CAL Wrapper (Ready for Testing)

Moderators: slegrand, Site Moderators, PandeGroup

CAL Wrapper (Ready for Testing)

Postby Shelnutt2 » Fri Jun 12, 2009 2:40 am

I created a new thread, for the simple fact to keep my code separate from his. It's easier to make sure that people are using mine, and not his if they are reporting a bug.

I don't have a ATI card capable of folding, and no one at ocforums has tested this yet. I figured I'd go ahead and post here though.

Alright, here we go:
http://shelnutt.twomurs.com/patches/cal/

Binaries:
http://shelnutt.twomurs.com/patches/cal/amdcalcl/amdcalcl.dll.so
http://shelnutt.twomurs.com/patches/cal/amdcalrt/amdcalrt.dll.so

Place those binaries in your /usr/lib/wine directory and then run the gpu2 client via wine. Make sure you have at least 9.3 or high Catalyst drivers installed, as I built this around CAL 1.4 beta. My code included some left out functions and it was built with a better linking makefile. Also there are some optional functions that I had trouble compiling, so they are commented out, but does anyone know if Stanford makes use of the calCtxCreateCounter, calCtxDestroyCounter, calCtxEndCounter, calCtxBeginCounter, calclImageWrite, calResCreate2D or calResCreate1D? Those are the seven functions not implemented.
Last edited by Shelnutt2 on Fri Jun 12, 2009 12:09 pm, edited 1 time in total.
OCforums, T32

Linux Lober!
Shelnutt2
 
Posts: 150
Joined: Sun Dec 02, 2007 5:37 pm

Re: CAL Wrapper (Ready for Testing)

Postby mhouston » Fri Jun 12, 2009 4:19 am

calCtxCreate, calCtxDestroy - yes
calCtxBegin/end - not by default
calResCreate1D/2D - yes

You can look at the Brook sources in the SDK to see which CAL functions are used. In general, ideally the full function map would be handled underneath. Sending you a PM in a sec with some code.
mhouston
 
Posts: 892
Joined: Sun Dec 02, 2007 8:19 pm

Re: CAL Wrapper (Ready for Testing)

Postby shatteredsilicon » Fri Jun 12, 2009 7:09 am

Is this a new wrapper from scratch, or an update to the old one? Any idea why the original version didn't work under WINE?
Image
1x Q6600 @ 3.2GHz, 4GB DDR3-1333
1x Phenom X4 9950 @ 2.6GHz, 4GB DDR2-1066
3x GeForce 9800GX2
1x GeForce 8800GT
CentOS 5 x86-64, WINE 1.x with CUDA wrappers
shatteredsilicon
 
Posts: 627
Joined: Tue Jul 08, 2008 2:27 pm

Re: CAL Wrapper (Ready for Testing)

Postby Anachron » Fri Jun 12, 2009 9:53 am

I downloaded the GPU client, and put the binaries in /usr/lib32/wine (I'm on 64 bit).
As you can see, it's not able to get a WU for some reason:
Code: Select all
anachron@supercomputer ~/.folding/folding-gpu $ wine fah6-gpu.exe -forcegpu r600 -oneunit -verbosity 9

Note: Please read the license agreement (fah6-gpu.exe -license). Further
use of this software requires that you have read and accepted this agreement.



--- Opening Log file [June 12 09:46:18 UTC]


# Windows GPU Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.23

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: Z:\home\anachron\.folding\folding-gpu
Executable: fah6-gpu.exe
Arguments: -forcegpu r600 -oneunit -verbosity 9

[09:46:18] - Ask before connecting: No
[09:46:18] - User name: Anonymous (Team 0)
[09:46:18] - User ID: 2EC0534E5D7AA47F
[09:46:18] - Machine ID: 2
[09:46:18]
[09:46:18] Loaded queue successfully.
[09:46:18] - Preparing to get new work unit...
[09:46:18] + Attempting to get work packet
[09:46:18] - Autosending finished units... [June 12 09:46:18 UTC]
[09:46:18] - Will indicate memory of 3900 MB
[09:46:18] Trying to send all finished work units
[09:46:18] - Detect CPU.[09:46:18] + No unsent completed units remaining.
 Vendor: AuthenticAMD, Family: 15, Model: 2, Stepping: 3
[09:46:18] - Autosend completed
[09:46:18] - Connecting to assignment server
[09:46:18] Connecting to http://assign-GPU.stanford.edu:8080/
[09:46:18] Posted data.
[09:46:18] Initial: 0000; + No appropriate work server was available; will try again in a bit.
[09:46:18] + Couldn't get work instructions.
[09:46:18] - Attempt #1  to get work failed, and no other work to do.
Waiting before retry.
[09:46:36] + Attempting to get work packet
[09:46:36] - Will indicate memory of 3900 MB
[09:46:36] - Connecting to assignment server
[09:46:36] Connecting to http://assign-GPU.stanford.edu:8080/
[09:46:37] Posted data.
[09:46:37] Initial: 0000; + No appropriate work server was available; will try again in a bit.
[09:46:37] + Couldn't get work instructions.
[09:46:37] - Attempt #2  to get work failed, and no other work to do.
Waiting before retry.
(...)


So, I restarted and logged into my XP install. I ran a gpu client there for a while with no problems.
I copied it back to Linux Mint, and ran it. This is what I get from it:
Code: Select all
anachron@supercomputer ~/.folding/folding-gpu2 $ wine fah6-gpu.exe -forcegpu r600 -oneunit -verbosity 9

Note: Please read the license agreement (fah6-gpu.exe -license). Further
use of this software requires that you have read and accepted this agreement.



--- Opening Log file [June 12 09:39:34 UTC]


# Windows GPU Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.23

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: Z:\home\anachron\.folding\folding-gpu2
Executable: fah6-gpu.exe
Arguments: -forcegpu r600 -oneunit -verbosity 9

[09:39:34] - Ask before connecting: No
[09:39:34] - User name: Anonymous (Team 0)
[09:39:34] - User ID: 2EC0534E5D7AA47F
[09:39:34] - Machine ID: 2
[09:39:34]
[09:39:34] Loaded queue successfully.
[09:39:34]
[09:39:34] + Processing work unit
[09:39:34] Core required: FahCore_11.exe
[09:39:34] - Autosending finished units... [June 12 09:39:34 UTC]
[09:39:34] Trying to send all finished work units
[09:39:34] Core found.
[09:39:34] + No unsent completed units remaining.
[09:39:34] - Autosend completed
[09:39:34] Working on queue slot 01 [June 12 09:39:34 UTC]
[09:39:34] + Working ...
[09:39:34] - Calling '.\FahCore_11.exe -dir work/ -suffix 01 -checkpoint 15 -verbose -lifeline 8 -version 623'

[09:39:34]
[09:39:34] *------------------------------*
[09:39:34] Folding@Home GPU Core - Beta
[09:39:34] Version 1.24 (Mon Feb 9 11:00:12 PST 2009)
[09:39:34]
[09:39:34] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[09:39:34] Build host: amoeba
[09:39:34] Board Type: AMD
[09:39:34] Core      :
[09:39:34] Preparing to commence simulation
[09:39:34] - Looking at optimizations...
[09:39:34] - Files status OK
[09:39:34] - Expanded 96467 -> 489152 (decompressed 507.0 percent)
[09:39:34] Called DecompressByteArray: compressed_data_size=96467 data_size=489152, decompressed_data_size=489152 diff=0
[09:39:34] - Digital signature verified
[09:39:34]
[09:39:34] Project: 5737 (Run 1, Clone 166, Gen 12)
[09:39:34]
[09:39:34] Assembly optimizations on if available.
[09:39:34] Entering M.D.
Reading file work/wudata_01.tpr, VERSION 3.1.4 (single precision)
Reading file work/wudata_01.tpr, VERSION 3.1.4 (single precision)
Reading sasa-enabled ir 0 0
Issues between syncs set at 16
In override calInit
calLoadBinary
[09:39:40] Will resume from checkpoint file
[09:39:40] Tpr hash work/wudata_01.tpr:  117149282 1306075925 1255274948 2794642518 1351495203
Failed to init CAL
Could not close cal device.
Could not shutdown cal.
[09:39:44] CoreStatus = FFFFFFF6 (-10)
[09:39:44] Client-core communications error: ERROR 0xfffffff6
[09:39:44] This is a sign of more serious problems, shutting down.


Hoping this helps.

EDIT I: Forgot to add: I have Wine 1.0.1, and fglrx 9.5 (8.612).
EDIT II: I get the same results with Wine 1.1.23.
Time flies like an arrow; fruit flies like a banana
Anachron
 
Posts: 50
Joined: Fri Mar 14, 2008 12:10 pm

Re: CAL Wrapper (Ready for Testing)

Postby Shelnutt2 » Fri Jun 12, 2009 12:21 pm

mhouston wrote:calCtxCreate, calCtxDestroy - yes
calCtxBegin/end - not by default
calResCreate1D/2D - yes

You can look at the Brook sources in the SDK to see which CAL functions are used. In general, ideally the full function map would be handled underneath. Sending you a PM in a sec with some code.


I had a typo, It's the counter functions I don't have implemented because when I tried compiling with them,
Code: Select all
amdcalrt.o: In function `wine_calclImageWrite':
amdcalrt.c:(.text+0x51d): undefined reference to `calclImageWrite'
amdcalrt.o: In function `wine_calCtxCreateCounter':
amdcalrt.c:(.text+0x721): undefined reference to `calCtxCreateCounter'
amdcalrt.o: In function `wine_calCtxDestroyCounter':
amdcalrt.c:(.text+0x74d): undefined reference to `calCtxDestroyCounter'
amdcalrt.o: In function `wine_calCtxBeginCounter':
amdcalrt.c:(.text+0x779): undefined reference to `calCtxBeginCounter'
amdcalrt.o: In function `wine_calCtxEndCounter':
amdcalrt.c:(.text+0x7a5): undefined reference to `calCtxEndCounter'
amdcalrt.o: In function `wine_calCtxGetCounter':
amdcalrt.c:(.text+0x7d8): undefined reference to `calCtxGetCounter'
amdcalrt.o: In function `wine_calResCreate2D':
amdcalrt.c:(.text+0x82e): undefined reference to `calResCreate2D'
amdcalrt.o: In function `wine_calResCreate1D':
amdcalrt.c:(.text+0x876): undefined reference to `calReCreate1D'
collect2: ld returned 1 exit status
winegcc: x86_64-pc-linux-gnu-gcc failed
make: *** [amdcalrt.dll.so] Error 2





shatteredsilicon wrote:Is this a new wrapper from scratch, or an update to the old one? Any idea why the original version didn't work under WINE?


This is a new wrapper, from the ground up. I opened the old wrapper, but it wasn't in order with the newer sdk API, the functions were not in order, so I just rewrote it completely, and kept the functions in the order of the newer API so that it's easier to debug.


Anachron wrote:I downloaded the GPU client, and put the binaries in /usr/lib32/wine (I'm on 64 bit).
As you can see, it's not able to get a WU for some reason:<snip>

So, I restarted and logged into my XP install. I ran a gpu client there for a while with no problems.
I copied it back to Linux Mint, and ran it. This is what I get from it:
Code: Select all
anachron@supercomputer ~/.folding/folding-gpu2 $ wine fah6-gpu.exe -forcegpu r600 -oneunit -verbosity 9

Note: Please read the license agreement (fah6-gpu.exe -license). Further
use of this software requires that you have read and accepted this agreement.



--- Opening Log file [June 12 09:39:34 UTC]


# Windows GPU Console Edition #################################################
###############################################################################

                       Folding@Home Client Version 6.23

                          http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: Z:\home\anachron\.folding\folding-gpu2
Executable: fah6-gpu.exe
Arguments: -forcegpu r600 -oneunit -verbosity 9

[09:39:34] - Ask before connecting: No
[09:39:34] - User name: Anonymous (Team 0)
[09:39:34] - User ID: 2EC0534E5D7AA47F
[09:39:34] - Machine ID: 2
[09:39:34]
[09:39:34] Loaded queue successfully.
[09:39:34]
[09:39:34] + Processing work unit
[09:39:34] Core required: FahCore_11.exe
[09:39:34] - Autosending finished units... [June 12 09:39:34 UTC]
[09:39:34] Trying to send all finished work units
[09:39:34] Core found.
[09:39:34] + No unsent completed units remaining.
[09:39:34] - Autosend completed
[09:39:34] Working on queue slot 01 [June 12 09:39:34 UTC]
[09:39:34] + Working ...
[09:39:34] - Calling '.\FahCore_11.exe -dir work/ -suffix 01 -checkpoint 15 -verbose -lifeline 8 -version 623'

[09:39:34]
[09:39:34] *------------------------------*
[09:39:34] Folding@Home GPU Core - Beta
[09:39:34] Version 1.24 (Mon Feb 9 11:00:12 PST 2009)
[09:39:34]
[09:39:34] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[09:39:34] Build host: amoeba
[09:39:34] Board Type: AMD
[09:39:34] Core      :
[09:39:34] Preparing to commence simulation
[09:39:34] - Looking at optimizations...
[09:39:34] - Files status OK
[09:39:34] - Expanded 96467 -> 489152 (decompressed 507.0 percent)
[09:39:34] Called DecompressByteArray: compressed_data_size=96467 data_size=489152, decompressed_data_size=489152 diff=0
[09:39:34] - Digital signature verified
[09:39:34]
[09:39:34] Project: 5737 (Run 1, Clone 166, Gen 12)
[09:39:34]
[09:39:34] Assembly optimizations on if available.
[09:39:34] Entering M.D.
Reading file work/wudata_01.tpr, VERSION 3.1.4 (single precision)
Reading file work/wudata_01.tpr, VERSION 3.1.4 (single precision)
Reading sasa-enabled ir 0 0
Issues between syncs set at 16
In override calInit
calLoadBinary
[09:39:40] Will resume from checkpoint file
[09:39:40] Tpr hash work/wudata_01.tpr:  117149282 1306075925 1255274948 2794642518 1351495203
Failed to init CAL
Could not close cal device.
Could not shutdown cal.
[09:39:44] CoreStatus = FFFFFFF6 (-10)
[09:39:44] Client-core communications error: ERROR 0xfffffff6
[09:39:44] This is a sign of more serious problems, shutting down.


