Hi community,

a teammate would like to let run 2 WUs/GTX460.

How need to look the app_info.xml file for to let run both CPU apps and the GPU app of MW@h?

Thanks.

---

I believe that it is me he is referring to also while useing 2 gpu I lost 1 of my 8 cores to feed the gpu. I would like to put that core to crunching rather than to feed the gpu also is there an opti app similar to lunatics that is used at seti. My gpu load is 98-99% with 1 wu per gpu and I am eating up the nvida wu at 7 min per wu. Also I am almost out of work I have 77 nvida wu left used up all of the 100 and 0.84 wu.

I just got some 11 wu 0.84 and 6 1.00 wu I still have 76 1.02 wu.
My other 2 projects are out of work seti and enigma. So I have been 24/7 for the last 3 days.

http://www.arkayn.us/forum/index.php?action=downloads;sa=view;down=59

Set the count to 0.5.

add this to the app_info

  <app>
    <name>milkyway</name>
  </app>
  <file_info>
    <name>milkyway_separation_1.00_windows_x86_64.exe</name>
    <executable />
  </file_info>
  <app_version>
    <app_name>milkyway</app_name>
    <version_num>100</version_num>
	<cmdline></cmdline>
    <file_ref>
      <file_name>milkyway_separation_1.00_windows_x86_64.exe</file_name>
      <main_program/>
    </file_ref>
  </app_version>
<app>
<name>milkyway_nbody</name>
<user_friendly_name>MilkyWay@Home nbody Simulation</user_friendly_name>
</app>
<file_info>
<name>milkyway_nbody_0.84_windows_x86_64__mt.exe</name>
<executable/>
</file_info>
<file_info>
<name>libgomp_64-1.dll</name>
<executable/>
</file_info>
<file_info>
<name>pthreadGC2_64.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>milkyway_nbody</app_name>
<version_num>84</version_num>
<plan_class>mt</plan_class>
<avg_ncpus>4</avg_ncpus>
<max_ncpus>4</max_ncpus>
<cmdline>--nthreads=4</cmdline>
<file_ref>
<file_name>milkyway_nbody_0.84_windows_x86_64__mt.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libgomp_64-1.dll</file_name>
</file_ref>
<file_ref>
<file_name>pthreadGC2_64.dll</file_name>
</file_ref>
</app_version>

http://milkyway.cs.rpi.edu/milkyway/download/libgomp_64-1.dll
http://milkyway.cs.rpi.edu/milkyway/download/pthreadGC2_64.dll
http://milkyway.cs.rpi.edu/milkyway/download/milkyway_nbody_0.84_windows_x86_64__mt.exe
http://milkyway.cs.rpi.edu/milkyway/download/milkyway_separation_1.00_windows_x86_64.exe

---

I downloaded the files

libgomp_64-1.dll
pthreadGC2_64.dll
milkyway_nbody_0.84_windows_x86_64__mt.exe
milkyway_separation_1.00_windows_x86_64.exe

and created your app_info file

I have question of will this work with my i7
<avg_ncpus>4</avg_ncpus>
<max_ncpus>4</max_ncpus>
<cmdline>--nthreads=4</cmdline>

I haven't installed anything yet.

app_info.xml

<app>
    <name>milkyway</name>
</app>
<file_info>
    <name>milkyway_separation_1.00_windows_x86_64.exe</name>
    <executable/>
</file_info>
<app_version>
    <app_name>milkyway</app_name>
    <version_num>100</version_num>
    <cmdline></cmdline>
    <file_ref>
        <file_name>milkyway_separation_1.00_windows_x86_64.exe</file_name>
        <main_program/>
    </file_ref>
</app_version>
<app>
    <name>milkyway_nbody</name>
    <user_friendly_name>MilkyWay@Home nbody Simulation</user_friendly_name>
</app>
<file_info>
    <name>milkyway_nbody_0.84_windows_x86_64__mt.exe</name>
    <executable/>
</file_info>
<file_info>
    <name>libgomp_64-1.dll</name>
    <executable/>
</file_info>
<file_info>
    <name>pthreadGC2_64.dll</name>
    <executable/>
</file_info>
<app_version>
    <app_name>milkyway_nbody</app_name>
    <version_num>84</version_num>
    <plan_class>mt</plan_class>
    <avg_ncpus>4</avg_ncpus>
    <max_ncpus>4</max_ncpus>
    <cmdline>--nthreads=4</cmdline>
    <file_ref>
        <file_name>milkyway_nbody_0.84_windows_x86_64__mt.exe</file_name>
        <main_program/>
    </file_ref>
    <file_ref>
        <file_name>libgomp_64-1.dll</file_name>
    </file_ref>
    <file_ref>
        <file_name>pthreadGC2_64.dll</file_name>
    </file_ref>
</app_version>

