中国分布式计算论坛

 找回密码
 新注册用户
搜索
查看: 7712|回复: 31

[新闻] core17将加速2倍

[复制链接]
发表于 2013-4-1 21:03:09 | 显示全部楼层 |阅读模式
We have been aggressively working on OpenMM (the key code used in the FAH GPU cores), creating new algorithms to increase performance on NVIDIA and AMD GPUs.  The results have been pretty exciting.  With OpenMM 5.1 (vs OpenMM 5.0, used in the current core 17 release), we are getting about a 2x speed up on typical FAH WU calculations, which will lead to an automatic 2x increase in PPD once this software is out of beta testing and integrated into core 17.  
There's a lot of testing to do and it's very possible that these numbers will change, but the results were so exciting that I wanted to give donors a heads up.  Here's some numbers that we're seeing:
OpenCL running on the GTX 680: The first 2 columns are nanoseconds per day (i.e. how much science gets done in a GPU day) and the 3rd column is the speedup of 5.1 over 5.0.
Type of Calculation
OpenMM 5.0
OpenMM 5.1
Speedup
Implicit hbonds
92
134
1.46
Implicit hangles
153
209
1.36
RF hbonds
31.4
78.1
2.49
RF hangles
58
113.0
1.95
PME hbonds
19.6
41.5
2.12
PME hangles
37.3
66.9
1.79

OpenCL running on a Radeon HD 7970:  The first 2 columns are nanoseconds per day (i.e. how much science gets done in a GPU day) and the 3rd column is the speedup of 5.1 over 5.0.
Type of Calculation
OpenMM 5.0
OpenMM 5.1
Speedup
Implicit hbonds
87
120
1.38
Implicit hangles
96
104
1.09
RF hbonds
33.5
83.5
2.49
RF hangles
51.8
90.2
1.74
PME hbonds
21.8
49.3
2.26
PME hangles
34.6
63.0
1.82

Note that "PME hbonds" is likely the most common calculation that we plan to run in the near term with core 17.   We're very excited about the way this is shaping up and think that donors would be curious to know where this is going.  


发表于 2013-4-2 07:42:31 | 显示全部楼层
最近斯坦福总算有几个拿得出手的优化了期待众多GPU党的核爆

PS:貌似几个优化都和香港高人有关,一个黄教授团队?还有一个年轻高手赵宇通?
发表于 2013-4-2 11:05:13 | 显示全部楼层
6系列N卡的福音???不过最多也就2倍 GPU核爆还差的远啊!!!
发表于 2013-4-2 11:50:53 | 显示全部楼层
请问各位前辈,怎样才能接到core17的包啊?
发表于 2013-4-2 19:32:50 | 显示全部楼层
tycolj 发表于 2013-4-2 11:50
请问各位前辈,怎样才能接到core17的包啊?

打开高级控制界面,点击configure>slots > X gpu > edit >add>上下各分别填入client-type和beta >ok>ok>save
然后退出客户端重开客户端

捕获.PNG
发表于 2013-4-2 22:22:23 | 显示全部楼层
那岂不是单卡能有100kPPD

点评

7970目前可以51K  发表于 2013-4-3 14:35
发表于 2013-4-3 14:26:18 来自手机 | 显示全部楼层
core17,A卡是不是只有7900系列能接到?

点评

都能接到,但是GCN架构很给力  发表于 2013-4-3 14:37
发表于 2013-4-3 18:10:37 | 显示全部楼层
貌似680性能挺好的嘛,怎么现在的表现这么挫。。
发表于 2013-4-3 23:15:29 | 显示全部楼层
按照金版的方法设置了参数,为清明赛热身

未命名.jpg

默频7770,好像比之前提高了好几倍

点评

求运行半小时后稳定PPD  发表于 2013-4-5 10:49
 楼主| 发表于 2013-4-4 00:18:30 | 显示全部楼层
dianci 发表于 2013-4-3 23:15
按照金版的方法设置了参数,为清明赛热身

刚启动时,软件估算ppd不准,你确定在稳定运行一段时间后,ppd也是3万多?
发表于 2013-4-5 10:41:27 | 显示全部楼层
- -早知道有这货学校电脑就改一下了..... 后悔回家前没看论坛
发表于 2013-4-5 11:22:00 | 显示全部楼层

——

本帖最后由 dianci 于 2013-4-5 14:17 编辑

CPU跑boinc,GPU跑FAH,目前稳定在5K——7K

另外,加了参数就是默认使用core17了吗?还是要看具体接到什么包?
 楼主| 发表于 2013-4-5 11:24:16 | 显示全部楼层
dianci 发表于 2013-4-5 11:22
CPU跑boinc,GPU跑FAH,目前稳定在5K

另外,加了参数就是默认使用core17了吗?还是要看具体接到什么包? ...

加参数表示具备接包能力,发不发包看发包服务器的分配
发表于 2013-4-6 18:02:16 | 显示全部楼层
金鹏 发表于 2013-4-2 19:32
打开高级控制界面,点击configure>slots > X gpu > edit >add>上下各分别填入client-type和beta >ok>ok>s ...

接了一个core17的包,ppd从4280猛降到1180,我的是GT620马甲卡,现在我重新设置了一下,还是算core15的包。
 楼主| 发表于 2013-4-6 18:09:30 | 显示全部楼层
tycolj 发表于 2013-4-6 18:02
接了一个core17的包,ppd从4280猛降到1180,我的是GT620马甲卡,现在我重新设置了一下,还是算core15的包 ...

N卡偏低,我这也是
您需要登录后才可以回帖 登录 | 新注册用户

本版积分规则

论坛官方淘宝店开业啦~

小黑屋|手机版|Archiver|中国分布式计算总站 ( 沪ICP备05042587号 )

GMT+8, 2019-11-22 05:57

Powered by Discuz! X3.4

© 2001-2017 Comsenz Inc.

快速回复 返回顶部 返回列表