Posts by [VENETO] boboviz

1) Message boards : Number crunching : Ralph support OpenCL ? (Message 7932)
Posted 27 May 2025 by Profile [VENETO] boboviz
Post:
Rocm 6.0 released


AMD is working on Rocm on Windows (and in-box support on Red Hat, Ubuntu and OpenSuse)
2) Message boards : Current tests : Feeder not running (Message 7931)
Posted 24 May 2025 by Profile [VENETO] boboviz
Post:
Are you using IPv6?

No

Can't be bothered to try and find the Hosts name & put it in my Hosts file incase it's a broken IPv4 configuration here....

The only server i see is ralph.bakerlab.org (128.95.160.141)
3) Message boards : Number crunching : Ralph and SSEx (Message 7930)
Posted 21 May 2025 by Profile [VENETO] boboviz
Post:
Visual Studio Code 1.84 introduces some smells of AI (Copilot)


Visual Studio Code is 1.100!!
4) Message boards : Current tests : Feeder not running (Message 7928)
Posted 16 May 2025 by Profile [VENETO] boboviz
Post:
Still....
5) Message boards : Current tests : Docker on Boinc (Message 7925)
Posted 9 May 2025 by Profile [VENETO] boboviz
Post:
The name of the functionality is BUDA (Boinc Universal Docker App) and will be released with the next boinc manager release.
6) Message boards : Current tests : Feeder not running (Message 7924)
Posted 9 May 2025 by Profile [VENETO] boboviz
Post:
On my boinc manager i have this error: "Server error: feeder not running. Project requested delay of 3600 seconds"
But, in the server page, the feeder is green and running....
7) Message boards : Number crunching : Ralph support OpenCL ? (Message 7923)
Posted 6 May 2025 by Profile [VENETO] boboviz
Post:
IWOCL 2025 Conference Program
8) Message boards : Number crunching : Ralph and SSEx (Message 7914)
Posted 25 Feb 2025 by Profile [VENETO] boboviz
Post:
Google released Gemini Code Assist as a free plugin for VisualStudioCode, Github, AndroidStudio and others
9) Message boards : Number crunching : Ralph support OpenCL ? (Message 7906)
Posted 7 Jan 2025 by Profile [VENETO] boboviz
Post:
ZLUDA is back!! And now is focused on AI/Machine Learning.


First usable version of Zluda (Zluda 4.0) even if with limited, for now, capabilities.

P.S.
I'm curious to see when Nvidia will remove it. Again.
10) Message boards : Number crunching : Ralph support OpenCL ? (Message 7905)
Posted 7 Jan 2025 by Profile [VENETO] boboviz
Post:
With hip-script you can run Nvidia Cuda and AMD Hip on your browser.
11) Message boards : Number crunching : Ralph support OpenCL ? (Message 7901)
Posted 12 Dec 2024 by Profile [VENETO] boboviz
Post:
Standard C/C++ code DIRECTLY on GPU (by AMD)
12) Message boards : Number crunching : Optimize or abandon the VMs? (Message 7898)
Posted 3 Dec 2024 by Profile [VENETO] boboviz
Post:
The name of the project is BUDA
13) Message boards : Number crunching : Ralph support OpenCL ? (Message 7897)
Posted 3 Dec 2024 by Profile [VENETO] boboviz
Post:
The Folding@Home team are migrating, for AMD gpu, from OpenCl to AMD HIP.
The tests are indicating that performances are doubled!!
14) Message boards : Number crunching : GPU tasks failing (Message 7887)
Posted 23 Oct 2024 by Profile [VENETO] boboviz
Post:
No new work or applications to test, and broken stats & Ralph preferences for well over a month now.


Maybe they are busy celebrating the Nobel Prize... :-P
15) Message boards : Number crunching : Ralph and SSEx (Message 7881)
Posted 16 Oct 2024 by Profile [VENETO] boboviz
Post:
AOCC 5 seems fast!
16) Message boards : Current tests : Docker on Boinc (Message 7880)
Posted 11 Oct 2024 by Profile [VENETO] boboviz
Post:
Boinc developers are working on Docker implementation to "abandon" Virtualbox, 'cause Docker is:
- faster
- easier to mantain
- gpus compliant

This could be a re-start of python application of Rosetta...
17) Message boards : Number crunching : GPU tasks failing (Message 7879)
Posted 11 Oct 2024 by Profile [VENETO] boboviz
Post:
Seems that the development of the gpu app is frozen...
18) Message boards : Number crunching : Ralph support OpenCL ? (Message 7878)
Posted 11 Oct 2024 by Profile [VENETO] boboviz
Post:
There is A LOT of tools to convert automatically (or semi-automatically) CUDA code to OpenCL/OneApi/SYCL/etc.
From AdaptiveCpp (using C++ as common language between Cuda/OpenCl/Sycl) to SpectralCompute, from HIPIFY to Coriander
But there are 2 problems:
- Nvidia could block your work (as they did with ZLUDA). I know, it's not a good thing for science to block others competitor, but CUDA is closed, so Nvidia can do it.


ZLUDA is back!! And now is focused on AI/Machine Learning.
19) Message boards : RALPH@home bug list : Rosetta 4.29 (Message 7859)
Posted 8 Sep 2024 by Profile [VENETO] boboviz
Post:
I've got few of these Rosetta 4.29 tasks too and they also Validate Error


Same here. Validation error despite, in the logs all seems ok...


At the end, seems that the validator is ok....
20) Message boards : RALPH@home bug list : RoseTTAFold All-Atom (Message 7846)
Posted 3 Sep 2024 by Profile [VENETO] boboviz
Post:
It is TOO BAD that volunteers get NO CREDIT for their 1000's of seconds of machine time when the failure is caused by the researcher submitting the broken work.
The same bug has been reported here for some time.


Some times ago, i proposed a "minimal credits" for bugged wus, a symbolical "premium" for volunteers that runs wus (and, sometimes, reports problems).
Obviously, no answer.


Next 20



©2025 University of Washington
http://www.bakerlab.org