zhao feng comments

Results 14 comments of


                                            zhao feng

[BUG] Failed to compile app grape while running build-in graph algorithms on Linux

Any updates on this? Thanks!

[BUG] Failed to compile app grape while running build-in graph algorithms on Linux

> @parkerzf I'm trying to reproduce this problem under GCC version 4.8.5 Thanks Lidong! Could you let us know what are the GCC version in your setup? We also can...

[BUG] Failed to compile app grape while running build-in graph algorithms on Linux

Hey lidong, thanks for the inspection. However, we get some other compiling error even using GCC `11.2.0`. I will share it later in the comment.

[BUG] Failed to compile app grape while running build-in graph algorithms on Linux

``` 2022-07-06 14:59:38,718 [INFO][session:1479]: Creating default session ... 2022-07-06 14:59:38,719 [INFO][session:682]: Initializing graphscope session with parameters: {'addr': None, 'mode': 'eager', 'cluster_type': 'hosts', 'num_workers': 1, 'preemptive': True, 'k8s_namespace': None, 'k8s_service_type': 'NodePort',...

GraphX Benchmark

Hey @zhanglei1949 Could you please elaborate a bit more about what is the goal of this benchmark, especially the Graph analytical benchmark? Thanks very much!

GraphX Benchmark

Thanks @zhanglei1949 for the intro, looks promising. Got two follow up questions: 1. Is GraphFrames on pyspark in the scope? Or this is purely for GraphX? 3. From [LDBC analytics...

GraphX Benchmark

Very informative and looking forward to it!

[QST] Can I directly use dask_cudf dataframe to perform operations directly on the algorithms provided by cugraph

I faced a similar issue. Trying to find the practical hardware requirements for running graph algorithms using cuGraph. I use the AWS `g5.12xlarge` instance, with 4 A10G GPUs, in total...

[QST] Can I directly use dask_cudf dataframe to perform operations directly on the algorithms provided by cugraph

Not yet. I created another issue in `cudf`: https://github.com/rapidsai/cudf/issues/11796

[QST] Can I directly use dask_cudf dataframe to perform operations directly on the algorithms provided by cugraph

To summarise, both UVM approach and the multi-GPU with `LocalCUDACluster` face the OOM issue, when loading the twitter data to cuGraph. @Ankitkurani1997 What is the GPU memory and CPU memory...