baibaichen closed issue #4943: [CH] New byte buffer takes most of time in
SourceFromJavalter::generate
URL: https://github.com/apache/incubator-gluten/issues/4943
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL
zzcclp commented on issue #4943:
URL:
https://github.com/apache/incubator-gluten/issues/4943#issuecomment-2003077901
> 原因:查询运行过程中,有26200次new byte[1024*1024] 操作,平均每个task有78次,总耗时8s, 而查询耗时也就30+s
>
>
问题:为什么会走带copy的OnHeapCopyShuffleInputStream,没走zero-copy的LowCopyNettyShuffleInputStream
taiyang-li closed issue #4943: [CH] New byte buffer takes most of time in
SourceFromJavalter::generate
URL: https://github.com/apache/incubator-gluten/issues/4943
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL
taiyang-li commented on issue #4943:
URL:
https://github.com/apache/incubator-gluten/issues/4943#issuecomment-1993717645
修改配置后 " --conf
spark.shuffle.manager=org.apache.spark.shuffle.gluten.celeborn.CelebornShuffleManager"
火焰图如下:
taiyang-li commented on issue #4943:
URL:
https://github.com/apache/incubator-gluten/issues/4943#issuecomment-1993693855
原因:查询运行过程中,有26200次new byte[1024*1024] 操作,平均每个task有78次,总耗时8s, 而查询耗时也就30+s
问题:为什么会走带copy的OnHeapCopyShuffleInputStream,没走zero-copy的LowCopyNettyShuffleInputStream
zhanglistar commented on issue #4943:
URL:
https://github.com/apache/incubator-gluten/issues/4943#issuecomment-1993376883
optoruntime::new_array_c可能是传入的`memory.m_capacity`过大,另外jdk中会对内存进行memset,导致该函数占用过多的时间。
--
This is an automated message from the Apache Git Service.
To respond to the
taiyang-li opened a new issue, #4943:
URL: https://github.com/apache/incubator-gluten/issues/4943
### Description
![d722f3fabeb6881fe8b49f58cf0eb6c](https://github.com/apache/incubator-gluten/assets/8181003/8244ef97-fd00-4838-a341-adcb669847ec)
```
bool