Posts by manalog
log in
1) Message boards : Number crunching : Apple M1 (Message 3081)
Posted 16 Feb 2023 by manalog
Good :) Yes, it definitely seems that it's running through Rosetta considered that it shows Intel SSE2 application... Performances seem good, it would be interesting to compare with a native ARM application to see how much overhead is introduced by binary translation.

Valter, I take the occasion to remind you to list the Android app ;) I disconnected that old phone ( http://gene.disi.unitn.it/test/show_host_detail.php?hostid=59291 )that I used for testing long ago cause at the end battery was going to explode :D but I confirm that the test went perfectly: I got 30.000 credits from it with not even a single invalid unit, and only using the official ARM binary.
2) Message boards : Number crunching : Windows SSE 32-bit Computation errors (Message 2361)
Posted 14 Aug 2021 by manalog
If these machines are intended to run boinc full time, then install Linux on them. I assure that tn-grid without optimizations runs on P4s (like yours), then if I find it I could give you an optimized app I made for P4 and SSE2 last year.
3) Message boards : Number crunching : Bad hosts topic (Message 2287)
Posted 19 Apr 2021 by manalog
Found a bad host: http://gene.disi.unitn.it/test/show_host_detail.php?hostid=25653
4) Message boards : Number crunching : Troubles with Orange Pi 4 (Message 2150)
Posted 28 Dec 2020 by manalog
Hi all,
I attached a brand new Orange Pi 4 to the grid. It is aery good board for boinc: rk3399, 4gb ram. It runs Armbian Ubuntu Bionic.
I noticed that a lot of tasks are invalid: http://gene.disi.unitn.it/test/results.php?hostid=64953&offset=0&show_names=0&state=5&appid=
the interesting thing is that if you see the ratio of valid/invalid tasks it looks like that the tasks are ok in 4/6 core but invalid for 2/6 cores.
At the moment I am writing there are 80 valid tasks and 25 invalid tasks.


Dual-core ARM Cortex-A72 MPCore processor and Quad-core ARM Cortex-A53 MPCore processor, both are high-performance, low-power and cached application processor.
Two CPU clusters. Big cluster with dual-core Cortex-A72 is optimized for high-performance and little cluster with quad-core Cortex-A53 is optimized for low power.
Full implementation of the ARM architecture v8-A instruction set, ARM Neon Advanced SIMD (single instruction, multiple data) support for accelerating media and signal processing
CCI500 ensures the memory coherency between the two clusters
Each Cortex-A72 integrates48KB L1 instruction cache and 32KB L1 data cache with 4-way set associative. Each Cortex A53 integrates 32KB L1 instruction cache and 32kB L1 data cache separately with 4-way set associative
1MB unified L2 Cache for Big cluster, 512KB unified L2 Cache for Little cluster
Trustzone technology support


At a first glance it seems to have problem with the twe a72 cores. I tried it on other projects: rosetta, mlc, wcg opn1. No errors reported, but the board was never attached to only one project for a long period of time as I am doing with tn-grid.
Any idea?
Thanks and Merry Christmas
5) Message boards : Number crunching : Android applications on ARM (Message 2128)
Posted 8 Dec 2020 by manalog
I rooted my Motorola Moto G5S+ and successfully crunched some Tn-grid's workunits.
Results are pretty interesting, 45000 seconds of cpu time is really a good result for a low-power ARM device (8 core). That's roughly the same time of my Orange Pi4 board with RK3399 chip (6 core).
http://gene.disi.unitn.it/test/show_host_detail.php?hostid=65018
6) Message boards : Number crunching : Android applications on ARM (Message 2107)
Posted 26 Nov 2020 by manalog
Hi, any update on this? I confirm that my old phone is still crunching h24 without any error. I think it's safe to add official support to Android in order to allow also users with unrooted phone to participate too.
7) Message boards : Science : AMD COVID-19 HPC Fund (Message 1943)
Posted 16 Sep 2020 by manalog
I am from Italy too and Universities in my country are going to re-open next week, so perhaps we will have an update soon :)
8) Message boards : Science : AMD COVID-19 HPC Fund (Message 1940)
Posted 16 Sep 2020 by manalog
Great news for this project!
https://www.amd.com/en/corporate/hpc-fund
AMD is going to donate computing systems or cloud-based computing capacity as a part of the second round of its COVID-19 HPC Fund initiative.
University of Trento, with the project we are supporting here, Gene@Home, is involved:
“The computational resources that AMD is providing will be inserted in our on-going gene@home project that aims to discover the causal relationships between the expression of genes. In particular, we plan to speed-up specific inquiries about genes interacting with SARS-CoV-2 in collaboration with the medical researchers and students in the Laboratory of Biological Data Mining. The AMD initiative is very timely for us and we expect that it will also push us forward in making further analysis methods feasible“ commented Enrico Blanzieri, professor, Department of Information Engineering and Computer Science, University of Trento.

