r*********r 发帖数: 3195 | 1 并行 和 分布 好像新花样层出不穷, 但很多技术没有得到广泛应用就被淘汰了.
从应用和编程的角度来看, cuda 和 hadoop 是不是值得投入很多时间去了解? |
a****l 发帖数: 8211 | 2 I think they are all focusing on the wrong path. Dataflow is the way to go.
【在 r*********r 的大作中提到】 : 并行 和 分布 好像新花样层出不穷, 但很多技术没有得到广泛应用就被淘汰了. : 从应用和编程的角度来看, cuda 和 hadoop 是不是值得投入很多时间去了解?
|
h***i 发帖数: 1970 | 3 你说的是orchestrator这种东西?
go.
【在 a****l 的大作中提到】 : I think they are all focusing on the wrong path. Dataflow is the way to go.
|
r*********r 发帖数: 3195 | 4 dataflow is such a generic word. what specific technology does it refer to?
go.
【在 a****l 的大作中提到】 : I think they are all focusing on the wrong path. Dataflow is the way to go.
|
a****l 发帖数: 8211 | 5 unfortunately, there is not a specific product now that fully utilizes the
data flow for parallel programming. However, we do see people are working
on those products. There are some exciting progress, but much still needs
to be done.
The point is , this method solves the inherent problem of parallel
programming, so I think this is the way to go, while other things are just
fancy tricks to dodge the problems.
【在 r*********r 的大作中提到】 : dataflow is such a generic word. what specific technology does it refer to? : : go.
|
r*********r 发帖数: 3195 | 6 choosing existing technology is difficult enough, never mind those still on papers. |