Hoping this helps.

EDIT I: Forgot to add: I have Wine 1.0.1, and fglrx 9.5 (8.612).
EDIT II: I get the same results with Wine 1.1.23.


Well the key is, "Failed to init CAL
Could not close cal device.
Could not shutdown cal."

It seems CAL did not load. I'm thinking the problem is, that the library in linux (and maybe windows, haven't checked) for the latest CAL (1.4) has been renamed to libaticalrt, from libamdcalrt. So in wine we have amdcalrt.dll, and wine is expecting libamdcalrt to be there, but it's not, it's libaticalrt. mhouston is sending me some code that should take care of this. For now though you could try renaming the amdcalrt.dll file to aticalrt.dll and the same with the wrapper. Rename it to aticalrt.dll.so and try again.
Shelnutt2
 
Posts: 150
Joined: Sun Dec 02, 2007 5:37 pm

Re: CAL Wrapper (Ready for Testing)

Postby bollix47 » Fri Jun 12, 2009 1:29 pm

Tried renaming amd* to ati* in the fah folder and the /usr/lib/wine folder but get basically the same as above:

Code: Select all
[13:26:59]
[13:26:59] Entering M.D.
[13:27:05] Tpr hash work/wudata_08.tpr:  1999441989 786770505 3174414183 3845918147 3506284447
Reading file work/wudata_08.tpr, VERSION 3.1.4 (single precision)
Reading file work/wudata_08.tpr, VERSION 3.1.4 (single precision)
Reading sasa-enabled ir 0 0
Issues between syncs set at 16
In override calInit
calLoadBinary
Failed to init CAL
Could not close cal device.
Could not shutdown cal.
[13:27:07] CoreStatus = FFFFFFF6 (-10)
[13:27:07] Client-core communications error: ERROR 0xfffffff6
[13:27:07] This is a sign of more serious problems, shutting down.


Running under wine from Ubuntu 8.04 repos.
Image
bollix47
 
Posts: 3463
Joined: Sun Dec 02, 2007 5:04 am
Location: Canada

Re: CAL Wrapper (Ready for Testing)

Postby shatteredsilicon » Fri Jun 12, 2009 6:42 pm

Yeah, that is the same problem as before. There seems to be something importantly different between the Windows and Linux CAL initialization that makes wrapping requirements a bit more than paper thin.
shatteredsilicon
 
Posts: 627
Joined: Tue Jul 08, 2008 2:27 pm

Re: CAL Wrapper (Ready for Testing)

Postby Anachron » Mon Jun 15, 2009 6:20 pm

I installed the latest CAL, and as it does not officially support Ubuntu I followed this guide: http://developer.amd.com/support/Knowle ... aspx?ID=28. (not sure if this is necessary for just running the client).

I tried the client again, with the same result as before. Running wine with WINEDEBUG=warn+all, however, shows this:
Code: Select all
warn:debugstr:OutputDebugStringA calLoadLibrary: loaded "amdcalcl.dll" ...
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calDeviceGetInfo missing in amdcalrt*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calImageFree missing in amdcalrt*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calGetErrorString missing in amdcalrt*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calModuleGetFuncInfo missing in amdcalrt*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclGetVersion missing in amdcalcl*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclFreeObject missing in amdcalcl*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclFreeImage missing in amdcalcl*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclDisassembleImage missing in amdcalcl*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclAssembleObject missing in amdcalcl*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclDisassembleObject missing in amdcalcl*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclImageGetSize missing in amdcalcl*.dll
warn:debugstr:OutputDebugStringA calLoadLibrary: ** ERROR: calclImageWrite missing in amdcalcl*.dll
Failed to init CAL
Could not release constant memory.
Could not free constant memory.
Could not close cal device.
Could not shutdown cal.


Seems something is amiss...

I added these functions to the source file, compiled, and replaced your .so's with my edited ones.

Now I get this (I have removed some irrelevant warns (I get them on freecell.exe for instance :) ):
Code: Select all
Issues between syncs set at 16
In override calInit
calLoadBinary
warn:debugstr:OutputDebugStringA calLoadLibrary: loaded "amdcalcl.dll" ...
CAL version: 1.4.283
CAL: 1 devices found
ASIC = RV770
Available board memory to CAL: 512MB
Available cached memory to CAL: 60MB
Available uncached memory to CAL: 1886MB
Engine clock: 0 (0.000000)
Natural vector width: 64
Recommended minimum thread count: 2560
Brook yield: 0
Don't flush kernels: 0
Pre flush command buffers: 0
Flush interval: 16

starting mdrun 'Protein'
10000000 steps,  20000.0 ps.

Calculated RepFac: 8 minSuggestedThreads=2560
Setting RepFac: 8
nbp->PForceStrWidth: 64, gpu->AtomStrSize: 1408 nbp->RepFac: 8 gpu->iUnroll: 4, nbp->PForceStrSize: 2816, nbp->PForceStrHeight: 44
[18:02:16] Working on Protein
[18:02:16] Client config found, loading data.
[18:02:16] Starting GUI Server
Run: exception thrown during GuardedRun
[18:02:16] Run: exception thrown during GuardedRun
[18:02:16] Run: exception thrown in GuardedRun -- Gromacs cannot continue further.
[18:02:16] Going to send back what have done -- stepsTotalG=10000000
[18:02:16] Work fraction=0.0000 steps=10000000.
[18:02:20] logfile size=16573 infoLength=16573 edr=0 trr=23
[18:02:20] - Writing 17109 bytes of core data to disk...
[18:02:20] Done: 16597 -> 5483 (compressed to 33.0 percent)
[18:02:20]   ... Done.
[18:02:20]
[18:02:20] Folding@home Core Shutdown: EARLY_UNIT_END
[18:02:24] CoreStatus = 72 (114)
[18:02:24] Sending work to server
[18:02:24] Project: 5737 (Run 1, Clone 166, Gen 12)


[18:02:24] + Attempting to send results [June 15 18:02:24 UTC]
[18:02:24] - Reading file work/wuresults_01.dat from core
[18:02:24]   (Read 5995 bytes from disk)
[18:02:24] - Presenting message box asking to network.



I obviously did something right!
This is about as far as we got with the previous try on a wrapper, I think.

What next?
Anachron
 
Posts: 50
Joined: Fri Mar 14, 2008 12:10 pm

Re: CAL Wrapper (Ready for Testing)

Postby toTOW » Tue Jun 16, 2009 12:56 pm

That might be good to know what wen wrong in the WU ... maybe you could look in the WU specific log file ... or if you can't get any information, it might be interesting to contact the project owner to ask him to look in the returned results to try to figure out what failed.
Folding@Home beta tester since 2002. Folding Forum moderator since July 2008.

FAH-Addict : latest news, tests and reviews about Folding@Home project.

Image
User avatar
toTOW
Site Moderator
 
Posts: 8624
Joined: Sun Dec 02, 2007 10:38 am
Location: Bordeaux, France

Re: CAL Wrapper (Ready for Testing)

Postby Shelnutt2 » Tue Jun 16, 2009 1:50 pm

That is a very good idea. If we know where to WU is failing maybe we can see what is tripping it up. Also Anachron, can you post up the source code? (So I can update what I have hosted, and to comply with lgpl license)
Shelnutt2
 
Posts: 150
Joined: Sun Dec 02, 2007 5:37 pm

Re: CAL Wrapper (Ready for Testing)

Postby Anachron » Tue Jun 16, 2009 4:50 pm

Yeah, I shot vvoelz a mail (He is listed as contact for the WU I uploaded).

Here's the source. I was not sure what order the functions should be in, so I just added them at the top or bottom.

amdcalrt.c:
Code: Select all
/*Code by Seth Shelnutt, released under the LGPL v2.1 or later in order to be compatiable with Wine
*/


#include <windef.h>

#include "cal.h"
#include "cal_ext.h"
#include "cal_ext_counter.h"
#include "calcl.h"


CALresult WINAPI wine_calModuleGetFuncInfo(CALfuncInfo* pInfo, CALcontext ctx, CALmodule module, CALfunc func) {
   return calModuleGetFuncInfo(pInfo, ctx, module, func);
}

const CALchar* WINAPI wine_calGetErrorString() {
   return calGetErrorString();
}


CALresult WINAPI wine_calImageFree(CALimage image) {
   return calImageFree(image);
}

CALresult WINAPI wine_calInit() {
        return calInit();
}

CALresult WINAPI wine_calShutdown() {
        return calShutdown();
}

CALresult WINAPI wine_calGetVersion(CALuint* major, CALuint* minor, CALuint* imp) {
        return calGetVersion(major, minor, imp);
}

CALresult WINAPI wine_calDeviceGetCount(CALuint* count) {
        return calDeviceGetCount(count);
}

CALresult WINAPI wine_calDeviceGetAttribs(CALdeviceattribs* attribs, CALuint ordinal) {
   CALresult res = calDeviceGetAttribs(attribs, ordinal);
   return res;
   
}

CALresult WINAPI wine_calDeviceOpen(CALdevice* dev, CALuint ordinal) {
        return calDeviceOpen(dev, ordinal);
}

CALresult WINAPI wine_calDeviceGetStatus(CALdevicestatus* status, CALdevice dev) {
        return calDeviceGetStatus(status, dev);
}

CALresult WINAPI wine_calDeviceGetInfo(CALdeviceinfo* info, CALuint ordinal) {
   return calDeviceGetInfo(info, ordinal);
}


CALresult WINAPI wine_calDeviceClose(CALdevice dev) {
        return calDeviceClose(dev);
}

CALresult WINAPI wine_calResAllocLocal2D(CALresource* res, CALdevice device, CALuint width, CALuint height, CALformat format, CALuint flags) {
        return calResAllocLocal2D(res, device, width, height, format, flags);
}

CALresult WINAPI wine_calResAllocRemote2D(CALresource* res, CALdevice* sharedDevices, CALuint deviceCount, CALuint width, CALuint height, CALformat format, CALuint flags) {
        return calResAllocRemote2D(res, sharedDevices, deviceCount, width, height, format, flags);
}

CALresult WINAPI wine_calResAllocLocal1D(CALresource* res, CALdevice device, CALuint width, CALformat format, CALuint flags) {
        return calResAllocLocal1D(res, device, width, format, flags);
}

CALresult WINAPI wine_calResAllocRemote1D(CALresource* res, CALdevice* sharedDevices, CALuint deviceCount, CALuint width, CALformat format, CALuint flags) {
        return calResAllocRemote1D(res, sharedDevices, deviceCount, width, format, flags);
}

CALresult WINAPI wine_calResFree(CALresource res) {
        return calResFree(res);
}

CALresult WINAPI wine_calResMap(CALvoid** pPtr, CALuint* pitch, CALresource res, CALuint flags) {
        return calResMap(pPtr, pitch, res, flags);
}

CALresult WINAPI wine_calResUnmap(CALresource res) {
        return calResUnmap(res);
}

CALresult WINAPI wine_calCtxCreate(CALcontext* ctx, CALdevice dev) {
        return calCtxCreate(ctx, dev);
}

CALresult WINAPI wine_calCtxDestroy(CALcontext ctx) {
        return calCtxDestroy(ctx);
}

CALresult WINAPI wine_calCtxGetMem(CALmem* mem, CALcontext ctx, CALresource res) {
        return calCtxGetMem(mem, ctx, res);
}

CALresult WINAPI wine_calCtxReleaseMem(CALcontext ctx, CALmem mem) {
        return calCtxReleaseMem(ctx, mem);
}

CALresult WINAPI wine_calCtxSetMem(CALcontext ctx, CALname name, CALmem mem) {
        return calCtxSetMem(ctx, name, mem);
}

CALresult WINAPI wine_calModuleLoad(CALmodule* module, CALcontext ctx, CALimage image) {
   return calModuleLoad(module, ctx, image);
}

CALresult WINAPI wine_calModuleUnload(CALcontext ctx, CALmodule module) {
        return calModuleUnload(ctx, module);
}

CALresult WINAPI wine_calModuleGetEntry(CALfunc* func, CALcontext ctx, CALmodule module, const CALchar* procName) {
        return calModuleGetEntry(func, ctx, module, procName);
}

CALresult WINAPI wine_calModuleGetName(CALname* name, CALcontext ctx, CALmodule module, const CALchar* symbolName) {
        return calModuleGetName(name, ctx, module, symbolName);
}

CALresult WINAPI wine_calImageRead(CALimage* image, CALvoid* buffer, CALuint size){
   return calImageRead(image, buffer, size);
}

/*CALresult WINAPI wine_calclImageWrite(CALvoid* buffer, CALuint size, CALimage image){
   return calclImageWrite(buffer, size, image);
}*/

CALresult WINAPI wine_calCtxRunProgram(CALevent* event, CALcontext ctx, CALfunc func, const CALdomain* rect) {
        return calCtxRunProgram(event, ctx, func, rect);
}

CALresult WINAPI wine_calCtxRunProgramGrid(CALevent* event, CALcontext ctx, CALprogramGrid* pProgramGrid) {
        return calCtxRunProgramGrid(event, ctx, pProgramGrid);
}

CALresult WINAPI wine_calCtxRunProgramGridArray(CALevent* event, CALcontext ctx, CALprogramGridArray* pGridArray) {
        return calCtxRunProgramGridArray(event, ctx, pGridArray);
}

CALresult WINAPI wine_calMemCopy(CALevent* event, CALcontext ctx, CALmem srcMem, CALmem dstMem, CALuint flags) {
        return calMemCopy(event, ctx, srcMem, dstMem, flags);
}

CALresult WINAPI wine_calCtxIsEventDone(CALcontext ctx, CALevent event) {
        return calCtxIsEventDone(ctx, event);
}

CALresult WINAPI wine_calCtxFlush(CALcontext ctx){
   return calCtxFlush(ctx);
}


CALresult WINAPI wine_calExtSupported(CALextid extid){
   return calExtSupported(extid);
}