- Enrico Blanzieri, professor, Department of Information Engineering and Computer Science, University of Trento


I hope these contributions will help to speed the pace of the researches going on this beautiful Boinc project.

Valterc, could you give us some news and details of this operation?
Today the number of "tasks ready to send" is 0... will you get also a new server from AMD?
How is the research going?
Grazie
9) Message boards : Number crunching : Help with invalid tasks and computation errors? (Message 1922)
Posted 15 Aug 2020 by manalog
I got these two versions:
Pentium 4 SSE2
Pentium 3 SSE
The first one should work also on your Pentium M.

If I remember well these binaries are compiled with a "O2" GCC optimization. Using another string (https://gene.disi.unitn.it/test/forum_thread.php?id=270&postid=1883) you can get a performance increase. Now I am in vacation and I only have these two binaries I've already compiled, in the next days I'll compile a version for Pentium 4 and Pentium M using these new optimizations ;)
By they way the version I am sending to you now should already give a great boost compared to the non-SIMD version.

Here the Drive folder where I will put also the new versions, stay tuned!:

https://drive.google.com/drive/folders/1l5MC2VGjXqyapUoxfo9Blw3I8_Faz4N4?usp=sharing
10) Message boards : Number crunching : Help with invalid tasks and computation errors? (Message 1918)
Posted 10 Aug 2020 by manalog
You are doing nothing wrong: the linux 32 app is compiled having a 'core2' processor target, so is not a 'pure' sse2 pentium4 compatible app ranher it probably ingorporates some newer extensions. I have compiled an sse2 version tdat runs on p4 and should run also on yours if it supports sse2. If not, I compiled also a sse version that sdould run even on pentium IiI. Now I am travelling, I will post tdem on this forum for you before friday ;)
Or you can compile it by yourself, just check my post 'sse3 optimization and android binary' on this post and figure out how to do it.
11) Message boards : Science : SARS-CoV-2 virus (Message 1905)
Posted 31 Jul 2020 by manalog
Hello,
any news from this project?
Is the University of Modena and Reggio Emilia happy with the results you provide them?
Are you still working on Covid?
What about the other researches, are they doing well?
Is there any news related to the grant to buy the new server?
Thanks!
12) Message boards : Number crunching : Lost Credit ? (Message 1888)
Posted 8 Jul 2020 by manalog
The main administrator is busy until July, 10 at least. Be patient.
13) Message boards : Number crunching : SSE3 optimization and Android binary (Message 1883)
Posted 7 Jul 2020 by manalog
Hi,
I have compiled an optimized version of the app for my cpu, an Intel Xeon L5420. I got a good increase in performance: the running time passed from an average of 25500 seconds to 19500.

I have used these flags for GCC:
ARCH += -march=native -mtune=native -Ofast -fprofile-use -fno-signed-zeros -fno-trapping-math -frename-registers -funroll-loops


I compiled the application twice:
first time I used -fprofile-generated and ran the application for a while; it generated some .gcda files in the src directory: these files are profile files that enable to further optimize the binary, in order to have it very tailored to your system.
So the second step was to use -fprofile-use in order to use the gcda files previously generated. Actually I didn't created very good profile files, I just ran four or five times the very rapid tests included in the source tree.

