on my machines with small nvidia GPU, this app pushed it's way onto the GPU which was already running something from another project - not good for what looks like a CPU app to the boinc-client. this needs to behave like other BOINC apps with separate CPU and GPU app versions. if not, this is the end of me running this app on such machines

I updated Boinc manually at 11:32:10 to get new Wus but the result was that Ralph@home wont send me Wus for the rest of the day. Curios!

It can't send you any work if there is no work to send.

From the Server Status page

Tasks ready to send  0

But it also says in his log he posted "This computer has finished a daily quota of 1 tasks"

on my machines with small nvidia GPU, this app pushed it's way onto the GPU which was already running something from another project - not good for what looks like a CPU app to the boinc-client. this needs to behave like other BOINC apps with separate CPU and GPU app versions. if not, this is the end of me running this app on such machines

It's a test project, they have to start somewhere. Although you'd think trying to run them on their own machine would be much faster for them to find basic problems.

No, this is precisely the same as say Einstein GPU tasks which need a CPU thread or a part of one. The task is given to me as 1GPU + 0.7 CPU threads.

You need to read what i am posting, not what you think i am posting.

---

Grant
Darwin NT

No, this is precisely the same as say Einstein GPU tasks which need a CPU thread or a part of one. The task is given to me as 1GPU + 0.7 CPU threads.

You need to read what i am posting, not what you think i am posting.

I understood what you wrote perfectly. You seem to think this is something never done before, a task which uses both GPU and CPU. What do you believe is different between Ralph tasks and Einstein GPU tasks?

I updated Boinc manually at 11:32:10 to get new Wus but the result was that Ralph@home wont send me Wus for the rest of the day. Curios!

It can't send you any work if there is no work to send.

From the Server Status page

Tasks ready to send  0

But it also says in his log he posted "This computer has finished a daily quota of 1 tasks"

And all the requests before those two, it didn't send out any work because there wasn't any to send.

---

Grant
Darwin NT

No, this is precisely the same as say Einstein GPU tasks which need a CPU thread or a part of one. The task is given to me as 1GPU + 0.7 CPU threads.

You need to read what i am posting, not what you think i am posting.

What do you believe is different between Ralph tasks and Einstein GPU tasks?

For the third time and final time-
Here there is one single application that does both CPU & GPU work for a given type of Task.
On all other projects they have two different applications when doing work for the same type of Task on the CPU and GPU.

---

Grant
Darwin NT

I updated Boinc manually at 11:32:10 to get new Wus but the result was that Ralph@home wont send me Wus for the rest of the day. Curios!

It can't send you any work if there is no work to send.

From the Server Status page

Tasks ready to send  0

But it also says in his log he posted "This computer has finished a daily quota of 1 tasks"

And all the request before that, it didn't send out any work because there wasn't any to send.

But presumably the quota message is what he's referring to.

No, this is precisely the same as say Einstein GPU tasks which need a CPU thread or a part of one. The task is given to me as 1GPU + 0.7 CPU threads.

You need to read what i am posting, not what you think i am posting.

What do you believe is different between Ralph tasks and Einstein GPU tasks?

For the third time and final time-
Here there is one single application that does both CPU & GPU work for a given type of Task.
On all other projects they have two different applications when doing work for the same type of Task on the CPU and GPU.

Forget the Einstein CPU tasks, this seems to be confusing you.

Look at the Einstein GPU tasks, they use BOTH the cpu and gpu. Just like Ralph, precisely no difference.

You seem to think this is something never done before, a task which uses both GPU and CPU. What do you believe is different between Ralph tasks and Einstein GPU tasks?

Once again, you haven't paid the slightest bit of attention to what i have posted.

I'll try to make it easier- Task and application are not the same thing- they are different, not the same, different.

A Task is the data that is processed.
The application is what does the processing.

On all other projects, for the same task, different applications are used for doing work on the CPU & work on the GPU for given hardware and Operating System.



Now try reading my posts again.

---

Grant
Darwin NT

Look at the Einstein GPU tasks, they use BOTH the cpu and gpu. Just like Ralph, precisely no difference.

That shows just how confused you are. The GPU processes the Task, the CPU supports the GPU by keeping it fed. The CPU doesn't actually do any processing, the GPU does that. Depending on the Task, with a very well written application, CPU support can be next to nothing.
And none of that has any relevance to the post i made & you misread & mis-quoted. Repeatedly, over and over again.