CALresult WINAPI wine_calExtGetVersion(CALuint* major,CALuint* minor, CALextid extid) {
        return calExtGetVersion(major, minor, extid);
}

CALresult WINAPI wine_calExtGetProc(CALextproc* proc, CALextid extid, const CALchar* procname) {
        return calExtGetProc(proc, extid, procname);
}

/*CALresult WINAPI wine_calCtxCreateCounter(CALcounter* counter, CALcontext ctx, CALcountertype type) {
        return calCtxCreateCounter(counter, ctx, type);
}

CALresult WINAPI wine_calCtxDestroyCounter(CALcontext ctx, CALcounter counter) {
        return calCtxDestroyCounter(ctx, counter);
}

CALresult WINAPI wine_calCtxBeginCounter(CALcontext ctx, CALcounter counter) {
        return calCtxBeginCounter(ctx, counter);
}

CALresult WINAPI wine_calCtxEndCounter(CALcontext ctx, CALcounter counter) {
        return calCtxEndCounter(ctx, counter);
}

CALresult WINAPI wine_calCtxGetCounter(CALfloat* result, CALcontext ctx, CALcounter counter) {
        return calCtxGetCounter(result, ctx, counter);
}

CALresult WINAPI wine_calResCreate2D(CALresource* res, CALdevice *dev, CALvoid* mem, CALuint width, CALuint height, CALformat format, CALuint size, CALuint flags) {
        return calResCreate2D(res, dev, mem, width, height, format, size, flags);
}

CALresult WINAPI wine_calResCreate1D(CALresource* res, CALdevice *dev, CALvoid* mem, CALuint width, CALformat format, CALuint size, CALuint flags) {
        return calReCreate1D(res, dev, width, format, size, flags);
}*/

amdcalrt.dll.spec:
Code: Select all
#*Code by Seth Shelnutt, released under the LGPL v2.1 or later in order to be compatiable with Wine


@  stdcall calImageFree(long) wine_calImageFree
@  stdcall calGetErrorString() wine_calGetErrorString
@  stdcall calModuleGetFuncInfo(ptr long long long) wine_calModuleGetFuncInfo


@  stdcall calInit() wine_calInit
@  stdcall calShutdown() wine_calShutdown
@  stdcall calGetVersion(ptr ptr ptr) wine_calGetVersion
@  stdcall calDeviceGetCount(ptr) wine_calDeviceGetCount
@  stdcall calDeviceGetAttribs(ptr long) wine_calDeviceGetAttribs
@  stdcall calDeviceOpen(ptr long) wine_calDeviceOpen
@  stdcall calDeviceGetStatus(ptr long) wine_calDeviceGetStatus

@  stdcall calDeviceGetInfo(ptr long) wine_calDeviceGetInfo

@  stdcall calDeviceClose(long) wine_calDeviceClose
@  stdcall calResAllocLocal2D(ptr long long long long long) wine_calResAllocLocal2D
@  stdcall calResAllocRemote2D(ptr ptr long long long long long) wine_calResAllocRemote2D
@  stdcall calResAllocLocal1D(ptr long long long long) wine_calResAllocLocal1D
@  stdcall calResAllocRemote1D(ptr ptr long long long long) wine_calResAllocRemote1D
@  stdcall calResFree(long) wine_calResFree
@  stdcall calResMap(ptr ptr long long) wine_calResMap
@  stdcall calResUnmap(long) wine_calResUnmap
@  stdcall calCtxCreate(ptr long) wine_calCtxCreate
@  stdcall calCtxDestroy(long) wine_calCtxDestroy
@  stdcall calCtxGetMem(ptr long long) wine_calCtxGetMem
@  stdcall calCtxReleaseMem(long long) wine_calCtxReleaseMem
@  stdcall calCtxSetMem(long long long) wine_calCtxSetMem
@  stdcall calModuleLoad(ptr long long) wine_calModuleLoad
@  stdcall calModuleUnload(long long) wine_calModuleUnload
@  stdcall calModuleGetEntry(ptr long long ptr) wine_calModuleGetEntry
@  stdcall calModuleGetName(ptr long long ptr) wine_calModuleGetName
@  stdcall calImageRead(ptr ptr long) wine_calImageRead
#@  stdcall calclImageWrite(ptr long long) wine_calclImageWrite
@  stdcall calCtxRunProgram(ptr long long ptr) wine_calCtxRunProgram
@  stdcall calCtxRunProgramGrid(ptr long ptr) wine_calCtxRunProgramGrid
@  stdcall calCtxRunProgramGridArray(ptr long ptr) wine_calCtxRunProgramGridArray
@  stdcall calMemCopy(ptr long long long long) wine_calMemCopy
@  stdcall calCtxIsEventDone(long long) wine_calCtxIsEventDone
@  stdcall calCtxFlush(long) wine_calCtxFlush

@  stdcall calExtSupported(long) wine_calExtSupported
@  stdcall calExtGetVersion(ptr ptr long) wine_calExtGetVersion
@  stdcall calExtGetProc(ptr long ptr) wine_calExtGetProc
#@  stdcall calCtxCreateCounter(ptr long long) wine_calCtxCreateCounter
#@  stdcall calCtxDestroyCounter(long long) wine_calCtxDestroyCounter
#@  stdcall calCtxBeginCounter(long long) wine_calCtxBeginCounter
#@  stdcall calCtxEndCounter(long long) wine_calCtxEndCounter
#@  stdcall calCtxGetCounter(ptr long long) wine_calCtxGetCounter
#@  stdcall calResCreate2D(ptr long ptr long long long long long) wine_calResCreate2D
#@  stdcall calResCreate1D(ptr long ptr long long long long) wine_calResCreate1D

amdcalcl.c:
Code: Select all
/*Code by Seth Shelnutt, released under the LGPL v2.1 or later in order to be compatiable with Wine
*/
#include <windef.h>

#include "calcl.h"

CALresult WINAPI wine_calclLink(CALimage* image, CALobject* obj, CALuint objCount) {
        return calclLink(image, obj, objCount);
}

const CALchar* WINAPI wine_calclGetErrorString() {
        return calclGetErrorString();
}

CALresult WINAPI wine_calclCompile(CALobject* obj, CALlanguage language, const CALchar* source, CALtarget target) {
        return calclCompile(obj, language, source, target);
}


CALresult WINAPI wine_calclGetVersion(CALuint* major, CALuint* minor, CALuint* imp) {
   return calclGetVersion(major, minor, imp);
}

CALresult WINAPI wine_calclFreeObject(CALobject obj) {
   return calclFreeObject(obj);
}

CALresult WINAPI wine_calclFreeImage(CALimage image) {
   return calclFreeImage(image);
}

void WINAPI wine_calclDisassembleImage(const CALimage image, CALLogFunction logfunc) {
   return calclDisassembleImage(image, logfunc);
}

CALresult WINAPI wine_calclAssembleObject(CALobject* obj, CALCLprogramType type, const CALchar* source, CALtarget target) {
   return calclAssembleObject(obj, type, source, target);
}

void WINAPI wine_calclDisassembleObject(const CALobject* obj, CALLogFunction logfunc){
   return calclDisassembleObject(obj, logfunc);
}
CALresult WINAPI wine_calclImageGetSize (CALuint* size, CALimage image) {
   return calclImageGetSize (size, image);
}

CALresult WINAPI wine_calclImageWrite(CALvoid* buffer, CALuint size, CALimage image) {
   return calclImageWrite(buffer, size, image);
}

amdcalcl.dll.spec:
Code: Select all
#*Code by Seth Shelnutt, released under the LGPL v2.1 or later in order to be compatiable with Wine



@  stdcall calclLink(ptr ptr long) wine_calclLink
@  stdcall calclGetErrorString() wine_calclGetErrorString
@  stdcall calclCompile(ptr long ptr long) wine_calclCompile




@  stdcall calclGetVersion(ptr ptr ptr) wine_calclGetVersion
@  stdcall calclFreeObject(long) wine_calclFreeObject
@  stdcall calclFreeImage(long) wine_calclFreeImage
@  stdcall calclDisassembleImage(long long) wine_calclDisassembleImage
@  stdcall calclAssembleObject(ptr long ptr long) wine_calclAssembleObject
@  stdcall calclDisassembleObject(ptr long) wine_calclDisassembleObject
@  stdcall calclImageGetSize (ptr long) wine_calclImageGetSize
@  stdcall calclImageWrite(ptr long long) wine_calclImageWrite
Anachron
 
Posts: 50
Joined: Fri Mar 14, 2008 12:10 pm

Re: CAL Wrapper (Ready for Testing)

Postby vvoelz » Tue Jun 16, 2009 5:10 pm

Bjornar et al:

So far, here's what I know about Project 5746 (Run 4, Clone 20, Gen 256):
The WU got sent to our ATI server (vspg2v2), and an error was reported for it:
Code: Select all
ERROR:  Client 88.88.40.161 (1/8/15/2000) reports Core error in WU (5746,4,256,20)
C7      Bjornar 31574   9341726068237910062     0.00    1.000   3       88.88.40.161    (5746,4,256,20) c414d92c4f1cf8a862e2c43bae087c25        12948   0       1       8       15      2000    12451582621245145314       1410715171


The only returned file was called "frame256.instability", which (appropriately enough) usually indicates that there's some instability in the integration, as small errors pile up. Some other WUs that were returned earlier (frame 251 for example) had problems in this area too, so it's possible you just got an "pre-unstabilized" WU :)

In any case, here's the output of "frame256.instability", if it's helpful (I'm not an expert on this, sorry):
Code: Select all
[server@vspg2 CLONE20]$ cat frame256.instability
-----
Log file opened: nodeid 0, nnodes = 1, host = unknown, process = 0
Gromacs code copyright(c) 1991-2002
University of Groningen, The Netherlands
GPU library and other code (c) Stanford University 2006
This program is distributed under a special license to Folding@Home
See http://folding.stanford.edu/gromacs.html for license details

++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
E. Lindahl and B. Hess and D. van der Spoel
GROMACS 3.0: A package for molecular simulation and trajectory analysis
J. Mol. Mod. 7 (2001) pp. 306-317
-------- -------- --- Thank You --- -------- --------


++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
H. J. C. Berendsen, D. van der Spoel and R. van Drunen
GROMACS: A message-passing parallel molecular dynamics implementation
Comp. Phys. Comm. 91 (1995) pp. 43-56
-------- -------- --- Thank You --- -------- --------

Main: Input args 14
 mdrun_gpu -quiet -s work/wudata_02.tpr -o work/wudata_02.trr -e work/wudata_02.edr -c work/wudata_02.gro -g work/wudata_02.log -x work/wudata_02.xtc

*-----------------------------------------------------------------------------------------*
GROMACS_FAH
Version:     1.24 (Fri Feb 6 11:30:37 PST 2009) ATI
Compilation: 11:30:43
Compiler:   
Build host:  amoeba

*-----------------------------------------------------------------------------------------*
There are 0 atoms for free energy perturbation
Input Parameters:
   integrator           = sd
   nsteps               = 10000000
   ns_type              = Simple
   nstlist              = 20000001
   ndelta               = 2
   bDomDecomp           = FALSE
   decomp_dir           = 0
   nstcomm              = -10
   nstlog               = 250000
   nstxout              = 0
   nstvout              = 0
   nstfout              = 0
   nstenergy            = 250000
   nstxtcout            = 250000
   init_t               = 5.12e+006
   delta_t              = 0.002
   xtcprec              = 1000
   nkx                  = 0
   nky                  = 0
   nkz                  = 0
   pme_order            = 4
   ewald_rtol           = 1e-005
   ewald_geometry       = 0
   epsilon_surface      = 0
   optimize_fft         = FALSE
   ePBC                 = no
   bUncStart            = TRUE
   bShakeSOR            = FALSE
   etc                  = No
   epc                  = No
   epctype              = Isotropic
   tau_p                = 1
   ref_p (3x3):
      ref_p[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   compress (3x3):
      compress[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   bSimAnn              = FALSE
   zero_temp_time       = 0
   rlist                = 100
   coulombtype          = Cut-off
   rcoulomb_switch      = 0
   rcoulomb             = 100
   vdwtype              = Cut-off
   rvdw_switch          = 0
   rvdw                 = 100
   epsilon_r            = 1
   DispCorr             = No
   fudgeQQ              = 0.8333
   free_energy          = no
   init_lambda          = 0
   sc_alpha             = 0
   sc_sigma             = 0.3
   delta_lambda         = 0
   disre_weighting      = Conservative
   disre_mixed          = FALSE
   dr_fc                = 1000
   dr_tau               = 0
   nstdisreout          = 100
   orires_fc            = 0
   orires_tau           = 0
   nstorireout          = 100
   em_stepsize          = 0.01
   em_tol               = 100
   niter                = 20
   fc_stepsize          = 0
   nstcgsteep           = 1000
   ConstAlg             = Shake
   shake_tol            = 0.0001
   lincs_order          = 4
   lincs_warnangle      = 30
   bd_temp              = 300
   bd_fric              = 0
   ld_seed              = 10122
   cos_accel            = 0
   userint1             = 0
   userint2             = 0
   userint3             = 0
   userint4             = 0
   userreal1            = 0
   userreal2            = 0
   userreal3            = 0
   userreal4            = 0
   solvent-model        = GBSA
   sasa                 = None
   nstsasa              = 1
   rsasa                = 0.55
   epsilon_dd           = 20
   nstsasaout           = 0
grpopts:
   nrdf:                2435
   ref_t:                370
   tau_t:            0.01099
   acc:            0           0           0
   nfreeze:           N           N           N
   energygrp_excl[  0]: 0
   efield-x:
      n = 0
   efield-xt:
      n = 0
   efield-y:
      n = 0
   efield-yt:
      n = 0
   efield-z:
      n = 0
   efield-zt:
      n = 0
CPU=  0, lastcg=  970, targetcg=  485, myshift=    0
nsb->shift =   1, nsb->bshift=  0
Neighbor Search Blocks
nsb->nodeid:         0
nsb->nnodes:      1
nsb->cgtotal:   971
nsb->natoms:    973
nsb->shift:       1
nsb->bshift:      0
Nodeid   index  homenr  cgload  workload
     0       0     973     971       971

