PPS: I would encourage all who can to donate a little something to Sesef for this great effort as well!

If he will set up a PayPal account, I will do my part.

I think this could be a boost to a lot of the BOINC projects, who may have good scientists but they may not be so familiar with programming. They need some expert help.

YES! Just noticed that....

Sesef, PLEASE setup a PayPal account if possible... I can't read polish...

8-)

I think this could be a boost to a lot of the BOINC projects, who may have good scientists but they may not be so familiar with programming. They need some expert help.

+1
See forum of Rosetta@home. After years of crunching, now they are searching for optimizations and they need help.

This week the university is close, we can't test it, but we are going to make all the test with it soon.

Is this app good for science? Can we continue to crunch this?

The first test are all ok. At the moment we didn't have see any differences in the results between this version and our 1.05. We can't support this version, and we haven't check if the app just do what if have to do.

We will contact with the developer soon and we hope that we can add this optimization to the official app.

Regards, Joel.

The first test are all ok. At the moment we didn't have see any differences in the results between this version and our 1.05. We can't support this version, and we haven't check if the app just do what if have to do.

We will contact with the developer soon and we hope that we can add this optimization to the official app.

Regards, Joel.

Sweet!

---

Great!

Sesef, you got $$$$$!!!

Thanks from all of us..

8-)

Great!
Sesef, you got $$$$$!!!

Yeap, great Sesef
If he have time, i like he take a look to this source and analize if possible to have optimized code...
dna@home

Moin,
My Quad 9550 with xp64 pro crash with Optimized app.
30.08.2015 08:03:41 | DENIS@home | [error] Process creation failed: (unknown error) - error code 193 (0xc1)


My i7 with Win7 pro with Optimized app ok.

Is Win XP64 pro not correct?

Sorry for my bad Language.
Greeting Rainer

Wow this is so much faster.
It definitely should become the standard app.
The project would get though a lot more work done using this version or at least using the optimizations.

The first test are all ok. At the moment we didn't have see any differences in the results between this version and our 1.05. We can't support this version, and we haven't check if the app just do what if have to do.

We will contact with the developer soon and we hope that we can add this optimization to the official app.

Regards, Joel.

Any news on the validation of the optimised application?

Hi!

Yes, we are working trying to port the optimization to other platforms. The Sesef version gives the same results as official. But the main problem of that optimization is that it modifies the model, one thing that we do automatically. So we are searching the way to use the optimized part of sesef app to the next version app.

Best regards, Joel

But almost everyone is using Sesef application. How it affects to the project ?

I think they are saying that they need to come out with versions for Linux, etc. and can't do that with just the optimized app itself.

I think they are saying that they need to come out with versions for Linux, etc. and can't do that with just the optimized app itself.

Maybe they should contact Sensef directly, he may be able to help them 'officially'.

I think they are saying that they need to come out with versions for Linux, etc. and can't do that with just the optimized app itself.

Maybe they should contact Sensef directly, he may be able to help them 'officially'.

Hi!

We have contact with him, and we are trying to reproduce the way he compile it in our systems. We want to create optimized version for each platform and also add this optimization for new version of the app.

We are a small develop team and we have also work the research part.

Best regards, Joel.

Maybe trying for AVX could bring an additional strong boost...

---

Maybe trying for AVX could bring an additional strong boost...

Yes, we think that sesef solution is based on it. But we have some problems to port this advantages to the linux version, also the main modification of Sesef is in the model code, which is generated automatically from a CellML model.

For the next version of the app we will include multimodel support and optimize each model would be a lot of work and some models are not public available (unpublished work),so we will see which improvements we can port to the next app version.

Best regards, Joel

Maybe trying for AVX could bring an additional strong boost...

Yes, we think that sesef solution is based on it. But we have some problems to port this advantages to the linux version, also the main modification of Sesef is in the model code, which is generated automatically from a CellML model.

For the next version of the app we will include multimodel support and optimize each model would be a lot of work and some models are not public available (unpublished work),so we will see which improvements we can port to the next app version.

Best regards, Joel

Maybe if you could specify what your problem is porting the app to linux, one could help doing that for you ;)

If anyone could make the port to Linux possible then Crunch3r would be the one to do it!

If anyone could make the port to Linux possible then Crunch3r would be the one to do it!

We are not expert ( at the moment :) ) On compilators and how they work, so we are learning from the work of others ( like Sesef).

We think that there is a problem with the class __md128 and the system. We have workaround that and it could be a conflict between this class and the linux system one that have the same name.

We are using the code directly from Sesef which can be found here: https://github.com/sesef/denis-boinc-baseapp

We are also working on the new version with multimodel support.

Best regards, Joel.