There is a script running in the background now ridding the database of all that crud, which is causing some slowdowns here as it's quite resource intensive.

Excellent.

Seconded.

---

Grant
Darwin NT

[quote]

The current stock NV & ATI apps are r3584.

But with crappy CPU apps which offset the r3584 gains.
/quote]
If you weren't using the CPU application would you recommend switching to the stock app r3584 for either Nvidia or ATI?

---

The current stock NV & ATI apps are r3584.

But with crappy CPU apps which offset the r3584 gains.

If you weren't using the CPU application would you recommend switching to the stock app r3584 for either Nvidia or ATI?

Why?

Cheers.

If you are only running GPUs in a system, then yes you would be a little better off reverting to the newer stock app.
Just rename your app_info and restart.

BUT remember to run your current cache dry, or you will lose it !!!!!!!

Editing the app_info really isn't that hard once you understand it.

Richard puts a LOT of work into ensuring that the installer will retain your current cache with each revision - that is why the file is so bloody long. In reality you need very little of it once you have run off any old versions of tasks after the install/reinstall. (i.e. If your running v8.12 NVidia from a previous install, you don't need that NVIDIA v8.12 definition after those tasks are gone and your running v8.21)

The easiest way to start is to get rid of all the old stuff like v8.12. Look at what version the BOINC manager says your running and remove the other versions for that same application type (i.e. MB NVIDIA.) and your list gets much shorter.

For CPU+NVIDIA MB+AP you need a total of 4 apps defined, about 60 lines, that's it. Once you widdel out the old definitions if becomes much more clear as to the layout of the file and very easy to see what to change.

Make a copy of the file and try it, you will understand it better.

The current stock NV & ATI apps are r3584.

But with crappy CPU apps which offset the r3584 gains.

If you weren't using the CPU application would you recommend switching to the stock app r3584 for either Nvidia or ATI?

Why?

Cheers.

My whole answer to this is, it depends on your GPU and CPU

For my part, I never run CPU tasks. Because using the CPU alongside the GPU slows down the GPU work units.

I run multiple work units on each GPU, having figured out the best ratio of that for my GPUs. Any credit that a CPU work unit would give doesn't compensate for the slowdown and productivity that is lost from the GPUs. The GPUs are more productive overall than the CPU.

You would need to see how well the system does with only GPU work per hour. Then add CPU task and see what the productivity is. If the overall number is lower, then you should drop the CPU work units.

My 2 cents......

Zalster

---

For my part, I never run CPU tasks. Because using the CPU alongside the GPU slows down the GPU work units.

That is why I prefer running the -hp flag much rather than use_sleep or the default priority - give the GPU the resources it wants, and don't run the CPU "balls to the wall" so you don't have to fight for CPU time.

I really noticed this in Linux where even using 3/4 of the CPU had a significant effect on the GPU, but have since found a way around the default CPU priorities assigned to tasks.

I found several hundreds of teams and user profiles...

A BIG thank You!

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

Hi all,

Regarless of the APP (Cuda/OpenCl) you are running the workload for GPU app needing for CPU is mainly 1) integer/boolean only and 2) used just to check if the GPU has finished processing.

This is why I run HT on. 6 cores becomes to 12 HT cores. I run 6 CPU tasks because the CPU has 6 FP/AVX units. The rest 6 HT cores can do the download from GPU and postprocessing of the results that my 4 GPUs produce. The GPUs need a HT core to be ready to notice a result is ready -- to minimize latency. I have set my CUDA app to loop actively (and placed it on a HT core) to wait the GPU. An active NOP loop does not consume much power nor does is hinder the performance of the real core. The Linux seems to be able to differentiate 'real' work from an active NOP-wait loop.

I do tell the Linux scheduler to run the CPU apps on defined CPU ID's and to run the GPU apps on some other subset of "CPU'"ID's. I nicely ask the sytem: "Please do run the CPU app on real core and the GPU app on the HT core and leave 2 of them HT cores to handle my network and system/everyday tasks.