Initializing gpu library w/ defaultStreamWidth=32
Test mode: disabled
Using runtime cal; initializing Brook
############

info_string:
CAL version: 1.4.255
CAL: 1 devices found
CAL: Using device 0
ASIC = RV770
Available board memory to CAL: 512MB
Available cached memory to CAL: 64MB
Available uncached memory to CAL: 256MB
Engine clock: 800 (0.000000)
Natural vector width: 64
Recommended minimum thread count: 2560
Brook yield: 0
Don't flush kernels: 0
Pre flush command buffers: 0
Flush interval: 16

############
Number of atoms is odd, won't work
  natoms:973 ceiling=976
  atoms: 31x32
  unroll: i=4 j=4
  useNew=4
Done w/ Gpu initialization
Adjusted gmx variables to match streams: atoms=973 new size=992
Max number of bonds per atom is 4
Table routines are used for coulomb: FALSE
Table routines are used for vdw:     FALSE
Cut-off's:   NS: 100   Coulomb: 100   LJ: 100
Generated table with 500 data points for COUL.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ6.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ12.
Tabscale = 500 points/nm
Going to determine what solvent types we have.
There are 1 molecules, 971 charge groups and 973 atoms
There are 0 optimized solvent molecules on node 0
There are 0 optimized water molecules on node 0
setupForRestartSave=0 restartStep=0
forceCheck=0 maxSteps=0
GPU Warning: Output energies are not correct
selfTestRandomI=2524719
gpu_center_molecule: Coordinates pre centering 0
0 0 x[     0.07476      0.26675      2.01031]
0 1 x[    -0.01407      0.28109      1.96443]
0 2 x[     0.05462      0.28191      2.10811]
0 3 x[     0.14441      0.33084      1.97507]
0 4 x[     0.11844      0.13006      1.98375]

gpu_center_molecule: Coordinates post centering 0 com[-0.1913 -0.2169 -0.1616]
0 0 x[     0.26601      0.48362      2.17188]
0 1 x[     0.17719      0.49796      2.12599]
0 2 x[     0.24588      0.49879      2.26967]
0 3 x[     0.33567      0.54772      2.13664]
0 4 x[     0.30969      0.34693      2.14532]

Centered molecule
Center of mass motion removal mode is Angular
We have the following groups for center of mass motion removal:
  0:  rest, initial mass: 7011.73
Started mdrun on node 0 Tue Jun 16 11:42:05 2009
Initial temperature: 430.496 K
system name is Protein
Downloaded x/v/vold/mInvU
0 x[     0.26601      0.48362      2.17188] v[     0.16320     -0.27768     -0.17488] vold[     0.16320     -0.27768     -0.17488] m[    14.01000]
1 x[     0.17719      0.49796      2.12599] v[    -0.78389     -3.17911      0.56918] vold[    -0.78389     -3.17911      0.56918] m[     1.00800]
2 x[     0.24588      0.49879      2.26967] v[     1.25266     -2.71580      0.50526] vold[     1.25266     -2.71580      0.50526] m[     1.00800]
3 x[     0.33567      0.54772      2.13664] v[     0.45676     -1.31498     -1.52614] vold[     0.45676     -1.31498     -1.52614] m[     1.00800]
4 x[     0.30969      0.34693      2.14532] v[    -0.35444      0.36603      0.23805] vold[    -0.35444      0.36603      0.23805] m[    12.01000]

Using normal Coulomb/LJ & GBSA
Calculated RepFac: 12 minSuggestedThreads=2560
Setting RepFac: 12
Initialized strExcl XX (height x width) 248 x 992=246016
Initialized NB parameters (height x width)
  force :47x64
  excl  :248x992
  jstrm :16x16 sz=256 256
  nAtomsCeiling=976 iUnroll=4 repfac=12
Done gpuInitNBCalc

Using 4 new gpuSetExclusionsDone gpuSetExclusions

Using new gpuSetVDWParamsDone gpuSetVDWParams
Calling cpuSetObcParameters 1 1 2 1.000 78.300

gpuSetObcParameters:  No atoms=973 agb file length=4135
GpuObc::GBSAStreamsFactory stream values appear to be in order.
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadiiWithDielectricOffset cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcScaledAtomicRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcChain cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcIntermediateForce cpuBlockSz=15872 float=4 size=992 [31 32] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii2 cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_1 cpuBlockSz=95232 float=4 size=5952 [93 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_2 cpuBlockSz=95232 float=4 size=5952 [93 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_1 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_2 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_3 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_4 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
 0. ObcAtomicRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 1. ObcAtomicRadiiWithDielectricOffset [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 2. ObcScaledAtomicRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 3. ObcBornRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 4. ObcChain [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 5. ObcIntermediateForce [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=3968
 6. ObcBornRadii2 [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 7. ObcForceLoop1_1 [64 93] Sz=93 Rep=12 Unroll=2 Chnk=12 Data=5952 BlkSz=23808
 8. ObcForceLoop1_2 [64 93] Sz=93 Rep=12 Unroll=2 Chnk=12 Data=5952 BlkSz=23808
 9. ObcForceLoop2_1 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
10. ObcForceLoop2_2 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
11. ObcForceLoop2_3 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
12. ObcForceLoop2_4 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
gpuSetObcParameters:  includeAce=1 soluteDiel=1.000 solventDiel=78.300.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii read 55 lines from agb parameter file.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii
   agb parameter file line=<# This file was created by edgar to supply Macromodel radii for AMBER atomypes> is being skipped.
   agb parameter file line=<#> is being skipped.
   agb parameter file line=<@   title "agbnp parameters"> is being skipped.
   agb parameter file line=<@   xaxis label "OPLS atomtype"> is being skipped.
   agb parameter file line=<@   yaxis label "parameter"> is being skipped.
   agb parameter file line=<@ s0 legend "radius"> is being skipped.
   agb parameter file line=<@ s1 legend "igamma"> is being skipped.
   agb parameter file line=<@ s2 legend "ialpha"> is being skipped.
   agb parameter file line=<@ s3 legend "idelta"> is being skipped.
   agb parameter file line=<@ s4 legend "sgamma"> is being skipped.
   agb parameter file line=<@ s5 legend "salpha"> is being skipped.
   agb parameter file line=<@ s6 legend "sdelta"> is being skipped.
   agb parameter file line=<@ s6 legend "ai"@ s7 legend "GBdistcorr"> is being skipped.
   agb parameter file line=<# the following values are from the default behaviour.> is being skipped.
gpuSetObcParameters:  obtained atomic radii from tar file
gpuSetObcParameters:  writing parameters to board
gpuSetObcParameters:  done OBC setup
do_md: OBC setup completed.
In gpuBondedParams_Init
In gpuBondedParams_Init Start param setup
In gpuBondedParams_Init Post param setup
nbondeds=2962 stream dimensions: [62 48] 2976
Writing params to board -- done
gpuBondedParams_Init: parameter setup complete, nbonds=2962.
gpuBondedParams_Init: calling LoadInvMaps
InvMap completed malloc natoms=973 nbondeds=2962 AtomStrSize=992 992 [31 32]
InvMap completed gpuCalcInvMap counts=4 [31 32] 992
InvMap I-stream nimaps=3 BONDED_INVMAP_ISTREAMS=9
Writing I-stream to board  -- done
I-Map:
InvMap J-stream njmaps=5 BONDED_INVMAP_JSTREAMS=6
Writing J-stream to board  -- done
J-Map
InvMap K-stream nkmaps=4 BONDED_INVMAP_KSTREAMS=6
Writing K-stream to board  -- done
InvMap calling gpuCalcInvMap(3, 4)
K-map

InvMap L-stream nlmaps=2 BONDED_INVMAP_LSTREAMS=9
Writing L-stream to board
L-map
  -- done
Bondeds use 3 inverse maps for i
Bondeds use 5 inverse maps for j
Bondeds use 4 inverse maps for k
Bondeds use 2 inverse maps for l
gpuBondedParams_Init: LoadInvMaps completed nbondeds=2962 bp->size=2976
gpuBondedParams_Init w i=2962 j=4 4i+j=11847
Writing atom indices to board -- done
Freeing buffers
gpuBondedParams_Init creating force streams
Done gpuBondedParams_Init
Done gpuSetBondedParams
bFFscan=0 -- if set, then original coords copied
bTCR=0 bGlas=0
generateTestHarnessFiles=0
COM removal: frequency==-10 Remove COM on Cpu=0
Sd parameters: ngtc=1 tau=1.099e-002 deltaT=2.000e-003 T=3.700e+002
StopCM every -10 steps -- requested ANGULAR
Neighbor list search=20000001 --  set to 0.
Merge Obc & CDLJ force calculations=0
force: Bonded=on NB=on OBC=on
           Step           Time         Lambda      Annealing
              0   5120000.00000        0.00000        1.00000

There are 973 atoms in your xtc output selection
Running low_constrain on GPU
gpuShakeHParams_Init: shakeh blocks 312
Initialized shakeH streams of size 20x16

++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
J. P. Ryckaert and G. Ciccotti and H. J. C. Berendsen
Numerical Integration of the Cartesian Equations of Motion of a System with
Constraints; Molecular Dynamics of n-Alkanes
J. Comp. Phys. 23 (1977) pp. 327-341
-------- -------- --- Thank You --- -------- --------

LoadGVStreams seed=10122 size rv texture=1048576 RngAlg=Kiss Rng
LoadGVStreamsKiss: reset state seeds stateInitialized=0 reseed=5
LoadGVStreamsKiss: count=1.0 ttl=6.291e+006 no./count=6291456.0 1048576 2
 -- done

gpuSDParams_Init: Setting sd2x w/ random values
Writing strSD2X -- done
Building permutation for shuffle...done
update: Calling gpuUpdateSD1 step=0
gpuUpdateSD1 using random value kernel.
update: Calling gpuShakeH(1) step=0
gpuShakeH: maxIterations=25 tolerance=1.000000e-004
gpuUpdateSD2 using random value kernel.
force: Bonded=on NB=on OBC=on
update: Calling gpuUpdateSD1 step=1
force: Bonded=on NB=on OBC=on
LoadGVStreamsKiss: count=2.0 ttl=1.258e+007 no./count=6291456.0 1048576 2

User avatar
vvoelz
Pande Group Member
 
Posts: 307
Joined: Sun Dec 02, 2007 8:07 pm
Location: Temple University, Philadelphia PA

Re: CAL Wrapper (Ready for Testing)

Postby Shelnutt2 » Wed Jun 17, 2009 3:05 pm

Reading through everything, it is very difficult to determine where the error is. I've got four theories. 1) It's in the wrapper . 2) It's in Wine's handling of the wrapper/ati card/the client. 3) It's the difference between windows and linux CAL, or 4) Something else.

From the output frame256.instability , I don't see any errors at all. From goolging guardedrun and core status 72. It leads me to believe that is possible that there are memory errors on the GPU, meaning between wine and the linux drivers it's not allocating the memory properly. However in the frame256.instability , it shows the right mb for memory, so I'm not 100%.

Try running wine with debug enabled again, and see if we can't get more output.
Shelnutt2
 
Posts: 150
Joined: Sun Dec 02, 2007 5:37 pm

Re: CAL Wrapper (Ready for Testing)

Postby bruce » Wed Jun 17, 2009 3:19 pm

My theory is that it's 4) Something else, but we don't have any evidence yet and the others are certainly possibilities.

There are a small percentage of WUs that are ill-conditioned and they spontaneously produce errors (EUEs). You probably got one of those WUs.

Please post FAHlog.txt so we can determine what was actually reported in the standard output.
bruce
 
Posts: 22022
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: CAL Wrapper (Ready for Testing)

Postby Anachron » Wed Jun 17, 2009 6:43 pm

There does not seem to be an error with the WU, as it runs fine in XP. Here are the logs for it running in Windows:

FahLog.txt:
Code: Select all

--- Opening Log file [June 17 17:35:09 UTC]





# Windows GPU Console Edition #################################################

###############################################################################



                       Folding@Home Client Version 6.23



                          http://folding.stanford.edu



###############################################################################

###############################################################################



Launch directory: C:\Documents and Settings\Hansen\Skrivebord\folding-gpu2

Executable: C:\Documents and Settings\Hansen\Skrivebord\folding-gpu2\fah6-gpu.exe

Arguments: -oneunit -forcegpu r600 -verbosity 9 -oneunit -verbosity 9



[17:35:09] - Ask before connecting: Yes

[17:35:09] - User name: Anonymous (Team 0)

[17:35:09] - User ID: 11FCF94873390A40

[17:35:09] - Machine ID: 2

[17:35:09]

[17:35:09] Loaded queue successfully.

[17:35:09]

[17:35:09] - Autosending finished units... [June 17 17:35:09 UTC]

[17:35:09] + Processing work unit

[17:35:09] Trying to send all finished work units

[17:35:09] Core required: FahCore_11.exe

[17:35:09] + No unsent completed units remaining.

[17:35:09] - Autosend completed

[17:35:09] Core found.

[17:35:09] Working on queue slot 02 [June 17 17:35:09 UTC]

[17:35:09] + Working ...

[17:35:09] - Calling '.\FahCore_11.exe -dir work/ -suffix 02 -checkpoint 15 -verbose -lifeline 1864 -version 623'


[17:35:10]

[17:35:10] *------------------------------*

[17:35:10] Folding@Home GPU Core - Beta

[17:35:10] Version 1.24 (Mon Feb 9 11:00:12 PST 2009)

[17:35:10]

[17:35:10] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86

[17:35:10] Build host: amoeba

[17:35:10] Board Type: AMD

[17:35:10] Core      :

[17:35:10] Preparing to commence simulation

[17:35:10] - Looking at optimizations...

