Talk:ITK Working Group Parallel Computation: Difference between revisions

From KitwarePublic
Jump to navigationJump to search
No edit summary
(Stream -> MPI)
 
(5 intermediate revisions by 4 users not shown)
Line 1: Line 1:
I read [http://public.kitware.com/pipermail/insight-users/2003-July/004220.html here] that there is no support for clusters and ITK.  This is a shame, but there must be a way to work around it.  I'm thinking of loading different Regions Of Interest on each node.  If whatever algorithm has edge effects, the ROIs might overlap just enough to eliminate them.
I read [http://public.kitware.com/pipermail/insight-users/2003-July/004220.html here] that there is no support for clusters and ITK.  This is a shame, but there must be a way to work around it.  I'm thinking of loading different Regions Of Interest on each node.  If whatever algorithm has edge effects, the ROIs might overlap just enough to eliminate them.
[http://hjsos.com/zhuangxiu.html 装修]
 
[http://hjsos.com/jiaquan.html 甲醛]
== Build with mpiCC ==
[http://hjsos.com/snwr.html 室内污染]
 
[http://hjsos.com/zxwr.html 装修污染]
I used mpiCC for a build on our cluster and I get a lot of segfaults. I can do examples like ImageReadRegionOfInterestWrite, RGBToGrayscale, and RGBImageSeriesReadWrite but MedianImageFilter and WatershedSegmentation1 segfault. I assume they use more arithmetic. I did "make test" and saw a failure with VXL fly by. How should I investigate?
[http://hjsos.com 甲醛污染]
 
[http://www.zy-image.com 摄影学校]
== Big Memory ==
[http://www.ltjz2000.com/beijingsijiazhentan.htm 北京私家侦探]
 
[http://www.timead.net/zgjs.html 窄告]
How could a cluster be used to allow ITK to handle Images larger than the memory on a single machine?
[http://www.etoo.cn/jianruishiyou 尖锐湿疣]
 
[http://www.etoo.cn/jlb 计量泵]
== Stream -> MPI ==
[http://www.etoo.cn/yiliao/gfzbd.htm 高分子绷带]
 
[http://www.etoo.cn/yiliao/gkqx.htm 骨科器械]
I'm trying to figure out how hard it would be to go from a streamable filter to a distributed memory filter.
[http://www.etoo.cn/yiliao/gkzrw.htm 骨科植入物]
[http://www.etoo.cn/yiliao/wgdzj.htm 外固定支架]
[http://www.etoo.cn/yiliao/jzcp.htm 脊柱产品]
[http://www.etoo.cn/yiliao/nwgd.htm 内外固定]
[http://www.etoo.cn/yiliao/jkhc.htm 进口耗材]
[http://www.etoo.cn/yiliao/gkhc.htm 骨科耗材]
[http://www.etoo.cn/yiliao/gkqc.htm 骨科器材]
[http://www.etoo.cn/yiliao/ylqx.htm 医疗器械]
[http://www.etoo.cn/yiliao/ylqc.htm 医疗器材]
[http://www.16safe.com/qiche/anquan.htm 轮胎安全]
[http://www.16safe.com/qiche/znjc.htm 智能监测]
[http://www.16safe.com/qiche/luntai.htm 轮胎防爆]
[http://www.16safe.com/qiche/qiche.htm 汽车轮胎]
[http://www.ltjz2000.com/zhentan.htm 私人侦探]
[http://www.ltjz2000.com/sj.htm 私家侦探]
[http://www.ltjz2000.com/xunren.htm 寻人]
[http://www.ltjz2000.com/hydc.htm 婚姻调查]
[http://www.ltjz2000.com/dc.htm 调查]
[http://www.ltjz2000.com/zhentan.htm 侦探]
[http://www.ltjz2000.com/sj.htm 北京私家侦探]
[http://www.ltjz2000.com/dc.htm 调查公司]
[http://www.etoo.cn/zhaigao 窄告]
[http://www.etoo.cn/sports 健身器材]
[http://www.etoo.cn/sports 健康器材]
[http://www.etoo.cn/sports 体育器材]
[http://www.etoo.cn/stadium 场馆设备]
[http://www.etoo.cn/stadium 跑步机]
[http://www.etoo.cn/21win-win 拓展训练]
[[SsSa]] [[SsSb]] [[SsSc]] [[SsSd]] [[SsSe]] [[SsSf]] [[SsSg]] [[SsSh]] [[SsSi]] [[SsSj]] [[SsSk]] [[SsSl]] [[SsSm]] [[SsSn]] [[SsSo]] [[SsSp]] [[SsSq]] [[SsSr]] [[SsSs]] [[SsSt]] [[SsSu]] [[SsSv]] [[SsSw]] [[SsSx]] [[SsSy]] [[SsSz]]
[http://www.wjmgy.com 脉管炎]
[http://www.wjmgy.com/1 脉管炎]

Latest revision as of 21:10, 18 January 2005

I read here that there is no support for clusters and ITK. This is a shame, but there must be a way to work around it. I'm thinking of loading different Regions Of Interest on each node. If whatever algorithm has edge effects, the ROIs might overlap just enough to eliminate them.

Build with mpiCC

I used mpiCC for a build on our cluster and I get a lot of segfaults. I can do examples like ImageReadRegionOfInterestWrite, RGBToGrayscale, and RGBImageSeriesReadWrite but MedianImageFilter and WatershedSegmentation1 segfault. I assume they use more arithmetic. I did "make test" and saw a failure with VXL fly by. How should I investigate?

Big Memory

How could a cluster be used to allow ITK to handle Images larger than the memory on a single machine?

Stream -> MPI

I'm trying to figure out how hard it would be to go from a streamable filter to a distributed memory filter.