Any update on this? Has anyone had any success with their Fermi?

I've got a water cooled one and I'd love to try to crunch with it. Can someone give me a step by step on how to config my app_info.xml file?

I downloaded a lunatics CUDA app that doesn't work. What do I need to modify?

Thanks!

---

And look who Jord was replying to!

Burt Bacharach, wasn't I?

---

Yes - host 2901600.

It's still experimental and temperamental, and I'm still telling Berkeley what they've done wrong and how to put it right. There were two mistakes made at the Beta site: one has been corrected, I'm still waiting to find out if they've done the other yet. Once we can demonstrate that both the app and the plumbing work (and we need more validations for that - bother these long caches), it should be easy to persuade Berkeley to make the app live on the Main project too.

Patience, Grasshopper.

My three GTX-480's are working ok - I got the files from the CUDA SDK and some from Rastimer. All of my results are completing and being validated - had 42k RAC and rising until the server crash over the weekend. Had 1000 WU's that were waiting to be sent in 24 hours - finally uploaded this morning.

Here is my host http://setiathome.berkeley.edu/show_host_detail.php?hostid=5220505

The files that I am using and an example app_info.xml file can be downloaded here http://rapidshare.com/files/388464735/setiathome.berkeley.edu.rar- your results may vary so please keep that in mind. Sorry for using Rapidshare - it was the quickest way for me to upload these files before I had to head out to a client site.
Todd

---

I have deployed out the five files from Todd's RAR file, and I added this file into the seti project directory

setiathome_6.09_windows_intelx86__cuda_fermi.exe

I modified the app_info.xml file and EXACTLY swapped out from line 27 down - all the CUDA info to be an exact copy of Todd's app_info.xml file.

When I relaunch BOINC I keep getting an error

5/17/2010 5:34:46 PM SETI@home [error] State file error: missing application file

and the application will not download any GPU units.

I could use some help here trying to troubleshoot what file is missing.

I am just trying to run the FERMI cuda app that Todd is running successfully

==
My setup is a single dual core extreme CPU + one Geforce 480.

---

In posted rar executable doesn't match binary used in app_info.
You should get needed binary from SETI beta project probably and place it into SETI main project dir.

Did I post the wrong file? - Which file is it failing on?

---

Todd-

I detached from Seti (regular), and re-attached.

then I quit out of the boinc program
extracted your files and copied them to my seti project directory
edited the app_info.xml file and copied over your entries for CUDA.

however it gets the "missing application file" error when I re-launch boinc.

I don't know how to troubleshoot this; I have no idea what file is missing.

Maybe you can RAR your entire seti directory (less the workunit files) because something is missing.

---

Ok, I will do that shortly and post the files on rapidshare.
Todd

---

Sorry the battery on my laptop died and I didn't have my charger with me. I'm out at a client site and we were doing work on their routers - no internet access was available.

Here are the files (No WU's) that are on my machine. http://rapidshare.com/files/388654390/SETI_Home.rar
I use the AK SSE3 app so if you don't need that disregard these entries. But there might be some correlation why it seems to work for me.

I'd like to give credit to Rastimer and Jason G for helping me get this to work properly. Without their assistance this host wouldn't be online!
Todd

---

Downloading now... will try it out and report back later today. Thanks for your assistance.

---

Happy to help! Sorry it took me a bit to get it posted. Let me know how it goes - either post here or PM.
Todd

---

[...]
Here are the files (No WU's) that are on my machine. http://rapidshare.com/files/388654390/SETI_Home.rar
I use the AK SSE3 app so if you don't need that disregard these entries. But there might be some correlation why it seems to work for me.
[...]

It's this machine?

I would guess this CPU can also SSSE3 & SSE4.1 .
I would use the SSSE3 app.

---

This processor is a generation back from the Nehalem and doesn't support SSSE4 extensions. It is a server chip nonetheless but that functionality didn't come til later.