[17:35:10] - Files status OK

[17:35:10] - Expanded 68563 -> 357580 (decompressed 521.5 percent)

[17:35:10] Called DecompressByteArray: compressed_data_size=68563 data_size=357580, decompressed_data_size=357580 diff=0

[17:35:10] - Digital signature verified

[17:35:10]

[17:35:10] Project: 5746 (Run 4, Clone 20, Gen 256)

[17:35:10]

[17:35:10] Assembly optimizations on if available.

[17:35:10] Entering M.D.

[17:35:16] Will resume from checkpoint file

[17:35:16] Tpr hash work/wudata_02.tpr:  2639730865 4031526054 2357533253 3358095351 301830008

[17:35:17] Working on Protein

[17:35:17] Client config found, loading data.

[17:35:17] Starting GUI Server

[17:35:22] Resuming from checkpoint

[17:35:22] fcCheckPointResume: retreived and current tpr file hash:

[17:35:22]    0   2639730865   2639730865

[17:35:22]    1   4031526054   4031526054

[17:35:22]    2   2357533253   2357533253

[17:35:22]    3   3358095351   3358095351

[17:35:22]    4    301830008    301830008

[17:35:22] Verified work/wudata_02.log

[17:35:22] Verified work/wudata_02.edr

[17:35:22] Verified work/wudata_02.xtc

[17:37:07] Completed 1%

[17:38:51] Completed 2%

[17:40:35] Completed 3%

[17:42:19] Completed 4%

[17:44:03] Completed 5%

[17:44:18] ***** Got a SIGTERM signal (2)

[17:44:18] Killing all core threads



Folding@Home Client Shutdown.



wudata_2.log:
Code: Select all
Log file opened: nodeid 0, nnodes = 1, host = unknown, process = 0
Gromacs code copyright(c) 1991-2002
University of Groningen, The Netherlands
GPU library and other code (c) Stanford University 2006
This program is distributed under a special license to Folding@Home
See http://folding.stanford.edu/gromacs.html for license details

++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
E. Lindahl and B. Hess and D. van der Spoel
GROMACS 3.0: A package for molecular simulation and trajectory analysis
J. Mol. Mod. 7 (2001) pp. 306-317
-------- -------- --- Thank You --- -------- --------


++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
H. J. C. Berendsen, D. van der Spoel and R. van Drunen
GROMACS: A message-passing parallel molecular dynamics implementation
Comp. Phys. Comm. 91 (1995) pp. 43-56
-------- -------- --- Thank You --- -------- --------

Main: Input args 14
 mdrun_gpu -quiet -s work/wudata_02.tpr -o work/wudata_02.trr -e work/wudata_02.edr -c work/wudata_02.gro -g work/wudata_02.log -x work/wudata_02.xtc

*-----------------------------------------------------------------------------------------*
GROMACS_FAH
Version:     1.24 (Fri Feb 6 11:30:37 PST 2009) ATI
Compilation: 11:30:43
Compiler:   
Build host:  amoeba

*-----------------------------------------------------------------------------------------*
There are 0 atoms for free energy perturbation
Input Parameters:
   integrator           = sd
   nsteps               = 10000000
   ns_type              = Simple
   nstlist              = 20000001
   ndelta               = 2
   bDomDecomp           = FALSE
   decomp_dir           = 0
   nstcomm              = -10
   nstlog               = 250000
   nstxout              = 0
   nstvout              = 0
   nstfout              = 0
   nstenergy            = 250000
   nstxtcout            = 250000
   init_t               = 5.12e+006
   delta_t              = 0.002
   xtcprec              = 1000
   nkx                  = 0
   nky                  = 0
   nkz                  = 0
   pme_order            = 4
   ewald_rtol           = 1e-005
   ewald_geometry       = 0
   epsilon_surface      = 0
   optimize_fft         = FALSE
   ePBC                 = no
   bUncStart            = TRUE
   bShakeSOR            = FALSE
   etc                  = No
   epc                  = No
   epctype              = Isotropic
   tau_p                = 1
   ref_p (3x3):
      ref_p[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   compress (3x3):
      compress[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   bSimAnn              = FALSE
   zero_temp_time       = 0
   rlist                = 100
   coulombtype          = Cut-off
   rcoulomb_switch      = 0
   rcoulomb             = 100
   vdwtype              = Cut-off
   rvdw_switch          = 0
   rvdw                 = 100
   epsilon_r            = 1
   DispCorr             = No
   fudgeQQ              = 0.8333
   free_energy          = no
   init_lambda          = 0
   sc_alpha             = 0
   sc_sigma             = 0.3
   delta_lambda         = 0
   disre_weighting      = Conservative
   disre_mixed          = FALSE
   dr_fc                = 1000
   dr_tau               = 0
   nstdisreout          = 100
   orires_fc            = 0
   orires_tau           = 0
   nstorireout          = 100
   em_stepsize          = 0.01
   em_tol               = 100
   niter                = 20
   fc_stepsize          = 0
   nstcgsteep           = 1000
   ConstAlg             = Shake
   shake_tol            = 0.0001
   lincs_order          = 4
   lincs_warnangle      = 30
   bd_temp              = 300
   bd_fric              = 0
   ld_seed              = 10122
   cos_accel            = 0
   userint1             = 0
   userint2             = 0
   userint3             = 0
   userint4             = 0
   userreal1            = 0
   userreal2            = 0
   userreal3            = 0
   userreal4            = 0
   solvent-model        = GBSA
   sasa                 = None
   nstsasa              = 1
   rsasa                = 0.55
   epsilon_dd           = 20
   nstsasaout           = 0
grpopts:
   nrdf:           2435
   ref_t:            370
   tau_t:        0.01099
   acc:              0           0           0
   nfreeze:           N           N           N
   energygrp_excl[  0]: 0
   efield-x:
      n = 0
   efield-xt:
      n = 0
   efield-y:
      n = 0
   efield-yt:
      n = 0
   efield-z:
      n = 0
   efield-zt:
      n = 0
CPU=  0, lastcg=  970, targetcg=  485, myshift=    0
nsb->shift =   1, nsb->bshift=  0
Neighbor Search Blocks
nsb->nodeid:         0
nsb->nnodes:      1
nsb->cgtotal:   971
nsb->natoms:    973
nsb->shift:       1
nsb->bshift:      0
Nodeid   index  homenr  cgload  workload
     0       0     973     971       971

Initializing gpu library w/ defaultStreamWidth=32
Test mode: disabled
Using runtime cal; initializing Brook
############

info_string:
CAL version: 1.4.255
CAL: 1 devices found
CAL: Using device 0
ASIC = RV770
Available board memory to CAL: 512MB
Available cached memory to CAL: 64MB
Available uncached memory to CAL: 256MB
Engine clock: 800 (0.000000)
Natural vector width: 64
Recommended minimum thread count: 2560
Brook yield: 0
Don't flush kernels: 0
Pre flush command buffers: 0
Flush interval: 16

############
Number of atoms is odd, won't work
  natoms:973 ceiling=976
  atoms: 31x32
  unroll: i=4 j=4
  useNew=4
Done w/ Gpu initialization
Adjusted gmx variables to match streams: atoms=973 new size=992
Max number of bonds per atom is 4
Table routines are used for coulomb: FALSE
Table routines are used for vdw:     FALSE
Cut-off's:   NS: 100   Coulomb: 100   LJ: 100
Generated table with 500 data points for COUL.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ6.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ12.
Tabscale = 500 points/nm
Going to determine what solvent types we have.
There are 1 molecules, 971 charge groups and 973 atoms
There are 0 optimized solvent molecules on node 0
There are 0 optimized water molecules on node 0
setupForRestartSave=0 restartStep=0
forceCheck=0 maxSteps=0
GPU Warning: Output energies are not correct
selfTestRandomI=2524719
gpu_center_molecule: Coordinates pre centering 0
0 0 x[     0.07476      0.26675      2.01031]
0 1 x[    -0.01407      0.28109      1.96443]
0 2 x[     0.05462      0.28191      2.10811]
0 3 x[     0.14441      0.33084      1.97507]
0 4 x[     0.11844      0.13006      1.98375]

gpu_center_molecule: Coordinates post centering 0 com[-0.1913 -0.2169 -0.1616]
0 0 x[     0.26601      0.48362      2.17188]
0 1 x[     0.17719      0.49796      2.12599]
0 2 x[     0.24588      0.49879      2.26967]
0 3 x[     0.33567      0.54772      2.13664]
0 4 x[     0.30969      0.34693      2.14532]

Centered molecule
Center of mass motion removal mode is Angular
We have the following groups for center of mass motion removal:
  0:  rest, initial mass: 7011.73
Started mdrun on node 0 Tue Jun 16 11:42:05 2009
Initial temperature: 430.496 K
system name is Protein
Downloaded x/v/vold/mInvU
0 x[     0.26601      0.48362      2.17188] v[     0.16320     -0.27768     -0.17488] vold[     0.16320     -0.27768     -0.17488] m[    14.01000]
1 x[     0.17719      0.49796      2.12599] v[    -0.78389     -3.17911      0.56918] vold[    -0.78389     -3.17911      0.56918] m[     1.00800]
2 x[     0.24588      0.49879      2.26967] v[     1.25266     -2.71580      0.50526] vold[     1.25266     -2.71580      0.50526] m[     1.00800]
3 x[     0.33567      0.54772      2.13664] v[     0.45676     -1.31498     -1.52614] vold[     0.45676     -1.31498     -1.52614] m[     1.00800]
4 x[     0.30969      0.34693      2.14532] v[    -0.35444      0.36603      0.23805] vold[    -0.35444      0.36603      0.23805] m[    12.01000]

Using normal Coulomb/LJ & GBSA
Calculated RepFac: 12 minSuggestedThreads=2560
Setting RepFac: 12
Initialized strExcl XX (height x width) 248 x 992=246016
Initialized NB parameters (height x width)
  force :47x64
  excl  :248x992
  jstrm :16x16 sz=256 256
  nAtomsCeiling=976 iUnroll=4 repfac=12
Done gpuInitNBCalc

Using 4 new gpuSetExclusionsDone gpuSetExclusions

Using new gpuSetVDWParamsDone gpuSetVDWParams
Calling cpuSetObcParameters 1 1 2 1.000 78.300

gpuSetObcParameters:  No atoms=973 agb file length=4135
GpuObc::GBSAStreamsFactory stream values appear to be in order.
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadiiWithDielectricOffset cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcScaledAtomicRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcChain cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcIntermediateForce cpuBlockSz=15872 float=4 size=992 [31 32] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii2 cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_1 cpuBlockSz=95232 float=4 size=5952 [93 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_2 cpuBlockSz=95232 float=4 size=5952 [93 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_1 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_2 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_3 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_4 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
 0. ObcAtomicRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 1. ObcAtomicRadiiWithDielectricOffset [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 2. ObcScaledAtomicRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 3. ObcBornRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 4. ObcChain [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 5. ObcIntermediateForce [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=3968
 6. ObcBornRadii2 [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 7. ObcForceLoop1_1 [64 93] Sz=93 Rep=12 Unroll=2 Chnk=12 Data=5952 BlkSz=23808
 8. ObcForceLoop1_2 [64 93] Sz=93 Rep=12 Unroll=2 Chnk=12 Data=5952 BlkSz=23808
 9. ObcForceLoop2_1 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
10. ObcForceLoop2_2 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
11. ObcForceLoop2_3 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
12. ObcForceLoop2_4 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
gpuSetObcParameters:  includeAce=1 soluteDiel=1.000 solventDiel=78.300.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii read 55 lines from agb parameter file.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii
   agb parameter file line=<# This file was created by edgar to supply Macromodel radii for AMBER atomypes> is being skipped.
   agb parameter file line=<#> is being skipped.
   agb parameter file line=<@   title "agbnp parameters"> is being skipped.
   agb parameter file line=<@   xaxis label "OPLS atomtype"> is being skipped.
   agb parameter file line=<@   yaxis label "parameter"> is being skipped.
   agb parameter file line=<@ s0 legend "radius"> is being skipped.
   agb parameter file line=<@ s1 legend "igamma"> is being skipped.
   agb parameter file line=<@ s2 legend "ialpha"> is being skipped.
   agb parameter file line=<@ s3 legend "idelta"> is being skipped.
   agb parameter file line=<@ s4 legend "sgamma"> is being skipped.
   agb parameter file line=<@ s5 legend "salpha"> is being skipped.
   agb parameter file line=<@ s6 legend "sdelta"> is being skipped.
   agb parameter file line=<@ s6 legend "ai"@ s7 legend "GBdistcorr"> is being skipped.
   agb parameter file line=<# the following values are from the default behaviour.> is being skipped.
gpuSetObcParameters:  obtained atomic radii from tar file
gpuSetObcParameters:  writing parameters to board
gpuSetObcParameters:  done OBC setup
do_md: OBC setup completed.
In gpuBondedParams_Init
In gpuBondedParams_Init Start param setup
In gpuBondedParams_Init Post param setup
nbondeds=2962 stream dimensions: [62 48] 2976
Writing params to board -- done
gpuBondedParams_Init: parameter setup complete, nbonds=2962.
gpuBondedParams_Init: calling LoadInvMaps
InvMap completed malloc natoms=973 nbondeds=2962 AtomStrSize=992 992 [31 32]
InvMap completed gpuCalcInvMap counts=4 [31 32] 992
InvMap I-stream nimaps=3 BONDED_INVMAP_ISTREAMS=9
Writing I-stream to board  -- done
I-Map:
InvMap J-stream njmaps=5 BONDED_INVMAP_JSTREAMS=6
Writing J-stream to board  -- done
J-Map
InvMap K-stream nkmaps=4 BONDED_INVMAP_KSTREAMS=6
Writing K-stream to board  -- done
InvMap calling gpuCalcInvMap(3, 4)
K-map

