這個好玩

看一下最新官方資料

2005.08.19

Dear members,

After some very constructive feedback, it has become obvious that the three day timeout is not acceptable due to slower machines. We absolutely do want to keep these machines crunching work units. Therefore, we will change the timeout. Here is the new configuration of the Rosetta jobs:

1) The maximum number of dispatches per workunit will be limited to 20.

2) The wall clock timeout will be set to 7 days. This means that after 7 days, your workunit will be discarded if it has not completed processing and uploaded to the grid server. Please do not grab more work than you can complete in a week (UD Mon or whatever) so we do not have to redispatch due to lost work.

3) After we have the minimum results (currently 3) for each and every workunit, the job will be considered completed. At that time we will wait 7 days for any outstanding workunits to be returned for credit. After 7 days, the job will be marked completed and no credit will be given for outstanding workunits.

4) We will enable a new job as soon as we have the minimum results for the first job. That way there will always be new work that needs to be processed.

Again we are in a trial and error mode while we attempt to find the optimal configuration that benefits both the work that needs to be done and the user community that is performing that work. If we need to make a change we will make it. Please offer suggestions if you have them and please keep the flames to a minimum. It is not our intent to invalidate work or keep users from getting the points they deserve.

Thank you for your patience and your contribution.
_________________
-----------------
Robby Brewer
Senior Support Engineer
United Devices

======================================================
yahoo翻譯
======================================================

親愛的成員,

在一些非常建設性的反饋以後, 它變得明顯, 3 天的暫停不可接受歸結於更慢的機器。我們absolutely 想要繼續這些機器咬嚼工作單位。所以, 我們將改變暫停。這Rosetta 工作的新配置:

1) 急件的最大數字每workunit 將被限制到20 。

2) 壁鐘暫停將被設置對7 天。這意味著, 在7 天以後, 您的workunit 將被擯除如果它未完成處理和上裝對柵格伺服器。比您能完成在一個星期請不要劫掠更多工作(UD 星期一或什麼) 我們不如此必須redispatch 由於失去的工作。

3) 在我們有之後極小值收效(當前3) 為各workunit, 工作將被認為被完成。那時我們將等7 天所有卓著的workunits 返回為credit 。在7 天以後, 工作將被標記完成了並且不會被相信為卓著的workunits 。

4) 我們將使能一個新工作當我們有極小的結果為第一工作。那個方式那裡總將是需要被處理的新工作。

我們再是在嘗試方式下當我們試圖發現有益於工作需要完成並且廣大用戶進行那工作的優選的配置。如果我們需要做變動我們將做它。請提供建議如果您有他們並且請保留火焰對極小值。這不是我們的意向退役工作或保留用戶從得到他們該當的點。

謝謝您的耐心和您的貢獻。
_________________
-----------------
Robby 釀酒者
高級支持工程師
團結的設備

*******************************************************

嗯...
又有了些應對的更新

但是讓我聯想到....
是否我們上次UD日回傳的資料過於異常

讓官方採取這些"預防措施"...(被抓包)
一切是我個人的猜想...

不過，大家一樣要繼續加油喔