Here is the CPU-Z validation page http://valid.canardpc.com/show_oc.php?id=1194889 that shows its properties.

Todd

---

This processor is a generation back from the Nehalem and doesn't support SSSE4 extensions. It is a server chip nonetheless but that functionality didn't come til later.

Here is the CPU-Z validation page http://valid.canardpc.com/show_oc.php?id=1194889 that shows its properties.

Todd

That's strange, my pair of X5450's show as having SSE 4.1

http://valid.canardpc.com/show_oc.php?id=1195147

This processor is a generation back from the Nehalem and doesn't support SSSE4 extensions. It is a server chip nonetheless but that functionality didn't come til later.

Here is the CPU-Z validation page http://valid.canardpc.com/show_oc.php?id=1194889 that shows its properties.

Todd

That's strange, my pair of X5450's show as having SSE 4.1

http://valid.canardpc.com/show_oc.php?id=1195147

Yours are showing as E5450, but Todd's as X5450. Different steppings as well.

I have to take what CPU-Z reports with this machine with a pinch of salt some times, it shows as an E5450 but there are a pair of X5450's in there. Even the BIOS and Windows says they're X5450 processors. I've seen CPU-Z report the core voltage at 32.7v before, that can't be right or the chip would have been blown to kingdom come lol

Back on point, the E0 stepping was manufactured some time after the C0 release so if CPU-Z is to be believed, how can an X5450 C0 have SSE 4.1 and an X5470 E0 doesn't?

This processor is a generation back from the Nehalem and doesn't support SSSE4 extensions. It is a server chip nonetheless but that functionality didn't come til later.

Here is the CPU-Z validation page http://valid.canardpc.com/show_oc.php?id=1194889 that shows its properties.

Todd

You mixed now SSSE3 and SSE4.1 .


You said in an upper message that you use SSE3 (2x S). (And it's in your ULed package of the apps).
But I thought your CPU should have at least SSSE3 (3x S).


If you look to the CPU-Z output:
CPU EXT : MMX, SSE (1, 2, 3, 3S), EM64T
Then SSE (x, x, x, 3S) say: SSSE3

But if SSE4.1 or not, your CPU is OCed and AFAIK you should use SSSE3.
The SSE4.1 run faster only on Core2 DUO CPUs.

After looking to a few results, you already use SSSE3 opt. app for CPU.
A result from 13 May 2010 15:27:20 UTC, I guess you have still this app installed.
http://setiathome.berkeley.edu/result.php?resultid=1608228127
Don't be surprised that the app say your CPU is only at ~ 2.6 GHz. If the app take the speed under heavy full load of the CPU, sometimes it don't detect the correct speed. This is also on my CPUs.


So - forget my upper message. ;-)

---

I tried the hosted files and it seems to be working for me. No failures anymore.

For my stock GTX 480, It's finishing a WU (175 credit) in just over 4000 seconds. I'm new to CUDA crunching...is this normal?

I stuck with the .33CPU dedication for the GPU computing, what is optimal?

One thing I noticed when watching GPU usage is that it rarely used 100% of the GPU during its run. It looked like a saw going for mid 60s% to 90% each second.

---

I have to take what CPU-Z reports with this machine with a pinch of salt some times, it shows as an E5450 but there are a pair of X5450's in there. Even the BIOS and Windows says they're X5450 processors. I've seen CPU-Z report the core voltage at 32.7v before, that can't be right or the chip would have been blown to kingdom come lol

Back on point, the E0 stepping was manufactured some time after the C0 release so if CPU-Z is to be believed, how can an X5450 C0 have SSE 4.1 and an X5470 E0 doesn't?

I suppose that all indicates CPU-Z is confused and shouldn't be believed on your system. You might try some other CPUID utility.

In the realm of curiosities, the Intel® Processor Identification Utility may also not report SSE4.1 if the documentation linked from that support page is accurate. The utility was supposedly updated in March 2010 so I suspect they have just missed updating that detail in the documentation.

                                                               Joe