InvMap L-stream nlmaps=2 BONDED_INVMAP_LSTREAMS=9
Writing L-stream to board
L-map
  -- done
Bondeds use 3 inverse maps for i
Bondeds use 5 inverse maps for j
Bondeds use 4 inverse maps for k
Bondeds use 2 inverse maps for l
gpuBondedParams_Init: LoadInvMaps completed nbondeds=2962 bp->size=2976
gpuBondedParams_Init w i=2962 j=4 4i+j=11847
Writing atom indices to board -- done
Freeing buffers
gpuBondedParams_Init creating force streams
Done gpuBondedParams_Init
Done gpuSetBondedParams
bFFscan=0 -- if set, then original coords copied
bTCR=0 bGlas=0
generateTestHarnessFiles=0
COM removal: frequency==-10 Remove COM on Cpu=0
Sd parameters: ngtc=1 tau=1.099e-002 deltaT=2.000e-003 T=3.700e+002
StopCM every -10 steps -- requested ANGULAR
Neighbor list search=20000001 --  set to 0.
Merge Obc & CDLJ force calculations=0
force: Bonded=on NB=on OBC=on
           Step           Time         Lambda      Annealing
              0   5120000.00000        0.00000        1.00000

There are 973 atoms in your xtc output selection
Running low_constrain on GPU
gpuShakeHParams_Init: shakeh blocks 312
Initialized shakeH streams of size 20x16

++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
J. P. Ryckaert and G. Ciccotti and H. J. C. Berendsen
Numerical Integration of the Cartesian Equations of Motion of a System with
Constraints; Molecular Dynamics of n-Alkanes
J. Comp. Phys. 23 (1977) pp. 327-341
-------- -------- --- Thank You --- -------- --------

LoadGVStreams seed=10122 size rv texture=1048576 RngAlg=Kiss Rng
LoadGVStreamsKiss: reset state seeds stateInitialized=0 reseed=5
LoadGVStreamsKiss: count=1.0 ttl=6.291e+006 no./count=6291456.0 1048576 2
 -- done

gpuSDParams_Init: Setting sd2x w/ random values
Writing strSD2X -- done
Building permutation for shuffle...done
update: Calling gpuUpdateSD1 step=0
gpuUpdateSD1 using random value kernel.
update: Calling gpuShakeH(1) step=0
gpuShakeH: maxIterations=25 tolerance=1.000000e-004
gpuUpdateSD2 using random value kernel.
force: Bonded=on NB=on OBC=on
update: Calling gpuUpdateSD1 step=1
force: Bonded=on NB=on OBC=on
Resumed from checkpoint, resent coordinates and velocities to GPU
LoadGVStreamsKiss: count=2.0 ttl=1.258e+007 no./count=6291456.0 1048576 2
LoadGVStreamsKiss: count=3.0 ttl=1.887e+007 no./count=6291456.0 1048576 2
LoadGVStreamsKiss: count=4.0 ttl=2.517e+007 no./count=6291456.0 1048576 2
LoadGVStreamsKiss: count=5.0 ttl=3.146e+007 no./count=6291456.0 1048576 2
           Step           Time         Lambda      Annealing
         250000   5120500.00000        0.00000        1.00000

LoadGVStreamsKiss: reset state seeds stateInitialized=5 reseed=5
LoadGVStreamsKiss: count=6.0 ttl=3.775e+007 no./count=6291456.0 1048576 2
LoadGVStreamsKiss: count=7.0 ttl=4.404e+007 no./count=6291456.0 1048576 2
LoadGVStreamsKiss: count=8.0 ttl=5.033e+007 no./count=6291456.0 1048576 2
LoadGVStreamsKiss: count=9.0 ttl=5.662e+007 no./count=6291456.0 1048576 2
LoadGVStreamsKiss: count=10.0 ttl=6.291e+007 no./count=6291456.0 1048576 2
           Step           Time         Lambda      Annealing
         500000   5121000.00000        0.00000        1.00000





And when continuing it in Linux Mint:

FahLog.txt:
Code: Select all

--- Opening Log file [June 17 18:15:15 UTC]





# Windows GPU Console Edition #################################################

###############################################################################



                       Folding@Home Client Version 6.23



                          http://folding.stanford.edu



###############################################################################

###############################################################################



Launch directory: C:\Program Files\folding-gpu

Executable: C:\Program Files\folding-gpu\fah6-gpu.exe

Arguments: -oneunit -verbosity 9 -forcegpu ati_r700



[18:15:15] - Ask before connecting: Yes

[18:15:15] - User name: Bjornar (Team 31574)

[18:15:15] - User ID: 2EC0534E5D7AA47F

[18:15:15] - Machine ID: 2

[18:15:15]

[18:15:16] Loaded queue successfully.

[18:15:16]

[18:15:16] + Processing work unit

[18:15:16] Core required: FahCore_11.exe

[18:15:16] + No unsent completed units remaining.

[18:15:16] - Autosend completed

[18:15:16] Core found.

[18:15:16] Working on queue slot 02 [June 17 18:15:16 UTC]

[18:15:16] + Working ...

[18:15:16] - Calling '.\FahCore_11.exe -dir work/ -suffix 02 -checkpoint 15 -verbose -lifeline 8 -version 623'


[18:15:16]

[18:15:16] *------------------------------*

[18:15:16] Folding@Home GPU Core - Beta

[18:15:16] Version 1.24 (Mon Feb 9 11:00:12 PST 2009)

[18:15:16]

[18:15:16] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86

[18:15:16] Build host: amoeba

[18:15:16] Board Type: AMD

[18:15:16] Core      :

[18:15:16] Preparing to commence simulation

[18:15:16] - Looking at optimizations...

[18:15:16] - Files status OK

[18:15:16] - Expanded 68563 -> 357580 (decompressed 521.5 percent)

[18:15:16] Called DecompressByteArray: compressed_data_size=68563 data_size=357580, decompressed_data_size=357580 diff=0

[18:15:16] - Digital signature verified

[18:15:16]

[18:15:16] Project: 5746 (Run 4, Clone 20, Gen 256)

[18:15:16]

[18:15:16] Assembly optimizations on if available.

[18:15:16] Entering M.D.

[18:15:22] Will resume from checkpoint file

[18:15:22] Tpr hash work/wudata_02.tpr:  2639730865 4031526054 2357533253 3358095351 301830008

[18:15:22] Working on Protein

[18:15:22] Client config found, loading data.

[18:15:22] Starting GUI Server

[18:15:22] Run: exception thrown during GuardedRun

[18:15:22] Run: exception thrown in GuardedRun -- Gromacs cannot continue further.

[18:15:22] Going to send back what have done -- stepsTotalG=10000000

[18:15:22] Work fraction=0.0000 steps=10000000.

[18:15:26] logfile size=16714 infoLength=16714 edr=0 trr=23

[18:15:26] - Writing 17250 bytes of core data to disk...

[18:15:26] Done: 16738 -> 5547 (compressed to 33.1 percent)

[18:15:27]   ... Done.

[18:15:27]

[18:15:27] Folding@home Core Shutdown: EARLY_UNIT_END

[18:15:30] CoreStatus = 72 (114)

[18:15:30] Sending work to server

[18:15:30] Project: 5746 (Run 4, Clone 20, Gen 256)





[18:15:30] + Attempting to send results [June 17 18:15:30 UTC]

[18:15:30] - Reading file work/wuresults_02.dat from core

[18:15:30]   (Read 6059 bytes from disk)

[18:15:30] - Presenting message box asking to network.



wudata_02.log_tmp:
Code: Select all
Log file opened: nodeid 0, nnodes = 1, host = unknown, process = 0
Gromacs code copyright(c) 1991-2002
University of Groningen, The Netherlands
GPU library and other code (c) Stanford University 2006
This program is distributed under a special license to Folding@Home
See http://folding.stanford.edu/gromacs.html for license details

++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
E. Lindahl and B. Hess and D. van der Spoel
GROMACS 3.0: A package for molecular simulation and trajectory analysis
J. Mol. Mod. 7 (2001) pp. 306-317
-------- -------- --- Thank You --- -------- --------


++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
H. J. C. Berendsen, D. van der Spoel and R. van Drunen
GROMACS: A message-passing parallel molecular dynamics implementation
Comp. Phys. Comm. 91 (1995) pp. 43-56
-------- -------- --- Thank You --- -------- --------

Main: Input args 14
 mdrun_gpu -quiet -s work/wudata_02.tpr -o work/wudata_02.trr -e work/wudata_02.edr -c work/wudata_02.gro -g work/wudata_02.log -x work/wudata_02.xtc

*-----------------------------------------------------------------------------------------*
GROMACS_FAH
Version:     1.24 (Fri Feb 6 11:30:37 PST 2009) ATI
Compilation: 11:30:43
Compiler:   
Build host:  amoeba

*-----------------------------------------------------------------------------------------*
There are 0 atoms for free energy perturbation
Input Parameters:
   integrator           = sd
   nsteps               = 10000000
   ns_type              = Simple
   nstlist              = 20000001
   ndelta               = 2
   bDomDecomp           = FALSE
   decomp_dir           = 0
   nstcomm              = -10
   nstlog               = 250000
   nstxout              = 0
   nstvout              = 0
   nstfout              = 0
   nstenergy            = 250000
   nstxtcout            = 250000
   init_t               = 5.12e+006
   delta_t              = 0.002
   xtcprec              = 1000
   nkx                  = 0
   nky                  = 0
   nkz                  = 0
   pme_order            = 4
   ewald_rtol           = 1e-005
   ewald_geometry       = 0
   epsilon_surface      = 0
   optimize_fft         = FALSE
   ePBC                 = no
   bUncStart            = TRUE
   bShakeSOR            = FALSE
   etc                  = No
   epc                  = No
   epctype              = Isotropic
   tau_p                = 1
   ref_p (3x3):
      ref_p[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   compress (3x3):
      compress[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   bSimAnn              = FALSE
   zero_temp_time       = 0
   rlist                = 100
   coulombtype          = Cut-off
   rcoulomb_switch      = 0
   rcoulomb             = 100
   vdwtype              = Cut-off
   rvdw_switch          = 0
   rvdw                 = 100
   epsilon_r            = 1
   DispCorr             = No
   fudgeQQ              = 0.8333
   free_energy          = no
   init_lambda          = 0
   sc_alpha             = 0
   sc_sigma             = 0.3
   delta_lambda         = 0
   disre_weighting      = Conservative
   disre_mixed          = FALSE
   dr_fc                = 1000
   dr_tau               = 0
   nstdisreout          = 100
   orires_fc            = 0
   orires_tau           = 0
   nstorireout          = 100
   em_stepsize          = 0.01
   em_tol               = 100
   niter                = 20
   fc_stepsize          = 0
   nstcgsteep           = 1000
   ConstAlg             = Shake
   shake_tol            = 0.0001
   lincs_order          = 4
   lincs_warnangle      = 30
   bd_temp              = 300
   bd_fric              = 0
   ld_seed              = 10122
   cos_accel            = 0
   userint1             = 0
   userint2             = 0
   userint3             = 0
   userint4             = 0
   userreal1            = 0
   userreal2            = 0
   userreal3            = 0
   userreal4            = 0
   solvent-model        = GBSA
   sasa                 = None
   nstsasa              = 1
   rsasa                = 0.55
   epsilon_dd           = 20
   nstsasaout           = 0
grpopts:
   nrdf:           2435
   ref_t:            370
   tau_t:        0.01099
   acc:              0           0           0
   nfreeze:           N           N           N
   energygrp_excl[  0]: 0
   efield-x:
      n = 0
   efield-xt:
      n = 0
   efield-y:
      n = 0
   efield-yt:
      n = 0
   efield-z:
      n = 0
   efield-zt:
      n = 0
CPU=  0, lastcg=  970, targetcg=  485, myshift=    0
nsb->shift =   1, nsb->bshift=  0
Neighbor Search Blocks
nsb->nodeid:         0
nsb->nnodes:      1
nsb->cgtotal:   971
nsb->natoms:    973
nsb->shift:       1
nsb->bshift:      0
Nodeid   index  homenr  cgload  workload
     0       0     973     971       971

Initializing gpu library w/ defaultStreamWidth=32
Test mode: disabled
Using runtime cal; initializing Brook
############

info_string:
CAL version: 1.4.283
CAL: 1 devices found
CAL: Using device 0
ASIC = RV770
Available board memory to CAL: 512MB
Available cached memory to CAL: 60MB
Available uncached memory to CAL: 1886MB
Engine clock: 0 (0.000000)
Natural vector width: 64
Recommended minimum thread count: 2560
Brook yield: 0
Don't flush kernels: 0
Pre flush command buffers: 0
Flush interval: 16

############
Number of atoms is odd, won't work
  natoms:973 ceiling=976
  atoms: 31x32
  unroll: i=4 j=4
  useNew=4
Done w/ Gpu initialization
Adjusted gmx variables to match streams: atoms=973 new size=992
Max number of bonds per atom is 4
Table routines are used for coulomb: FALSE
Table routines are used for vdw:     FALSE
Cut-off's:   NS: 100   Coulomb: 100   LJ: 100
Generated table with 500 data points for COUL.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ6.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ12.
Tabscale = 500 points/nm
Going to determine what solvent types we have.
There are 1 molecules, 971 charge groups and 973 atoms
There are 0 optimized solvent molecules on node 0
There are 0 optimized water molecules on node 0
setupForRestartSave=0 restartStep=0
forceCheck=0 maxSteps=0
GPU Warning: Output energies are not correct
selfTestRandomI=2524719
gpu_center_molecule: Coordinates pre centering 0
0 0 x[     0.07476      0.26675      2.01031]
0 1 x[    -0.01407      0.28109      1.96443]
0 2 x[     0.05462      0.28191      2.10811]
0 3 x[     0.14441      0.33084      1.97507]
0 4 x[     0.11844      0.13006      1.98375]

gpu_center_molecule: Coordinates post centering 0 com[-0.1913 -0.2169 -0.1616]
0 0 x[     0.26601      0.48362      2.17188]
0 1 x[     0.17719      0.49796      2.12599]
0 2 x[     0.24588      0.49879      2.26967]
0 3 x[     0.33567      0.54772      2.13664]
0 4 x[     0.30969      0.34693      2.14532]