---

Grant
Darwin NT

Look at the Einstein GPU tasks, they use BOTH the cpu and gpu. Just like Ralph, precisely no difference.

That shows just how confused you are. The GPU processes the Task, the CPU supports the GPU by keeping it fed. The CPU doesn't actually do any processing, the GPU does that. Depending on the Task, with a very well written application, CPU support can be next to nothing.
And none of that has any relevance to the post i made & you misread & mis-quoted. Repeatedly, over and over again.

Jesus Christ this is absurd. Firstly, the CPU doesn't just feed it. Especially in Einstein a lot of processing is done on the CPU before it can hand data to the GPU. As to your other post, who cares what applications are used? All we need is Boinc to be told, this will require x GPUs and y CPUs.

In LHC the same application runs ATLAS tasks on anything from 1 to 8 CPU threads. You don't need a different application just to change the requirements!!

Jesus Christ this is absurd.

It's gone well beyond that.

As to your other post, who cares what applications are used?

If you had read any of my previous posts and understood them, you would have your answer already.

All we need is Boinc to be told, this will require x GPUs and y CPUs.

For a given application. Not for a given Task, for a given application.
Is this starting to sink in? At all?

In LHC the same application runs ATLAS tasks on anything from 1 to 8 CPU threads. You don't need a different application just to change the requirements!!

You really don't have the slightest clue do you?
Are you even paying attention to what you are typing? Or do you just not have the slightest idea of what it is you are discussing, as you continually keep conflating things.

The Atlas CPU application is for doing work on the CPU, and the CPU only. it may use one or 50 threads, but it's all done on the CPU. Comprende?
If there were a GPU application for ATLAS, then the work would be done on the GPU, most likely requiring CPU support, but the work world be done on the GPU.
Notice how i am taking bout 2 different applications, to process the same type of Task?

Just because the Einstein application needs the CPU to do some work for it, that doesn't make it a CPU application. It's a GPU application, for doing work on the GPU. No GPU- no work.
If there was a CPU application, then it could do the same work, without the need for the GPU.
Notice how i am talking about two different applications, that can process the same Tasks?

---

Grant
Darwin NT

on my machines with small nvidia GPU, this app pushed it's way onto the GPU which was already running something from another project - not good for what looks like a CPU app to the boinc-client. this needs to behave like other BOINC apps with separate CPU and GPU app versions. if not, this is the end of me running this app on such machines

Is it trying to run those Tasks on your GT 1030???

---

Grant
Darwin NT

Jesus Christ this is absurd.

It's gone well beyond that.

As to your other post, who cares what applications are used?

If you had read any of my previous posts and understood them, you would have your answer already.

All we need is Boinc to be told, this will require x GPUs and y CPUs.

For a given application. Not for a given Task, for a given application.
Is this starting to sink in? At all?

In LHC the same application runs ATLAS tasks on anything from 1 to 8 CPU threads. You don't need a different application just to change the requirements!!

You really don't have the slightest clue do you?
Are you even paying attention to what you are typing? Or do you just not have the slightest idea of what it is you are discussing, as you continually keep conflating things.

The Atlas CPU application is for doing work on the CPU, and the CPU only. it may use one or 50 threads, but it's all done on the CPU. Comprende?
If there were a GPU application for ATLAS, then the work would be done on the GPU, most likely requiring CPU support, but the work world be done on the GPU.
Notice how i am taking bout 2 different applications, to process the same type of Task?

Just because the Einstein application needs the CPU to do some work for it, that doesn't make it a CPU application. It's a GPU application, for doing work on the GPU. No GPU- no work.
If there was a CPU application, then it could do the same work, without the need for the GPU.
Notice how i am talking about two different applications, that can process the same Tasks?

Stop getting so annoyed. Why do we need two applications? If the server can see you have a compatible GPU, and the client asks for GPU work, it gives it what we're currently getting and marks it as 1GPU and 1CPU. If the client asks for CPU work only, you're given THE SAME APPLICATION marked as needing 8 CPUs. Very simple, stop overthinking things and complicating matters. All we need and nothing else whatsoever, is the task to get marked as needing what it's going to use.

Stop getting so annoyed.

