We have tested locally and this version works for us.
Still 30% faster over current application on G200 cards.

gdf
PS: if all works Fermi application coming out in the afternoon.

We have tested locally and this version works for us.
Still 30% faster over current application on G200 cards.

gdf
PS: if all works Fermi application coming out in the afternoon.

Picked up 3 tonight. 2 out of 3 crashed after about 15 seconds. Third is still going.

SWAN : FATAL : Failure executing kernel [mshake_position_kernel_1] [2] [10,1,1][64,1,1]
Assertion failed: 0, file swanlib_nv.cpp, line 281

1st wu
2nd wu

Both gave a windows pop-up saying task has stopped responding. BOINC 6.10.43, Win7 x64, Driver 196.21

---

BOINC blog

Tried to crunch two tasks:

1. Host ID: 31329 - WinXP x64 GTX260 65nm, driver 19713, one of unusable GTX260 cards for GPUGRID. Underclocked according to basic settings by wiki - the card is factory OCed. The task erroed out after 2,972s. Very low CPU load (max. 8% of one core used).

<core_client_version>6.10.36</core_client_version>
<![CDATA[
<message>
The system cannot find the path specified. (0x3) - exit code 3 (0x3)
</message>
<stderr_txt>
# Using device 0
# There is 1 device supporting CUDA
# Device 0: "GeForce GTX 260"
# Clock rate: 1.24 GHz
# Total amount of global memory:                 939196416 bytes
# Number of multiprocessors:                     27
# Number of cores:                               216
MDIO ERROR: cannot open file "restart.coor"
SWAN : FATAL : Failure executing kernel sync [reduce_kernel4] [702]
Assertion failed: 0, file swanlib_nv.cpp, line 121

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.

</stderr_txt>
]]>

2. Host ID 38345 - Win XP x86, 9600 GT, driver 19562, a little OCed. No problem when crunchig 6.03 tasks. The task errored out after 230s.

<core_client_version>6.10.18</core_client_version>
<![CDATA[
<stderr_txt>
# Using device 0
# There is 1 device supporting CUDA
# Device 0: "GeForce 9600 GT"
# Clock rate: 1.62 GHz
# Total amount of global memory:                 536543232 bytes
# Number of multiprocessors:                     8
# Number of cores:                               64
MDIO ERROR: cannot open file "restart.coor"
# Time per step (avg over 5000 steps): 	45.297 ms
# Approximate elapsed time for entire WU:  	226.484 s
called boinc_finish

</stderr_txt>
]]>

It did not error out. It finished successfully!
The application seems to be working fine.

g

Tried to crunch two tasks:

1. Host ID: 31329 - WinXP x64 GTX260 65nm, driver 19713, one of unusable GTX260 cards for GPUGRID. Underclocked according to basic settings by wiki - the card is factory OCed. The task erroed out after 2,972s. Very low CPU load (max. 8% of one core used).

<core_client_version>6.10.36</core_client_version>
<![CDATA[
<message>
The system cannot find the path specified. (0x3) - exit code 3 (0x3)
</message>
<stderr_txt>
# Using device 0
# There is 1 device supporting CUDA
# Device 0: "GeForce GTX 260"
# Clock rate: 1.24 GHz
# Total amount of global memory:                 939196416 bytes
# Number of multiprocessors:                     27
# Number of cores:                               216
MDIO ERROR: cannot open file "restart.coor"
SWAN : FATAL : Failure executing kernel sync [reduce_kernel4] [702]
Assertion failed: 0, file swanlib_nv.cpp, line 121

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.

</stderr_txt>
]]>

2. Host ID 38345 - Win XP x86, 9600 GT, driver 19562, a little OCed. No problem when crunchig 6.03 tasks. The task errored out after 230s.

<core_client_version>6.10.18</core_client_version>
<![CDATA[
<stderr_txt>
# Using device 0
# There is 1 device supporting CUDA
# Device 0: "GeForce 9600 GT"
# Clock rate: 1.62 GHz
# Total amount of global memory:                 536543232 bytes
# Number of multiprocessors:                     8
# Number of cores:                               64
MDIO ERROR: cannot open file "restart.coor"
# Time per step (avg over 5000 steps): 	45.297 ms
# Approximate elapsed time for entire WU:  	226.484 s
called boinc_finish

</stderr_txt>
]]>

You are right GDF, claimed 18.73, granted 28.09 for 280s elapsed time. Very high credit for 9600 GT, nice.
Sorry for my thoughtlessness, wall-clock time 280s is unexpected for succefully finished task.

