Another thing, firing up my "slower" laptop, it started to process Cuda WU's again. It's not using the 6.08 though, but 6.06.

Just got home and the "slower" computers are trying for 6.08 too now. Hope the guys (and gals??) at Berkeley can sort things out soon.

---

Matt has got the message - hopefully it should be a quick and easy problem to fix.

Is anyone successfully able to download 6.08?

---

@SETIEric@qoto.org (Mastodon)

I think I've solved the problem... Does it work now?

When generating our cryptographics signatures on the executable files, the files get copied from our workstations to a USB drive. That drive gets connected to a secure machine (no network, locked up in a cage) which contains our signing key. The files are signed and the files and signatures are copied to the download server. Apparently in that final copy a single bit error crept into the file "cudart.dll". I've overwritten it with a clean copy.

Eric

---

@SETIEric@qoto.org (Mastodon)

It still does not work.

The applications are still with checksum and signature error, but every single wu gets an dl-error too now.

---

Just tried it, still no luck:-(
1/21/2009 1:58:18 PM|SETI@home|Finished download of setiathome_6.08_windows_intelx86__cuda.exe
1/21/2009 1:58:18 PM|SETI@home|[error] Signature verification failed for setiathome_6.08_windows_intelx86__cuda.exe
1/21/2009 1:58:18 PM|SETI@home|[error] Checksum or signature error for setiathome_6.08_windows_intelx86__cuda.exe

---

I think I've solved the problem... Does it work now?

When generating our cryptographics signatures on the executable files, the files get copied from our workstations to a USB drive. That drive gets connected to a secure machine (no network, locked up in a cage) which contains our signing key. The files are signed and the files and signatures are copied to the download server. Apparently in that final copy a single bit error crept into the file "cudart.dll". I've overwritten it with a clean copy.

Eric

Eric,

If you look down to the opening post in this thread, you'll see that the initial report was for

Signature verification failed for setiathome_6.08_windows_intelx86__cuda.exe

- maybe that file dropped a bit, as well as the .DLL?

Edit - maybe we should all club together to buy Eric a new USB stick ;-) ?

Still getting the same errors:-(
01/21/09 14:39:37|SETI@home|Finished download of setiathome_6.08_windows_intelx86__cuda.exe
01/21/09 14:39:37|SETI@home|[error] Signature verification failed for setiathome_6.08_windows_intelx86__cuda.exe
01/21/09 14:39:37|SETI@home|[error] Checksum or signature error for setiathome_6.08_windows_intelx86__cuda.exe

---

How do you force the download of version 6.08?

---

I just abort some pending work units (usually the largest ones), then click update

---

THANKS!
That works great.

---

It's working now.

Thanks Eric.

---

Thanks Eric,

Everything works fine now :-))

---

It works now for me too:-)
Thanks Eric!

---

It works now for me too:-)
Thanks Eric!

same here, everything running again!

---

I just abort some pending work units (usually the largest ones), then click update

I have just a few MB WUs remaiming to complete in my Seti@Home-Main (non-beta) project client, so I want to let them finish (and not Abort them). I drained my cache earlier today, Allowed New Tasks, but did not automatically download the 6.08-Cuda application.

So what do I need to do to auto-download the 6.08-Cuda app in Main project? I do have an app_info.xml file that refers to AK_v8 optimized MB app and to the optimized AP-V500-R69 app. My S@H Compute Preferences (in my web account) is set to Allow Cuda Task= yes.

(I did get 6.08-Cuda automatically in my S@H-Beta account ~14Jan09, and have crunched many Cuda WUs in Beta on my 8800GT/512mb video card since mid-Dec08 using 6.06-Cuda, 6.07-Cuda and now 6.08-Cuda).

Please advise.

---

Sabertooth Z77, i7-3770K@4.2GHz, GTX680, W8.1Pro x64
P5N32-E SLI, C2D E8400@3Ghz, GTX580, Win7SP1Pro x64 & PCLinuxOS2015 x64

So what do I need to do to auto-download the 6.08-Cuda app in Main project? I do have an app_info.xml file that refers to AK_v8 optimized MB app and to the optimized AP-V500-R69 app. My S@H Compute Preferences (in my web account) is set to Allow Cuda Task= yes.