Don't forget the one at the top of the post.

Or the complete app_info would be

<app_info>
  <app>
    <name>milkyway</name>
  </app>
  <file_info>
    <name>milkyway_separation_1.02_windows_x86_64__opencl_nvidia.exe</name>
    <executable />
  </file_info>
  <app_version>
    <app_name>milkyway</app_name>
    <version_num>102</version_num>
    <flops>1.0e11</flops>
    <avg_ncpus>0.05</avg_ncpus>
    <max_ncpus>1</max_ncpus>
    <plan_class>cuda</plan_class>
    <coproc>
      <type>CUDA</type>
      <count>0.5</count>
    </coproc>
	<cmdline></cmdline>
    <file_ref>
      <file_name>milkyway_separation_1.02_windows_x86_64__opencl_nvidia.exe</file_name>
      <main_program/>
    </file_ref>
  </app_version>
<app>
<name>milkyway</name>
</app>
<file_info>
<name>milkyway_separation_1.00_windows_x86_64.exe</name>
<executable />
</file_info>
<app_version>
<app_name>milkyway</app_name>
<version_num>100</version_num>
<cmdline></cmdline>
<file_ref>
<file_name>milkyway_separation_1.00_windows_x86_64.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>milkyway_nbody</name>
<user_friendly_name>MilkyWay@Home nbody Simulation</user_friendly_name>
</app>
<file_info>
<name>milkyway_nbody_0.84_windows_x86_64__mt.exe</name>
<executable/>
</file_info>
<file_info>
<name>libgomp_64-1.dll</name>
<executable/>
</file_info>
<file_info>
<name>pthreadGC2_64.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>milkyway_nbody</app_name>
<version_num>84</version_num>
<plan_class>mt</plan_class>
<avg_ncpus>4</avg_ncpus>
<max_ncpus>4</max_ncpus>
<cmdline>--nthreads=4</cmdline>
<file_ref>
<file_name>milkyway_nbody_0.84_windows_x86_64__mt.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libgomp_64-1.dll</file_name>
</file_ref>
<file_ref>
<file_name>pthreadGC2_64.dll</file_name>
</file_ref>
</app_version>
</app_info>

And yes it will work just fine on the i7.

---

Thanks it is working just fine now doing 8 wu per cpu and 2 wu per gpu.
gpu work units are now labeled cuda
all of my opencl nvida 11 of them got errored out though
Well done I am happy cruncher

I noticed on your sit crunchers anonymous the downloads section and noticed the different app so which one I am using?

The apps are the exact same thing, I just include the app_info so you can tune it to your system.

There is also the AMD app there.

ATI/AMD still kicks Nvidias behind in double precision processing.

---

I believe that it is me he is referring to also while useing 2 gpu I lost 1 of my 8 cores to feed the gpu. I would like to put that core to crunching rather than to feed the gpu also is there an opti app similar to lunatics that is used at seti. My gpu load is 98-99% with 1 wu per gpu and I am eating up the nvida wu at 7 min per wu. Also I am almost out of work I have 77 nvida wu left used up all of the 100 and 0.84 wu.

Your problem is also here: "Memory 3062.12 MB"! You have an Intel i7 cpu running 8 cpu cores and 2 gpu's all in just 4gb of ram! AND 4 of the cpu cores are not physical!!

I believe that it is me he is referring to also while useing 2 gpu I lost 1 of my 8 cores to feed the gpu. I would like to put that core to crunching rather than to feed the gpu also is there an opti app similar to lunatics that is used at seti. My gpu load is 98-99% with 1 wu per gpu and I am eating up the nvida wu at 7 min per wu. Also I am almost out of work I have 77 nvida wu left used up all of the 100 and 0.84 wu.