I'd like to share this binary with you: https://drive.google.com/file/d/1nGfHSGxbS3QXKp8dIPIqGh5HyJLeRn4g/view?usp=sharing

I have seen several hosts using cpu similar to mine on this project, so other users could use it to increase credit and science ;)
14) Message boards : Number crunching : Orange Pi PC: suddenly all tasks are invalid (Message 1881)
Posted 3 Jul 2020 by manalog
Hi all,
I got a small SBC, an Orange Pi PC, based on the Alwinner H3 SoC (quad-core cortex A7 1,3GHz). It crunched Tn-Grid successfully for a lot of time: a couple of year ago, achieving more than 150.000 credits (http://gene.disi.unitn.it/test/show_host_detail.php?hostid=1841), and this year for the last three months, together with OPN1 (http://gene.disi.unitn.it/test/show_host_detail.php?hostid=58468).
A couple of weeks ago it started returning all invalid tasks, with all the application provided by the server (vfp4, vfp3) that previously worked. It could seem like a problem of the board, but the strange thing is that it works perfectly on several other projects: OPN1, Universe@Home, Einstein@Home, WEP-M+2.
Thus I tried to compile an app targeted for this device
ARCH += -mcpu=cortex-a7 -mfpu=neon-vfpv4 -ftree-vectorize -march=native -mtune=native -Ofast -fprofile-use -fno-signed-zeros -fno-trapping-math -frename-registers -funroll-loops

The speed of execution improved a bit but same problem: all the results are invalid.

Are other people with ARM boards experiencing the same problem? My old android phone crunching on Tn-Grid is still working well...
What kind of test shall I do to figure out if the problem is in my board or in the project?
Thank you.
15) Message boards : Number crunching : Android applications on ARM (Message 1870)
Posted 27 Jun 2020 by manalog
I wish you all the best for your works :)
16) Message boards : Number crunching : Android applications on ARM (Message 1865)
Posted 24 Jun 2020 by manalog
Hi!
The three android devices we put on Tn-Grid seems still to be working well. Mine had only one invalid task in one month running full time.
Valter, do you think it's time to provide official support to Android or it's better to wait more?
Perhaps if you are still concerned about potential validation problem you could send Android WUs only to users who opt in the "allow beta test applications" flag.
17) Message boards : Science : Escherichia Coli and Biobutanol (Message 1853)
Posted 31 May 2020 by manalog
Hi all, after some google researches about biofuels I have found experiments about using Escherichia Coli, genetically modified by adding in it the functions of Clostridium, to produce biobutanol, a biofuel much better than ethanol to be used in gasoline engines.
Here an article about it, more can be found just by looking for biobutanol on a search engine: https://pubmed.ncbi.nlm.nih.gov/25662903/
I know that Tn-grid is working also on E.Choli. Any possible connection with this kind of researches?
Climate change in perspective is something much worser than Covid...
18) Message boards : Number crunching : Android applications on ARM (Message 1849)
Posted 25 May 2020 by manalog
One more host added to testing ;)
http://gene.disi.unitn.it/test/results.php?hostid=59291
19) Message boards : Number crunching : Android applications on ARM (Message 1848)
Posted 23 May 2020 by manalog
Good, probably it should be a good idea to try to start sending the openwrt executable to android devices. Valterc, what do you think about it? I am ready to try this executable on my device too.
Then if in the future someone will compile a faster app, we will think about it again.
20) Message boards : Number crunching : SSE3 optimization and Android binary (Message 1835)
Posted 17 May 2020 by manalog
You have to put in the project's director an app_info.xml file with the name of your executable and then restart boinc.
https://boinc.berkeley.edu/wiki/Anonymous_platform
let us know if everything goes fine!

Then, if your procedure worked, you could explain all the passage you've done so that we will be able to compile also for arm32.
Did you compile the app with some extension? (ex. NEON, PIE)


Next 20

Main page · Your account · Message boards


Copyright © 2024 CNR-TN & UniTN