Start paying attention.

Why do we need two applications?

Once again, you ask a question that i have repeatedly answered.

There are none so deaf and blind as those that refuse to listen or look.

If the server can see you have a compatible GPU, and the client asks for GPU work, it gives it what we're currently getting and marks it as 1GPU and 1CPU. If the client asks for CPU work only, you're given THE SAME APPLICATION marked as needing 8 CPUs. Very simple, stop overthinking things and complicating matters. All we need and nothing else whatsoever

And here you are suggesting what i had already suggested as part of my list of things that need fixing, right at the very end as number 8.

Yet you forgot the most important part of that- the time taken to do a given Task by the slowest of CPUs must be the same as the time taken to do that very same Task by the most powerful of GPUs.
If not, all Scheduling is screwed & deadlines will be missed, Resource Share balancing will take forever, if it were to occur at all (oh, and i forgot about the random nature of the amount of Credit being awarded).

Which is why on every other BOINC project where the Runtime of Tasks is not fixed, they have different applications for CPU work and GPU work, for processing the same Tasks, on the same hardware, with the same OS (and even if like here the application is the same, the application names are different for the CPU & that GPU- effectively providing different CPU & GPU applications).

If you had read my earlier post on what needs fixing, the above would look rather familiar.

---

Grant
Darwin NT

Stop getting so annoyed.

Start paying attention.

Maybe if you didn't waffle so much....

Why do we need two applications?

Once again, you ask a question that i have repeatedly answered.

There are none so deaf and blind as those that refuse to listen or look.

Somewhere in your screeds of text.

If the server can see you have a compatible GPU, and the client asks for GPU work, it gives it what we're currently getting and marks it as 1GPU and 1CPU. If the client asks for CPU work only, you're given THE SAME APPLICATION marked as needing 8 CPUs. Very simple, stop overthinking things and complicating matters. All we need and nothing else whatsoever

And here you are suggesting what i had already suggested as part of my list of things that need fixing, right at the very end as number 8. "This will probably be the single biggest issue, as history has shown us that without separate tracking of actual processing times for separate initial estimated completion times, honouring Resource shares becomes near impossible, and systems download more work than they can finish, as well as download less work than they can handle, all depending on how much work is done on what compute resource."

Boinc fails even without this, it can't cope with GPU and CPU tasks from any project, the times go everywhere. Primegrid is a good example. It only records one correction factor for the whole host.

Yet you forgot the most important part of that- the time taken to do a given Task by the slowest of CPUs must be the same as the time taken to do that very same Task by the most powerful of GPUs.
If not, all Scheduling is screwed & deadlines will be missed, Resource Share balancing will take forever, if it were to occur at all.

Rubbish. Most projects allow time for the slowest CPU, and the faster ones finish earlier.

I'm assuming you're American and having difficulties with our language. I give up, someone else can explain it to you. Unsubscribed from this thread and you, permanently.

Boinc fails even without this, it can't cope with GPU and CPU tasks from any project, the times go everywhere. Primegrid is a good example. It only records one correction factor for the whole host.

That isn't a BOINC fail, that is a Prime Grid fail.
Duration Correction Factor hasn't been used by BOINC for over a decade. If they made use of the Credit New runtime estimates function, then the initial Estimated processing time & Scheduling would work as intended.
And thank you for providing the proof that what i have been saying all along is correct. A perfect example of why a single application for CPU & GPU with variable length runtime Tasks won't work, and how having one application for CPU work and another for GPU work does actually function correctly for those types of work.

Rubbish. Most projects allow time for the slowest CPU, and the faster ones finish earlier.

And once again you make a response that shows you either didn't read what i posted, or weren't able to understand it.
I posted about Rosetta type work where all Tasks run for the same length of time, regardless of what processes it. Then you talk about different processing times for different capability processors.
You continue to conflate things.

I'm assuming you're American and having difficulties with our language.

And your excuse for your difficulties in comprehending straight forward English?

---

Grant
Darwin NT

No, this is precisely the same as say Einstein GPU tasks which need a CPU thread or a part of one. The task is given to me as 1GPU + 0.7 CPU threads.

No, this is not the same as Einstein: here we are an 8-cores cpu app+gpu.
And, if you have multiple, concurrent wus all goes wrong

It is like GPU grid Python apps for GPU hosts