Your problem is also here: "Memory 3062.12 MB"! You have an Intel i7 cpu running 8 cpu cores and 2 gpu's all in just 4gb of ram! AND 4 of the cpu cores are not physical!!

I often wondered about my memory and was told I have enough I am using 86% what would you suggest I upgrade to?

I am currently at 12GB in my AMD machine, my i3 machine has 8GB.

I think 8GB is probably the new sweet spot.

---

I am currently at 12GB in my AMD machine, my i3 machine has 8GB.

I think 8GB is probably the new sweet spot.

It sort of depends on which Boinc Projects you want to crunch for, if you are going to crunch for Rosetta I would suggest 12 or more, otherwise 8 is enough for now. Check the prices, if you can get 12, or even 16, for only a small amount more than 8 go for it, otherwise 8 should be plenty. I have two AMD X6 machines and they both have 16gb of ram, they crunch with all 6 cores and a gpu and don't even breathe hard! The 16gb for me was only $60US for four 4gb sticks, so it was a no brainer!

If you are using 85% that is a STRONG indication of a memory bound system! Ram is cheap but cached memory is fake and since it uses the hard drive slows down your system. In Win7 Windows itself is an okay memory manager using free memory as cache memory, but that is different than when you are running out of memory and have to use the hard drive to supplement it.

I have 12GB using an i7 2600K (8 vitual cores) and I ran out of memory because I had set my swap file to 16 MB. Also I run Cosmology on 7 CPU cores (leaving 1 core free for supplying data to my 2 (soon 3) GPU's. But Cosmology uses 700-800 MB of memory so that adds up. And I do (not a lot) more than crunching on this machine. Browse the net and using it with my TV dual-tuner card to record/watch TV an play music/movies/DVD's. Although the music stuff I prefer doing on my older dual core pc as it stutters sometimes on my i7 2600K. I'm using Winamp for listening to music and internet radio and it's priority is even on realtime (not adviced to use high or realtime on other programs). But the stutter/crackling sound remains.

Anthony.

---

The longer I live, the more reasons I develop for wanting to die.

I am not doing any crunching at all now and it is at 73% so when it was crunching 3 projects seti, enigma and milkyway it used 83% or less so I think I am ok.

I have decided to upgrade my memory to

Corsair 24 GB Vengeance PC3-12800 1600mhz 240-pin Triple Channel DDR3 Memory Kit CMZ24GX3M6A1600C9

My old memory is

G.SKILL
F3-1066 CLT7T
3 6BPK
DDR3-1333
CL7-7-7-18
1.5V-1.6V

I am running it in XMP mode
MHz 666
CL 7
tRCD 7
tRP 7
tRAS 18
CR 2T

I am using 83% memory
I have 5.98gb virtual memory
and have 1.82 gb of virtual memory left
so this should improve

I have decided to upgrade my memory to

Corsair 24 GB Vengeance PC3-12800 1600mhz 240-pin Triple Channel DDR3 Memory Kit CMZ24GX3M6A1600C9

My old memory is

G.SKILL
F3-1066 CLT7T
3 6BPK
DDR3-1333
CL7-7-7-18
1.5V-1.6V

I am running it in XMP mode
MHz 666
CL 7
tRCD 7
tRP 7
tRAS 18
CR 2T

I am using 83% memory
I have 5.98gb virtual memory
and have 1.82 gb of virtual memory left
so this should improve

Those % numbers should improve dramatically now!!!

How do I figure my flops for milkyway is it similar to the way it is done at seti such as Average processing rate 17.590351299988 so flops is rounded off to 17e09 this is for my seti host so how do I do it for milkyway?

How do I figure my flops for milkyway is it similar to the way it is done at seti such as Average processing rate 17.590351299988 so flops is rounded off to 17e09 this is for my seti host so how do I do it for milkyway?

That I have no clue about.

You should not have to worry about it here as the DCF does not swing wildly here. The only reason we set it at SETI is because of the the wild DCF swings.

---