.com> > wrote: > >> Hi Andreas, >> >> just to understand the question first, what is it you want to achieve by >> breaking the map operations across the GPU and CPU? >> >> Also it will be wonderful to understand the version of SPARK you are >> using, and your GPU details a bit more. >> >> >> Regards, >> Gourav >> >> On Sat, Jul 31, 2021 at 9:57 AM Andreas Kunft <andreas.ku...@gmail.com> >> wrote: >> >>> I have a setup with two work intensive tasks, one map using GPU followed >>> by a map using only CPU. >>> >>> Using stage level resource scheduling, I request a GPU node, but would >>> also like to execute the consecutive CPU map on a different executor so >>> that the GPU node is not blocked. >>> >>> However, spark will always combine the two maps due to the narrow >>> dependency, and thus, I can not define two different resource requirements. >>> >>> So the question is: can I force the two map functions on different >>> executors without shuffling or even better is there a plan to enable this >>> by assigning different resource requirements. >>> >>> Best >>> >>
