Hello

Coming to acquire a gtx 470, I wanted to know that your units will be compatible with?

For all your units 0.24 go astray ...

I understand that the thought 0.03 would be compatible but I do not get ....

Sincerely,

---

Team Alliance francophone, boinc: 7.0.18

GA-P55-UD5, i7 860, Win 7 64 bits, 8g DDR3, GTX 470

always the same, even after I detached and attached ....

Someone just answer me, thank you ...

---

Team Alliance francophone, boinc: 7.0.18

GA-P55-UD5, i7 860, Win 7 64 bits, 8g DDR3, GTX 470

Hi,
here http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1841&nowrap=true#41418 is some info for you.

Alexander

FYI, I believe this is the only change necessary to get the app working, at least minimally, on Fermi:

--- evaluation_gpu6.cu.old	2010-01-24 18:53:01.000000000 -0700
+++ evaluation_gpu6.cu	2010-11-13 18:20:09.000000000 -0700
@@ -156,7 +156,7 @@
       cudaDeviceProp deviceProp;
       cutilSafeCall(cudaGetDeviceProperties(&deviceProp, idx));
       fprintf(stderr, "Found a %s\n", deviceProp.name);
-      if (deviceProp.major == 1 && deviceProp.minor == 3)
+      if ((deviceProp.major >= 2 && deviceProp.major < 9999) || (deviceProp.major == 1 && deviceProp.minor == 3))
 	{
 	  eligable_devices[eligable_device_idx++] = idx;
 	  fprintf(stderr, "Device can be used it has compute capability 1.3 support\n");
@@ -217,7 +217,7 @@
   if (device_arg != -1) {
     cudaDeviceProp deviceProp;
     cutilSafeCall(cudaGetDeviceProperties(&deviceProp, device_arg));
-    if (deviceProp.major == 1 && deviceProp.minor == 3) {
+    if ((deviceProp.major >= 2 && deviceProp.major < 9999) || (deviceProp.major == 1 && deviceProp.minor == 3)) {
       cutilSafeCall(cudaSetDevice(device_arg));
       fprintf(stderr, "The device %s specified on the command line can be used\n", deviceProp.name);
     } else {

However, I haven't been able to build libcutil.so to build this yet.

Of course, you could just download the optimized app for "Fermi" cards. The url for it can be found in this thread.

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1297

Of course, you could just download the optimized app for "Fermi" cards. The url for it can be found in this thread.

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1297

I'm trying that App on 1 of my GTX 460's and I keep getting >>> 11/18/2010 4:59:37 AM | Milkyway@home | Message from Milkyway@home: Your app_info.xml file doesn't have a version of MilkyWay@Home N-Body Simulation. <<<

What do I have to do now to get the Wu's to run ... Thanks

EDIT: Seems to be working now, around 10 Mins, slow but it's working anyway & I didn't expect any speed out of it when I bought it ...

Of course, you could just download the optimized app for "Fermi" cards. The url for it can be found in this thread.

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=1297

I'm trying that App on 1 of my GTX 460's and I keep getting >>> 11/18/2010 4:59:37 AM | Milkyway@home | Message from Milkyway@home: Your app_info.xml file doesn't have a version of MilkyWay@Home N-Body Simulation. <<<

What do I have to do now to get the Wu's to run ... Thanks

EDIT: Seems to be working now, around 10 Mins, slow but it's working anyway & I didn't expect any speed out of it when I bought it ...

Steve, the problem with running with an .XML file for the GPU is that an app for the CPU isn't specified (normally).

You were getting that message because there wasn't any GPU WUs available at the time of the download request, just the pesky N-Body WUs were available. I get that message quite a bit, also. I do not run Milkyway on my CPU, just the GPUs.

-Mike

You were getting that message because there wasn't any GPU WUs available at the time of the download request, just the pesky N-Body WUs were available. I get that message quite a bit, also. I do not run Milkyway on my CPU, just the GPUs.

-Mike

Same here Mike, I don't run any CPU Tasks at the GPU Projects ...

---

STE\/E