Centered molecule
Center of mass motion removal mode is Angular
We have the following groups for center of mass motion removal:
  0:  rest, initial mass: 7011.73
Started mdrun on node 0 Wed Jun 17 20:15:22 2009
Initial temperature: 430.496 K
system name is Protein
Downloaded x/v/vold/mInvU
0 x[     0.26601      0.48362      2.17188] v[     0.16320     -0.27768     -0.17488] vold[     0.16320     -0.27768     -0.17488] m[    14.01000]
1 x[     0.17719      0.49796      2.12599] v[    -0.78389     -3.17911      0.56918] vold[    -0.78389     -3.17911      0.56918] m[     1.00800]
2 x[     0.24588      0.49879      2.26967] v[     1.25266     -2.71580      0.50526] vold[     1.25266     -2.71580      0.50526] m[     1.00800]
3 x[     0.33567      0.54772      2.13664] v[     0.45676     -1.31498     -1.52614] vold[     0.45676     -1.31498     -1.52614] m[     1.00800]
4 x[     0.30969      0.34693      2.14532] v[    -0.35444      0.36603      0.23805] vold[    -0.35444      0.36603      0.23805] m[    12.01000]

Using normal Coulomb/LJ & GBSA
Calculated RepFac: 12 minSuggestedThreads=2560
Setting RepFac: 12
Initialized strExcl XX (height x width) 248 x 992=246016
Initialized NB parameters (height x width)
  force :47x64
  excl  :248x992
  jstrm :16x16 sz=256 256
  nAtomsCeiling=976 iUnroll=4 repfac=12
Done gpuInitNBCalc

Using 4 new gpuSetExclusionsDone gpuSetExclusions

Using new gpuSetVDWParamsDone gpuSetVDWParams
Calling cpuSetObcParameters 1 1 2 1.000 78.300

gpuSetObcParameters:  No atoms=973 agb file length=4135
GpuObc::GBSAStreamsFactory stream values appear to be in order.
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadiiWithDielectricOffset cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcScaledAtomicRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcChain cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcIntermediateForce cpuBlockSz=15872 float=4 size=992 [31 32] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii2 cpuBlockSz=3968 float=4 size=992 [31 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_1 cpuBlockSz=95232 float=4 size=5952 [93 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_2 cpuBlockSz=95232 float=4 size=5952 [93 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_1 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_2 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_3 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_4 cpuBlockSz=48128 float=4 size=3008 [47 64] chunk=4  <float4>
 0. ObcAtomicRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 1. ObcAtomicRadiiWithDielectricOffset [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 2. ObcScaledAtomicRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 3. ObcBornRadii [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 4. ObcChain [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 5. ObcIntermediateForce [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=3968
 6. ObcBornRadii2 [32 31] Sz=31 Rep=1 Unroll=0 Chnk=1 Data=973 BlkSz=992
 7. ObcForceLoop1_1 [64 93] Sz=93 Rep=12 Unroll=2 Chnk=12 Data=5952 BlkSz=23808
 8. ObcForceLoop1_2 [64 93] Sz=93 Rep=12 Unroll=2 Chnk=12 Data=5952 BlkSz=23808
 9. ObcForceLoop2_1 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
10. ObcForceLoop2_2 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
11. ObcForceLoop2_3 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
12. ObcForceLoop2_4 [64 47] Sz=47 Rep=12 Unroll=4 Chnk=12 Data=3008 BlkSz=12032
gpuSetObcParameters:  includeAce=1 soluteDiel=1.000 solventDiel=78.300.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii read 55 lines from agb parameter file.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii
   agb parameter file line=<# This file was created by edgar to supply Macromodel radii for AMBER atomypes> is being skipped.
   agb parameter file line=<#> is being skipped.
   agb parameter file line=<@   title "agbnp parameters"> is being skipped.
   agb parameter file line=<@   xaxis label "OPLS atomtype"> is being skipped.
   agb parameter file line=<@   yaxis label "parameter"> is being skipped.
   agb parameter file line=<@ s0 legend "radius"> is being skipped.
   agb parameter file line=<@ s1 legend "igamma"> is being skipped.
   agb parameter file line=<@ s2 legend "ialpha"> is being skipped.
   agb parameter file line=<@ s3 legend "idelta"> is being skipped.
   agb parameter file line=<@ s4 legend "sgamma"> is being skipped.
   agb parameter file line=<@ s5 legend "salpha"> is being skipped.
   agb parameter file line=<@ s6 legend "sdelta"> is being skipped.
   agb parameter file line=<@ s6 legend "ai"@ s7 legend "GBdistcorr"> is being skipped.
   agb parameter file line=<# the following values are from the default behaviour.> is being skipped.
gpuSetObcParameters:  obtained atomic radii from tar file
gpuSetObcParameters:  writing parameters to board
gpuSetObcParameters:  done OBC setup
do_md: OBC setup completed.
In gpuBondedParams_Init
In gpuBondedParams_Init Start param setup
In gpuBondedParams_Init Post param setup
nbondeds=2962 stream dimensions: [62 48] 2976
Writing params to board -- done
gpuBondedParams_Init: parameter setup complete, nbonds=2962.
gpuBondedParams_Init: calling LoadInvMaps
InvMap completed malloc natoms=973 nbondeds=2962 AtomStrSize=992 992 [31 32]
InvMap completed gpuCalcInvMap counts=4 [31 32] 992
InvMap I-stream nimaps=3 BONDED_INVMAP_ISTREAMS=9
Writing I-stream to board  -- done
I-Map:
InvMap J-stream njmaps=5 BONDED_INVMAP_JSTREAMS=6
Writing J-stream to board  -- done
J-Map
InvMap K-stream nkmaps=4 BONDED_INVMAP_KSTREAMS=6
Writing K-stream to board  -- done
InvMap calling gpuCalcInvMap(3, 4)
K-map

InvMap L-stream nlmaps=2 BONDED_INVMAP_LSTREAMS=9
Writing L-stream to board
L-map
  -- done
Bondeds use 3 inverse maps for i
Bondeds use 5 inverse maps for j
Bondeds use 4 inverse maps for k
Bondeds use 2 inverse maps for l
gpuBondedParams_Init: LoadInvMaps completed nbondeds=2962 bp->size=2976
gpuBondedParams_Init w i=2962 j=4 4i+j=11847
Writing atom indices to board -- done
Freeing buffers
gpuBondedParams_Init creating force streams
Done gpuBondedParams_Init
Done gpuSetBondedParams
bFFscan=0 -- if set, then original coords copied
bTCR=0 bGlas=0
generateTestHarnessFiles=0
COM removal: frequency==-10 Remove COM on Cpu=0
Sd parameters: ngtc=1 tau=1.099e-002 deltaT=2.000e-003 T=3.700e+002
StopCM every -10 steps -- requested ANGULAR
Neighbor list search=20000001 --  set to 0.
Merge Obc & CDLJ force calculations=0


The wudata log seems to cut off abruptly at the end. Indicates an error, I guess.

Here are the logs for a new WU that was downloaded by the client in Linux:
FahLog.txt:
Code: Select all


--- Opening Log file [June 17 17:57:52 UTC]





# Windows GPU Console Edition #################################################

###############################################################################



                       Folding@Home Client Version 6.23



                          http://folding.stanford.edu



###############################################################################

###############################################################################



Launch directory: C:\Program Files\folding-gpu

Executable: C:\Program Files\folding-gpu\fah6-gpu.exe

Arguments: -oneunit -verbosity 9 -forcegpu ati_r700



[17:57:52] - Ask before connecting: Yes

[17:57:52] - User name: Bjornar (Team 31574)

[17:57:52] - User ID: 2EC0534E5D7AA47F

[17:57:52] - Machine ID: 2

[17:57:52]

[17:57:52] Work directory not found. Creating...

[17:57:52] Could not open work queue, generating new queue...

[17:57:52] - Preparing to get new work unit...

[17:57:52] - Presenting message box asking to network.

[17:57:52] - Autosending finished units... [June 17 17:57:52 UTC]

[17:57:52] Trying to send all finished work units

[17:57:52] + No unsent completed units remaining.

[17:57:52] - Autosend completed

[17:57:56] + Attempting to get work packet

[17:57:56] - Will indicate memory of 3900 MB

[17:57:56] - Detect CPU. Vendor: AuthenticAMD, Family: 15, Model: 2, Stepping: 3

[17:57:56] - Connecting to assignment server

[17:57:56] Connecting to http://assign-GPU.stanford.edu:8080/

[17:57:57] Posted data.

[17:57:57] Initial: 40AB; - Successful: assigned to (171.64.65.102).

[17:57:57] + News From Folding@Home: Welcome to Folding@Home

[17:57:57] Loaded queue successfully.

[17:57:57] Connecting to http://171.64.65.102:8080/

[17:57:57] Posted data.

[17:57:57] Initial: 0000; - Receiving payload (expected size: 99079)

[17:57:59] - Downloaded at ~48 kB/s

[17:57:59] - Averaged speed for that direction ~48 kB/s

[17:57:59] + Received work.

[17:57:59] + Connections closed: You may now disconnect

[17:57:59]

[17:57:59] + Processing work unit

[17:57:59] Core required: FahCore_11.exe

[17:57:59] Core found.

[17:57:59] Working on queue slot 01 [June 17 17:57:59 UTC]

[17:57:59] + Working ...

[17:57:59] - Calling '.\FahCore_11.exe -dir work/ -suffix 01 -checkpoint 15 -verbose -lifeline 8 -version 623'


[17:57:59]

[17:57:59] *------------------------------*

[17:57:59] Folding@Home GPU Core - Beta

[17:57:59] Version 1.24 (Mon Feb 9 11:00:12 PST 2009)

[17:57:59]

[17:57:59] Compiler  : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86

[17:57:59] Build host: amoeba

[17:57:59] Board Type: AMD

[17:57:59] Core      :

[17:57:59] Preparing to commence simulation

[17:57:59] - Looking at optimizations...

[17:57:59] - Created dyn

[17:57:59] - Files status OK

[17:57:59] - Expanded 98567 -> 492188 (decompressed 499.3 percent)

[17:57:59] Called DecompressByteArray: compressed_data_size=98567 data_size=492188, decompressed_data_size=492188 diff=0

[17:57:59] - Digital signature verified

[17:57:59]

[17:57:59] Project: 5732 (Run 0, Clone 139, Gen 38)

[17:57:59]

[17:57:59] Assembly optimizations on if available.

[17:57:59] Entering M.D.

[17:58:05] Tpr hash work/wudata_01.tpr:  530599094 2200634762 728672176 432296281 1059477897

[17:58:05] Working on Protein

[17:58:06] Client config found, loading data.

[17:58:06] Starting GUI Server

[17:58:06] Run: exception thrown during GuardedRun

[17:58:06] Run: exception thrown in GuardedRun -- Gromacs cannot continue further.

[17:58:06] Going to send back what have done -- stepsTotalG=10000000

[17:58:06] Work fraction=0.0000 steps=10000000.

[17:58:10] logfile size=14319 infoLength=14319 edr=0 trr=23

[17:58:10] - Writing 14855 bytes of core data to disk...

[17:58:10] Done: 14343 -> 4866 (compressed to 33.9 percent)

[17:58:10]   ... Done.

[17:58:10]

[17:58:10] Folding@home Core Shutdown: EARLY_UNIT_END

[17:58:13] CoreStatus = 72 (114)

[17:58:13] Sending work to server

[17:58:13] Project: 5732 (Run 0, Clone 139, Gen 38)

[17:58:13] - Read packet limit of 540015616... Set to 524286976.





[17:58:13] + Attempting to send results [June 17 17:58:13 UTC]

[17:58:13] - Reading file work/wuresults_01.dat from core

[17:58:13]   (Read 5378 bytes from disk)

[17:58:13] - Presenting message box asking to network.


wudata_01.log:
Code: Select all
Log file opened: nodeid 0, nnodes = 1, host = unknown, process = 0
Gromacs code copyright(c) 1991-2002
University of Groningen, The Netherlands
GPU library and other code (c) Stanford University 2006
This program is distributed under a special license to Folding@Home
See http://folding.stanford.edu/gromacs.html for license details

++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
E. Lindahl and B. Hess and D. van der Spoel
GROMACS 3.0: A package for molecular simulation and trajectory analysis
J. Mol. Mod. 7 (2001) pp. 306-317
-------- -------- --- Thank You --- -------- --------


++++++++ PLEASE CITE THE FOLLOWING REFERENCE ++++++++
H. J. C. Berendsen, D. van der Spoel and R. van Drunen
GROMACS: A message-passing parallel molecular dynamics implementation
Comp. Phys. Comm. 91 (1995) pp. 43-56
-------- -------- --- Thank You --- -------- --------

Main: Input args 14
 mdrun_gpu -quiet -s work/wudata_01.tpr -o work/wudata_01.trr -e work/wudata_01.edr -c work/wudata_01.gro -g work/wudata_01.log -x work/wudata_01.xtc

*-----------------------------------------------------------------------------------------*
GROMACS_FAH
Version:     1.24 (Fri Feb 6 11:30:37 PST 2009) ATI
Compilation: 11:30:43
Compiler:   
Build host:  amoeba