You may do it some other way. And that is OK.

Petri

---

To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones

Hi all,

Regarless of the APP (Cuda/OpenCl) you are running the workload for GPU app needing for CPU is mainly 1) integer/boolean only and 2) used just to check if the GPU has finished processing.

This is why I run HT on. 6 cores becomes to 12 HT cores. I run 6 CPU tasks because the CPU has 6 FP/AVX units. The rest 6 HT cores can do the download from GPU and postprocessing of the results that my 4 GPUs produce. The GPUs need a HT core to be ready to notice a result is ready -- to minimize latency. I have set my CUDA app to loop actively (and placed it on a HT core) to wait the GPU. An active NOP loop does not consume much power nor does is hinder the performance of the real core. The Linux seems to be able to differentiate 'real' work from an active NOP-wait loop.

I do tell the Linux scheduler to run the CPU apps on defined CPU ID's and to run the GPU apps on some other subset of "CPU'"ID's. I nicely ask the sytem: "Please do run the CPU app on real core and the GPU app on the HT core and leave 2 of them HT cores to handle my network and system/everyday tasks.

You may do it some other way. And that is OK.

Petri

I accomplish the same thing on Windows with the ProcessLasso application. I have all CPU tasks [MB8_win_x64_AVX_VS2010_r3330.exe] assigned to real CPU ID's [i.e. CPU #0, CPU#2, etc] with CPU affinity. The rest of the virtual cores can service the desktop, apps and feed the GPU's.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

The script finished and caught most of them. Because the spammers were immediately repopulating the database, new security measures have been added:
1. Team creation via web RPC has been disabled.
2. To be able to create a team, one now has to finish a reCaptcha.

Let's hope that holds up for a while. At least this way they won't be slowing the forums down again, when that happens it'll be something else.

---

Good Job Jord!

+1

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Oh, I didn't do anything. Just found the myriad of bad teams and users. Reported that all to David who ran the script and added the security. So you'll have to thank him for once. :)

---

The current stock NV & ATI apps are r3584.

But with crappy CPU apps which offset the r3584 gains.

If you weren't using the CPU application would you recommend switching to the stock app r3584 for either Nvidia or ATI?

Why?

Cheers.

I was wanting to see if anybody had tried this and noticed a speed increase. Hence the reason why I was asking.

---

I was wanting to see if anybody had tried this and noticed a speed increase. Hence the reason why I was asking.

I didn't see any speed increase. The benefit of r3584 is a decrease in invalids.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I was wanting to see if anybody had tried this and noticed a speed increase. Hence the reason why I was asking.

I didn't see any speed increase. The benefit of r3584 is a decrease in invalids.

Thanks Keith

---

I didn't see any speed increase. The benefit of r3584 is a decrease in invalids.

Invalids or inconclusives?
I'm running r3557, and it's been ages since I've had a GPU Invalid. With some of the previous SoG applications there used to almost always be one or 2 (or more) there in my task list.

---

Grant
Darwin NT

I didn't see any speed increase. The benefit of r3584 is a decrease in invalids.

Invalids or inconclusives?
I'm running r3557, and it's been ages since I've had a GPU Invalid. With some of the previous SoG applications there used to almost always be one or 2 (or more) there in my task list.

Oops. Misspoke. Yes, inconclusives. I don't create invalids unless I am doing something stupid to the computer while it is crunching and crash it.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

I didn't see any speed increase. The benefit of r3584 is a decrease in invalids.

Invalids or inconclusives?
I'm running r3557, and it's been ages since I've had a GPU Invalid. With some of the previous SoG applications there used to almost always be one or 2 (or more) there in my task list.

Oops. Misspoke. Yes, inconclusives. I don't create invalids unless I am doing something stupid to the computer while it is crunching and crash it.

The only inconclusives I get now while running r3557 are created by others still using the older SoG apps or rigs that are just plain bad.

Cheers.