Picked up 3 tonight. 2 out of 3 crashed after about 15 seconds. Third is still going.

Third one completed successfully. Link is here

---

BOINC blog

For what it's worth,
6h and 3h into 6.22 runs on two seperate systems that use GT240s.
Estimated run time 11h and 13h roughly.

1st beta 6.22 WU completed. ^^

http://www.gpugrid.net/result.php?resultid=2073498

Approximately 3000 seconds faster on my GTX 275, a performance increase of 16%

Good work!!

---

cuda30 version is also out now. This should work on Fermi.

gdf

Success.

Two completed results taking approximately 19 & 20 minutes each.

Windows 7, i7 920, 2X GTX260 (55nm, 216 core).

Running 8 cpu tasks simultaneously.

2076209 1307386 31 Mar 2010 0:37:59 UTC 31 Mar 2010 1:09:04 UTC Completed and validated 1,221.97 207.51 187.28 280.92 ACEMD beta version v6.22 (cuda30)

2076073 1307228 31 Mar 2010 0:20:28 UTC 31 Mar 2010 1:07:05 UTC Completed and validated 1,144.30 198.67 187.28 280.92 ACEMD beta version v6.22 (cuda30)

Using about 15% of one cpu thread (about 2% of total cpu)per WU.

This is the best I've seen yet.

6.22 cuda 3.0 erroed out (both GTX260 and 9600GT).
ID 31329

<core_client_version>6.10.18</core_client_version>
<![CDATA[
<message>
Incorrect function. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using device 0
# There is 1 device supporting CUDA
# Device 0: "GeForce GTX 260"
# Clock rate: 1.24 GHz
# Total amount of global memory:                 939196416 bytes
# Number of multiprocessors:                     27
# Number of cores:                               216
MDIO ERROR: cannot open file "restart.coor"
</stderr_txt>
]]>

ID 38345

<core_client_version>6.10.18</core_client_version>
<![CDATA[
<message>
 - exit code -40 (0xffffffd8)
</message>
<stderr_txt>
# Using device 0
# There is 1 device supporting CUDA
# Device 0: "GeForce 9600 GT"
# Clock rate: 1.62 GHz
# Total amount of global memory:                 536543232 bytes
# Number of multiprocessors:                     8
# Number of cores:                               64
SWAN : Module load result [.fastfill.cu.] [200]
SWAN: FATAL : Module load failed
</stderr_txt>
]]>

nenym wrote:

6.22 cuda 3.0 erroed out (both GTX260 and 9600GT).

6.22 cuda 3.0 works fine on GTX260 now, wow. The first task finished on my weird GTX260 after updating from 181.xx drivers.

After updating 9600GT drivers to WHQL 197.13 6.22 cuda 3.0 runs fine.

What's the oldest driver required for the CUDA 3.0 app?

MrS

---

Scanning for our furry friends since Jan 2002

It gets shipped only if you have 197.13.

gdf

6.22 cuda and cuda3.0 works fine although I'd like to test cuda3.0 on a full blown WU (not the shorties).

The 6.22 cuda doesnt seem to give any performance boost on GTX 295 (2003572)

---

It gets shipped only if you have 197.13.

gdf

Cuda toolkit 3 was supported from 195.62 drivers onwards. You could give the cuda 3 DLL's to machines with this or later.

---

BOINC blog

Just finished my 1 WU
http://www.gpugrid.net/result.php?resultid=2082769
Really surprised to see that the system wasn't too lagy when I ran medium graphics software while crunching.
So far it seams to be on par with Collatz, thinking of system responsiveness, and much better than SETI or Einstein.
Really happy I could put the FX 580 Quadro to work, and doing currently only ACEMD beta :)

It gets shipped only if you have 197.13.

gdf

This is not correct. My computer downloaded 2 of these units and I am running version 195.62. They both crashed after about 2 seconds.

It gets shipped only if you have 197.13.

gdf

This is not correct. My computer downloaded 2 of these units and I am running version 195.62. They both crashed after about 2 seconds.

If you read a couple of posts further up I mentioned that according to to nvidia 195.62 supports cuda toolkit 3. Its likely the guys changed it so we get the the cuda 3 DLL's and beta app.

It may be that the latest cuda 3 toolkit doesn't work properly with 195.62, but it may with later drivers. I have 196 drivers, but haven't picked up any beta work since the cuda 3 version came out. I will see if I can get one.

Its also possible its a bug with the beta app and nothing to do with the driver version.

---

BOINC blog