*-----------------------------------------------------------------------------------------*
There are 0 atoms for free energy perturbation
Input Parameters:
   integrator           = sd
   nsteps               = 10000000
   ns_type              = Simple
   nstlist              = 20000001
   ndelta               = 2
   bDomDecomp           = FALSE
   decomp_dir           = 0
   nstcomm              = -10
   nstlog               = 250000
   nstxout              = 0
   nstvout              = 0
   nstfout              = 0
   nstenergy            = 250000
   nstxtcout            = 250000
   init_t               = 760000
   delta_t              = 0.002
   xtcprec              = 1000
   nkx                  = 0
   nky                  = 0
   nkz                  = 0
   pme_order            = 4
   ewald_rtol           = 1e-005
   ewald_geometry       = 0
   epsilon_surface      = 0
   optimize_fft         = FALSE
   ePBC                 = no
   bUncStart            = TRUE
   bShakeSOR            = FALSE
   etc                  = No
   epc                  = No
   epctype              = Isotropic
   tau_p                = 1
   ref_p (3x3):
      ref_p[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      ref_p[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   compress (3x3):
      compress[    0]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    1]={0.00000e+000, 0.00000e+000, 0.00000e+000}
      compress[    2]={0.00000e+000, 0.00000e+000, 0.00000e+000}
   bSimAnn              = FALSE
   zero_temp_time       = 0
   rlist                = 100
   coulombtype          = Cut-off
   rcoulomb_switch      = 0
   rcoulomb             = 100
   vdwtype              = Cut-off
   rvdw_switch          = 0
   rvdw                 = 100
   epsilon_r            = 1
   DispCorr             = No
   fudgeQQ              = 0.8333
   free_energy          = no
   init_lambda          = 0
   sc_alpha             = 0
   sc_sigma             = 0.3
   delta_lambda         = 0
   disre_weighting      = Conservative
   disre_mixed          = FALSE
   dr_fc                = 1000
   dr_tau               = 0
   nstdisreout          = 100
   orires_fc            = 0
   orires_tau           = 0
   nstorireout          = 100
   em_stepsize          = 0.01
   em_tol               = 100
   niter                = 20
   fc_stepsize          = 0
   nstcgsteep           = 1000
   ConstAlg             = Shake
   shake_tol            = 0.0001
   lincs_order          = 4
   lincs_warnangle      = 30
   bd_temp              = 300
   bd_fric              = 0
   ld_seed              = 6009
   cos_accel            = 0
   userint1             = 0
   userint2             = 0
   userint3             = 0
   userint4             = 0
   userreal1            = 0
   userreal2            = 0
   userreal3            = 0
   userreal4            = 0
   solvent-model        = GBSA
   sasa                 = None
   nstsasa              = 1
   rsasa                = 0.55
   epsilon_dd           = 20
   nstsasaout           = 0
grpopts:
   nrdf:           3476
   ref_t:            300
   tau_t:        0.01099
   acc:              0           0           0
   nfreeze:           N           N           N
   energygrp_excl[  0]: 0
   efield-x:
      n = 0
   efield-xt:
      n = 0
   efield-y:
      n = 0
   efield-yt:
      n = 0
   efield-z:
      n = 0
   efield-zt:
      n = 0
CPU=  0, lastcg= 1389, targetcg=  694, myshift=    0
nsb->shift =   1, nsb->bshift=  0
Neighbor Search Blocks
nsb->nodeid:         0
nsb->nnodes:      1
nsb->cgtotal:  1390
nsb->natoms:   1392
nsb->shift:       1
nsb->bshift:      0
Nodeid   index  homenr  cgload  workload
     0       0    1392    1390      1390

Initializing gpu library w/ defaultStreamWidth=32
Test mode: disabled
Using runtime cal; initializing Brook
############

info_string:
CAL version: 1.4.283
CAL: 1 devices found
CAL: Using device 0
ASIC = RV770
Available board memory to CAL: 512MB
Available cached memory to CAL: 60MB
Available uncached memory to CAL: 1886MB
Engine clock: 0 (0.000000)
Natural vector width: 64
Recommended minimum thread count: 2560
Brook yield: 0
Don't flush kernels: 0
Pre flush command buffers: 0
Flush interval: 16

############
  natoms:1392 ceiling=1392
  atoms: 44x32
  unroll: i=4 j=4
  useNew=4
Done w/ Gpu initialization
Adjusted gmx variables to match streams: atoms=1392 new size=1408
Max number of bonds per atom is 4
Table routines are used for coulomb: FALSE
Table routines are used for vdw:     FALSE
Cut-off's:   NS: 100   Coulomb: 100   LJ: 100
Generated table with 500 data points for COUL.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ6.
Tabscale = 500 points/nm
Generated table with 500 data points for LJ12.
Tabscale = 500 points/nm
Going to determine what solvent types we have.
There are 1 molecules, 1390 charge groups and 1392 atoms
There are 0 optimized solvent molecules on node 0
There are 0 optimized water molecules on node 0
setupForRestartSave=0 restartStep=0
forceCheck=0 maxSteps=0
GPU Warning: Output energies are not correct
selfTestRandomI=4000061
gpu_center_molecule: Coordinates pre centering 0
0 0 x[     0.64678      1.26401     -0.10210]
0 1 x[     0.59737      1.31632     -0.17297]
0 2 x[     0.56815      1.21369     -0.06355]
0 3 x[     0.68401      1.32710     -0.03256]
0 4 x[     0.75398      1.17219     -0.15641]

gpu_center_molecule: Coordinates post centering 0 com[-0.0579 -0.0382 0.0256]
0 0 x[     0.70472      1.30221     -0.12770]
0 1 x[     0.65531      1.35452     -0.19857]
0 2 x[     0.62608      1.25189     -0.08915]
0 3 x[     0.74194      1.36530     -0.05816]
0 4 x[     0.81191      1.21039     -0.18202]

Centered molecule
Center of mass motion removal mode is Angular
We have the following groups for center of mass motion removal:
  0:  rest, initial mass: 9913.18
Started mdrun on node 0 Wed Jun 17 19:58:05 2009
Initial temperature: 356.863 K
system name is Protein
Downloaded x/v/vold/mInvU
0 x[     0.70472      1.30221     -0.12770] v[    -0.19417     -0.68968     -0.26841] vold[    -0.19417     -0.68968     -0.26841] m[    14.01000]
1 x[     0.65531      1.35452     -0.19857] v[    -2.80945      2.50558      3.44882] vold[    -2.80945      2.50558      3.44882] m[     1.00800]
2 x[     0.62608      1.25189     -0.08915] v[    -1.21496     -0.63496     -2.16548] vold[    -1.21496     -0.63496     -2.16548] m[     1.00800]
3 x[     0.74194      1.36530     -0.05816] v[    -1.06821      0.50974     -0.84490] vold[    -1.06821      0.50974     -0.84490] m[     1.00800]
4 x[     0.81191      1.21039     -0.18202] v[     0.19799      0.43578     -0.03345] vold[     0.19799      0.43578     -0.03345] m[    12.01000]

Using normal Coulomb/LJ & GBSA
Calculated RepFac: 8 minSuggestedThreads=2560
Setting RepFac: 8
Initialized strExcl XX (height x width) 352 x 1408=495616
Initialized NB parameters (height x width)
  force :44x64
  excl  :352x1408
  jstrm :22x16 sz=352 352
  nAtomsCeiling=1392 iUnroll=4 repfac=8
Done gpuInitNBCalc

Using 4 new gpuSetExclusionsDone gpuSetExclusions

Using new gpuSetVDWParamsDone gpuSetVDWParams
Calling cpuSetObcParameters 1 1 2 1.000 78.300

gpuSetObcParameters:  No atoms=1392 agb file length=4135
GpuObc::GBSAStreamsFactory stream values appear to be in order.
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadii cpuBlockSz=5632 float=4 size=1408 [44 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcAtomicRadiiWithDielectricOffset cpuBlockSz=5632 float=4 size=1408 [44 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcScaledAtomicRadii cpuBlockSz=5632 float=4 size=1408 [44 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii cpuBlockSz=5632 float=4 size=1408 [44 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcChain cpuBlockSz=5632 float=4 size=1408 [44 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcIntermediateForce cpuBlockSz=22528 float=4 size=1408 [44 32] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcBornRadii2 cpuBlockSz=5632 float=4 size=1408 [44 32] chunk=1  <float>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_1 cpuBlockSz=90112 float=4 size=5632 [88 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop1_2 cpuBlockSz=90112 float=4 size=5632 [88 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_1 cpuBlockSz=45056 float=4 size=2816 [44 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_2 cpuBlockSz=45056 float=4 size=2816 [44 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_3 cpuBlockSz=45056 float=4 size=2816 [44 64] chunk=4  <float4>
BrookStreamWrapper::allocateMemory Creating stream ObcForceLoop2_4 cpuBlockSz=45056 float=4 size=2816 [44 64] chunk=4  <float4>
 0. ObcAtomicRadii [32 44] Sz=44 Rep=1 Unroll=0 Chnk=1 Data=1392 BlkSz=1408
 1. ObcAtomicRadiiWithDielectricOffset [32 44] Sz=44 Rep=1 Unroll=0 Chnk=1 Data=1392 BlkSz=1408
 2. ObcScaledAtomicRadii [32 44] Sz=44 Rep=1 Unroll=0 Chnk=1 Data=1392 BlkSz=1408
 3. ObcBornRadii [32 44] Sz=44 Rep=1 Unroll=0 Chnk=1 Data=1392 BlkSz=1408
 4. ObcChain [32 44] Sz=44 Rep=1 Unroll=0 Chnk=1 Data=1392 BlkSz=1408
 5. ObcIntermediateForce [32 44] Sz=44 Rep=1 Unroll=0 Chnk=1 Data=1392 BlkSz=5632
 6. ObcBornRadii2 [32 44] Sz=44 Rep=1 Unroll=0 Chnk=1 Data=1392 BlkSz=1408
 7. ObcForceLoop1_1 [64 88] Sz=88 Rep=8 Unroll=2 Chnk=8 Data=5632 BlkSz=22528
 8. ObcForceLoop1_2 [64 88] Sz=88 Rep=8 Unroll=2 Chnk=8 Data=5632 BlkSz=22528
 9. ObcForceLoop2_1 [64 44] Sz=44 Rep=8 Unroll=4 Chnk=8 Data=2816 BlkSz=11264
10. ObcForceLoop2_2 [64 44] Sz=44 Rep=8 Unroll=4 Chnk=8 Data=2816 BlkSz=11264
11. ObcForceLoop2_3 [64 44] Sz=44 Rep=8 Unroll=4 Chnk=8 Data=2816 BlkSz=11264
12. ObcForceLoop2_4 [64 44] Sz=44 Rep=8 Unroll=4 Chnk=8 Data=2816 BlkSz=11264
gpuSetObcParameters:  includeAce=1 soluteDiel=1.000 solventDiel=78.300.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii read 55 lines from agb parameter file.
SimTKOpenMMGromacsUtilities::getMacroModelAtomicRadii
   agb parameter file line=<# This file was created by edgar to supply Macromodel radii for AMBER atomypes> is being skipped.
   agb parameter file line=<#> is being skipped.
   agb parameter file line=<@   title "agbnp parameters"> is being skipped.
   agb parameter file line=<@   xaxis label "OPLS atomtype"> is being skipped.
   agb parameter file line=<@   yaxis label "parameter"> is being skipped.
   agb parameter file line=<@ s0 legend "radius"> is being skipped.
   agb parameter file line=<@ s1 legend "igamma"> is being skipped.
   agb parameter file line=<@ s2 legend "ialpha"> is being skipped.
   agb parameter file line=<@ s3 legend "idelta"> is being skipped.
   agb parameter file line=<@ s4 legend "sgamma"> is being skipped.
   agb parameter file line=<@ s5 legend "salpha"> is being skipped.
   agb parameter file line=<@ s6 legend "sdelta"> is being skipped.
   agb parameter file line=<@ s6 legend "ai"@ s7 legend "GBdistcorr"> is being skipped.
   agb parameter file line=<# the following values are from the default behaviour.> is being skipped.
gpuSetObcParameters:  obtained atomic radii from tar file
gpuSetObcParameters:  writing parameters to board
gpuSetObcParameters:  done OBC setup
do_md: OBC setup completed.
In gpuBondedParams_Init
In gpuBondedParams_Init Start param setup
In gpuBondedParams_Init Post param setup
nbondeds=4286 stream dimensions: [67 64] 4288
Writing params to board -- done
gpuBondedParams_Init: parameter setup complete, nbonds=4286.
gpuBondedParams_Init: calling LoadInvMaps
InvMap completed malloc natoms=1392 nbondeds=4286 AtomStrSize=1408 1408 [44 32]
InvMap completed gpuCalcInvMap counts=4 [44 32] 1408
InvMap I-stream nimaps=3 BONDED_INVMAP_ISTREAMS=9
Writing I-stream to board  -- done
I-Map:
InvMap J-stream njmaps=5 BONDED_INVMAP_JSTREAMS=6
Writing J-stream to board  -- done
J-Map
InvMap K-stream nkmaps=4 BONDED_INVMAP_KSTREAMS=6
Writing K-stream to board  -- done
InvMap calling gpuCalcInvMap(3, 4)
K-map

InvMap L-stream nlmaps=2 BONDED_INVMAP_LSTREAMS=9
Writing L-stream to board
L-map
  -- done
Bondeds use 3 inverse maps for i
Bondeds use 5 inverse maps for j
Bondeds use 4 inverse maps for k
Bondeds use 2 inverse maps for l
gpuBondedParams_Init: LoadInvMaps completed nbondeds=4286 bp->size=4288
gpuBondedParams_Init w i=4286 j=4 4i+j=17143
Writing atom indices to board -- done
Freeing buffers
gpuBondedParams_Init creating force streams
Done gpuBondedParams_Init
Done gpuSetBondedParams
bFFscan=0 -- if set, then original coords copied
bTCR=0 bGlas=0
generateTestHarnessFiles=0
COM removal: frequency==-10 Remove COM on Cpu=0
Sd parameters: ngtc=1 tau=1.099e-002 deltaT=2.000e-003 T=3.000e+002
StopCM every -10 steps -- requested ANGULAR
Neighbor list search=20000001 --  set to 0.
Merge Obc & CDLJ force calculations=0



I have the output produced in the terminal when running WINEDEBUG=warn+all wine fah6-gpu.exe, but it is too long to post.

One thing I have noticed is that the engine clock is always reported incorrectly:
Code: Select all
Engine clock: 0 (0.000000)
Last edited by Anachron on Wed Jun 17, 2009 7:04 pm, edited 1 time in total.
Anachron
 
Posts: 50
Joined: Fri Mar 14, 2008 12:10 pm

Next

Return to unOfficial Linux GPU (WINE wrapper) (3rd party support)

Who is online

Users browsing this forum: No registered users and 1 guest

cron