The presence of "app_info.xml" is preventing apps auto-download.

So the choice is yours:

• Do not use optimized apps (delete optimized apps and app_info.xml).
  Pro: simpliest way, allows future auto-update of stock apps
  Contra: AP crunching reverts to slow stock app.
  
  
• Or do manual copy of existing in beta folder bunch of ver. 6.08 files into main S@H folder and edit existing app_info.xml (replace everywhere filename of your optimized MultiBeam app, for example, "AK_v8_win_SSSE3x.exe" with the filename of stock app, for example, "setiathome_6.08_windows_intelx86__cuda.exe")
  Pro: AP crunching is still optimized
  Contra: no auto-update
  

Leopoldo:

Thanks for your reply and detailed two options for me to get 6.08-Cuda onto my non-beta S@H project.

Since I have a dual-core E8400 CPU, I'm wondering if there's a third possible option which would allow 1 core to support the GPU-based 6.08-Cuda app (Cuda would show "0.1 CPU" feeding the GPU), and 1 whole core to concurrently run optimized MultiBeam AK_v8 app? Could I do this by:
(1) first either renaming my app_info.xml file and auto-downloading 6.08-Cuda into my seti folder, OR first copying bunch of 6.08 files + cudart.dll + cufft.dll files into my ...\BOINC\projects\setiathome.berkeley.edu folder; and then
(2) ADDING a new <app> section for "setiathome_6.08_windows_intelx86__cuda.exe" to my app_info.xml file (rather than Replacing all references to my "AK_v8_win_SSE41.exe" app);
so that 0.1CPU would feed 6.08-Cuda on GPU and the other core would run AK-v8 optimized MB app? I recognize that future apps would not be auto-updated if I can implement this proposed third option.

(Incidentally, I'm using BOINC v6.4.5, and presently I have set Allow AP= NO for both S@H main and S@H-Beta Preferences.)

When I run a Cuda WU in Beta project (0.1CPUs 1 CUDA), I can concurrently run a MB WU using optimized AK-v8 in S@H main project, so I am wanting to do Cuda+Opt all in main project.

Please advise if my proposed 3rd option (or any variation of it) will work.

---

Sabertooth Z77, i7-3770K@4.2GHz, GTX680, W8.1Pro x64
P5N32-E SLI, C2D E8400@3Ghz, GTX580, Win7SP1Pro x64 & PCLinuxOS2015 x64

Leopoldo:

Thanks for your reply and detailed two options for me to get 6.08-Cuda onto my non-beta S@H project.

Since I have a dual-core E8400 CPU, I'm wondering if there's a third possible option which would allow 1 core to support the GPU-based 6.08-Cuda app (Cuda would show "0.1 CPU" feeding the GPU), and 1 whole core to concurrently run optimized MultiBeam AK_v8 app? Could I do this by:
(1) first either renaming my app_info.xml file and auto-downloading 6.08-Cuda into my seti folder, OR first copying bunch of 6.08 files + cudart.dll + cufft.dll files into my ...\BOINC\projects\setiathome.berkeley.edu folder; and then
(2) ADDING a new <app> section for "setiathome_6.08_windows_intelx86__cuda.exe" to my app_info.xml file (rather than Replacing all references to my "AK_v8_win_SSE41.exe" app);
so that 0.1CPU would feed 6.09-Cuda on GPU and the other core would run AK-v8 optimized MB app? I recognize that future apps would not be auto-updated if I can implement this proposed third option.

(Incidentally, I'm using BOINC v6.4.5, and presently I have set Allow AP= NO for both S@H main and S@H-Beta Preferences.)

Please advise if my proposed 3rd option (or any variation of it) will work.

Not possible at present, you can eithier run AK_V8 and Optimised AP with an app_info,
or you can run Cuda and Optimised AP with an app_info,
but you can't run Cuda and AK_V8 with an app_info.

Claggy