I have a Titan is a number of Boinc projects run smoothly with my Titan and any version of Boinc and nVidia driver, for example,
Moo!, DistrRTgen, PrimGrid, WCG HCC, Einstein, Seti.
Is it really a problem of nVidia driver and Boinc?

This do not mean anything. Distrrtgen is very simple. Also other ones. Genefer is also very specific. Acemd is much complex and requires cpu-gpu as I think. From point of developer I can talk about distrrtgen only and this is very straitforward short app which runs on gpu only.

Folding is now optimized for GeForce Titan, and performance are now multiplied by 2. Future nVidia driver will support the compilation just in time "JIT" and still allow new significant performance gains.

http://folding.typepad.com/news/2013/06/welcome-to-fahcore-17.html

"Welcome to FahCore 17!

We are proud to announce that our latest GPU core, FahCore 17, was recently moved from beta to advanced testing, the last quality assurance step before a full release. As we previously mentioned, this core is a significant step for us. FahCore 17 is a complete overhaul from our previous GPU cores. It brings a cleaner and more streamlined codebase, new serialization mechanisms that allow us to set up diverse simulations, and improved stability. Its use of OpenCL has united our development, allowing the single core to run on both Nvidia and AMD cards, and theoretically any OpenCL-capable device. It is also our first GPU core to run natively in Linux, although we are only supporting Nvidia GPUs there for the time being as we wait for AMD's Linux drivers to mature a bit more. Overall, this core sets a strong foundation for the future of GPU core development.
On AMD cards, FahCore 17 is about 10 times faster than the old GPU cores, and on Nvidia it's about twice as fast. This is mainly due to its OpenMM 5.1 base, which contains many optimizations which deliver a significant speedup. One optimization in particular that we are waiting for is CUDA JIT, a just-in-time compiler that Nvidia may be introducing into its drivers in the coming future. Not only will this technology allow us to offer support for the CUDA platform with FahCore 17, but the JIT compiler is likely to deliver a massive speedup. For the time being, we continue to work at finding additional optimizations on our end. We have also successfully tested FahCore 17 with extremely large proteins (500,000+ atoms), which are on par with the ones used by "bigadv" CPU projects.

To run FahCore 17, you need a Fermi GPU or better and Windows or Linux, or a AMD HD5000 or better and Windows. It also currently requires proprietary drivers from these vendors. You can test FahCore 17 by adding the "client-type = advanced" setting into the extra core options in the V7 client, as in the Configuration FAQ. Another excellent resource is the GPU FAQ which describes why GPUs are so helpful to us.

We'd like to thank all the alpha testers on FreeNode's #fah IRC channel, as well as the beta testers on foldingforum.org, who have all helped us bring the core to this point!"

---

You mention Titans working at Primegrid...it's true that they work, but not at all as expected. I've done some PG testing of my own with my Titan on GFN and some PSA apps, which require FP64, something the Titan is supposed to excel at. It does not. Turning the FP64 ability on and of in the Control Panel adds 4-5% speed increase at best, when it should be closer to 650% (taking into account the clock speed reduction when enabling FP64). So although it "works", it does not work properly.

I also did some testing at Albert@home. Again, the Titan "works", but not properly. WUs take 2x longer to complete on the Titan than on a GTX 560. I haven't tested if the results are the same for Einstein, but you might assume they are.

There is something seriously wrong somewhere, and hopefully Nvidia finally gets it right sooner rather than later, especially since i spent the extra cash and got the EVGA Hydro Copper version. It never runs above 39C at full load, so temps have nothing to do with these issues.

Edit: Also i've had no luck at Poem either. Instant crashes everytime, same as with GPUgrid.

Einstein does not use FP64. So, that is a non issue there. The gk110s do work there, and that is where mine are currently at. However, due to a bug that it would appear NVIDIA is unwilling to fix, they cannot move forward optimizing their CUDA app. AMD excels there on their latest tasks because of it. It's not that they don't perform well, they do, just not as well.

If you are going to use a Titan at Einstein I think it's best to run several WU's concurrently. IIRC running 7 Einstein WU's takes as long as running 1, but you get 7 times the credits. It's probably the same situation at Albert too.

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Einstein is running new WUs, and for whatever reason, I'm having scaling issues on my 780. They don't scale at a 50% increase in time anymore. In fact, mine were pretty much double.

Again, I can only assume this has something to do with the new WUs, and the bug NVIDIA seems unwilling to fix for them. They're stuck on CUDA 3 or something like that.

If you are going to use a Titan at Einstein I think it's best to run several WU's consecutively. IIRC running 7 Einstein WU's takes as long as running 1, but you get 7 times the credits. It's probably the same situation at Albert too.

I think you mean concurrently (all at the same time). Consecutively means one after another, the way BOINC normally does things.

Einstein is running new WUs, and for whatever reason, I'm having scaling issues on my 780. They don't scale at a 50% increase in time anymore. In fact, mine were pretty much double.

Again, I can only assume this has something to do with the new WUs, and the bug NVIDIA seems unwilling to fix for them. They're stuck on CUDA 3 or something like that.

Have a look at http://einstein.phys.uwm.edu/forum_thread.php?id=10028&nowrap=true#125738, posted a week ago in Einstein's "technical news".

This seems to be finally fixed in the current release candidate of CUDA 5.5. We'll see how it comes out when they actually release it.

BM

(the context - the post Bernd was replying to - confirms that 'this' was the problem which has kept them locked down to cuda 3.2)

New driver for nVidia, 326.29

OPENGL 4.4 DRIVER RELEASE NOTES

The Desktop Version for Windows 7 and Windows 8 also supports.

https://developer.nvidia.com/opengl-driver

@+
*_*

---

POEM is working fine on my GTX 780:
http://imgur.com/hGjEn51

POEM is working fine on my GTX 780:
http://imgur.com/hGjEn51

Hello: I have seen his work in POEM and it seems that there are variations in the execution of several simultaneous, suppose that various tests.

If I take as a reference the past that seem to be three tasks simultaneously, it has underperformed my GTX770 that even in Ubuntu, I see much difference, I am surprised.

just upgraded to 326.19 and tried on my titan. it has cuda 5.5.1 with file version 8.17.13.2619. the NATHAN_KIDKIX wu failed the same way as the 320 version of the driver - immediately upon start. thanks for all the discussion and support - please keep the info streaming.

just upgraded to 326.19 and tried on my titan. it has cuda 5.5.1 with file version 8.17.13.2619. the NATHAN_KIDKIX wu failed the same way as the 320 version of the driver - immediately upon start. thanks for all the discussion and support - please keep the info streaming.

If You are trying official app, there is no support yet. Beta tasks are not in queue for now.

We have tried the latest driver and still the same situation.

No news yet of the specific patch from nvidia.

gdf

We have received today the indication of a driver that fixes the problem for both GTX780 and Titan.
We are testing now. Stay tuned.

gdf

Good news.

http://www.nvidia.com/object/linux-display-amd64-325.15-driver.html

This driver works for titan and gtx780 in the sense that simulation is now stable.

We will be getting out a new application asap.

gdf

Good news.

http://www.nvidia.com/object/linux-display-amd64-325.15-driver.html

This driver works for titan and gtx780 in the sense that simulation is now stable.

We will be getting out a new application asap.

gdf

Have you had any luck with the Windows v326.41 driver, released in Beta a week ago?

I have asked which one is the windows version.

gdf

Well thats one down. Fingers crosser windows is released and stable.

This is the one:
http://www.nvidia.com/object/win8-win7-winvista-64bit-326.41-beta-driver.html

gdf