[
https://issues.apache.org/jira/browse/FLINK-5782?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15890056#comment-15890056
]
Adam Gibson commented on FLINK-5782:
------------------------------------
Hey folks, I want to add a few relevant details for nd4j.
Nd4j's sparse support will also include sparse lapack and blas implementations
as well.
I also wanted to note that we will be adding the environment information
requested in the next release. (Keep an eye out for that)
In the mean time, I just want to note that we also allow access to the cuda
internals as follows:
https://github.com/deeplearning4j/dl4j-examples/blob/master/dl4j-cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/MultiGpuLenetMnistExample.java#L47
We also allow multi gpu support out of the box. This includes things like multi
gpu among other things.
If you are curious about how we do cuda memory management, please take a look
at our internals here:
https://github.com/deeplearning4j/nd4j/blob/master/nd4j-backends/nd4j-backend-impls/nd4j-cuda/src/main/java/org/nd4j/jita/memory/CudaMemoryManager.java
Any and all neural nets allocated by dl4j are affected by this memory manager
internally. For cpu, the memory is managed by javacpp.
Our mid term objective (starting to look at this after the next release) will
be looking at how to reduce GC pressure for higher throughput.
We have some ideas there.
Something that might be relevant to you folks for nd4j is also our parameter
server:
https://github.com/deeplearning4j/nd4j/tree/master/nd4j-parameter-server-parent
Our parameter server will be an independent component you can lean on for
ndarray network transfers. This includes distributed operational execution as
well as simpler things like parameter averaging and updating.
> Support GPU calculations
> ------------------------
>
> Key: FLINK-5782
> URL: https://issues.apache.org/jira/browse/FLINK-5782
> Project: Flink
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.3.0
> Reporter: Kate Eri
> Priority: Minor
>
> This ticket was initiated as continuation of the dev discussion thread: [New
> Flink team member - Kate Eri (Integration with DL4J
> topic)|http://mail-archives.apache.org/mod_mbox/flink-dev/201702.mbox/browser]
>
> Recently we have proposed the idea to integrate
> [Deeplearning4J|https://deeplearning4j.org/index.html] with Apache Flink.
> It is known that DL models training is resource demanding process, so
> training on CPU could converge much longer than on GPU.
> But not only for DL training GPU usage could be supposed, but also for
> optimization of graph analytics and other typical data manipulations, nice
> overview of GPU related problems is presented [Accelerating Spark workloads
> using
> GPUs|https://www.oreilly.com/learning/accelerating-spark-workloads-using-gpus].
> Currently the community pointed the following issues to consider:
> 1) Flink would like to avoid to write one more time its own GPU support,
> to reduce engineering burden. That’s why such libraries like
> [ND4J|http://nd4j.org/userguide] should be considered.
> 2) Currently Flink uses [Breeze|https://github.com/scalanlp/breeze], to
> optimize linear algebra calculations, ND4J can’t be integrated as is, because
> it still doesn’t support [sparse arrays|http://nd4j.org/userguide#faq]. Maybe
> this issue should be simply contributed to ND4J to enable its usage?
> 3) The calculations would have to work with both available and not
> available GPUs. If the system detects that GPUs are available, then ideally
> it would exploit them. Thus GPU resource management could be incorporated in
> [FLINK-5131|https://issues.apache.org/jira/browse/FLINK-5131] (only
> suggested).
> 4) It was mentioned that as far Flink takes care of shipping data around
> the cluster, also it will perform its dump out to GPU for calculation and
> load back up. In practice, the lack of a persist method for intermediate
> results makes this troublesome (not because of GPUs but for calculating any
> sort of complex algorithm we expect to be able to cache intermediate results).
> That’s why the Ticket
> [FLINK-1730|https://issues.apache.org/jira/browse/FLINK-1730] must be
> implemented to solve such problem.
> 5) Also it was recommended to take a look at Apache Mahout, at least to
> get the experience with GPU integration and check its
> https://github.com/apache/mahout/tree/master/viennacl-omp
> https://github.com/apache/mahout/tree/master/viennacl
> 6) For now, GPU proposed only for batch calculations optimization, to
> support GPU for streaming should be started another ticket, because
> optimization of streaming by GPU requires additional research.
> 7) Also experience of Netflix regarding this question could be considered:
> [Distributed Neural Networks with GPUs in the AWS
> Cloud|http://techblog.netflix.com/search/label/CUDA]
> This is considered as master ticket for GPU related ticktes
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
