Posts by ebahapo

1) Message boards : Feedback : AMD64 (Message 3552)
Posted 17 Dec 2007 by ebahapo
Post:
Even though there's an official AMD64 client for Linux, it refers to too many dynamic libraries and requires a fairly recent Linux setup to run on.

So, one more time, I'm making available the AMD64 Linux client here. It refers to a minimal set of standard dynamic libraries whose version requirements should be satisfied by Linux systems up to 2 or 3 years old, however it was built with a fairly recent version of GCC, 4.1.2.

The drill's still the same:


The official AMD64 Windows client can be found here.

For more information, see the BoincStats Forum.

HTH

2) Message boards : Feedback : AMD64 (Message 3322)
Posted 13 Aug 2007 by ebahapo
Post:
Here's a development version of the x86-64 Linux client:


The official client for x64 Windows client can be found at boinc_5.10.13_windows_x86_64.exe.

The BOINC client 5.10 can now get 32-bit applications from projects that haven't added support for AMD64 (e.g., Lattice, QMC, etc), provided that they run at least the BOINC server 5.0.9. However, such AMD64 clients for Windows may not get applications from some projects that supported AMD64 before due to a platform name change, at least until such projects are updated.

For more information, see the BoincStats Forum.

HTH

3) Message boards : Feedback : AMD64 (Message 2376)
Posted 10 Oct 2006 by ebahapo
Post:
So far, two projects have added support for x86-64: SETI sends the x86 application and SIMAP, a true x86-64 application. Would you consider supporting the platform x86_64-pc-linux-gnu or x86_64-unknown-linux-gnu similarly?

Chess960 now supports AMD64 with a native Linux application too...

HashClash now supports AMD64 with a 32-bit Linux application...

HashClash now supports AMD64 on Windows with a 32-bit application...

Guess what? Leiden Classical now supports AMD64 on Linux with a 32-bit application. ;-)
4) Message boards : Feedback : AMD64 (Message 2298)
Posted 28 Sep 2006 by ebahapo
Post:
FWIW, running two instances of the client, one the 32-bit client, the other, the 64-bit client, on the same 4-core system, but limiting each client to 2 cores, I can compare the relative performance of 32-bit and 64-bit SIMAP's HMMER: the 64-bit version is about 7% faster.

By enabling vectorization (supported by default on AMD64), the SIMAP developers observed other 8% improvement.

Bottom line: porting the project application to AMD64 has the potential to improve performance by 15%!
5) Message boards : Feedback : AMD64 (Message 2294)
Posted 27 Sep 2006 by ebahapo
Post:
So far, two projects have added support for x86-64: SETI sends the x86 application and SIMAP, a true x86-64 application. Would you consider supporting the platform x86_64-pc-linux-gnu or x86_64-unknown-linux-gnu similarly?

Chess960 now supports AMD64 with a native Linux application too...

HashClash now supports AMD64 with a 32-bit Linux application...

HashClash now supports AMD64 on Windows with a 32-bit application...
6) Message boards : Feedback : AMD64 (Message 2289)
Posted 26 Sep 2006 by ebahapo
Post:
I have a 4P AMD64 system running both the 32-bit and the 64-bit core clients, with 2P each....

I wish I had SIMAP results too, but the AMD64 system is a production server whose kernel version is deemed too old by the AMD64 application. The advantage is that SIMAP WUs have a fixed amount of processing each, making comparisons straight-forward.

Now that SIMAP is releasing WUs for the HMMER application, I can compare the performance of the 32-bit and the 64-bit applications. Again, the 64-bit application. running on the same server, is about 7% faster.

They used GCC 3.3.3 though. Had they used GCC 4, they could have enabled vectorization for the 64-bit application and the difference would be even greater. All that using a freely available compiler with no hoops to go through...
7) Message boards : Feedback : AMD64 (Message 2277)
Posted 20 Sep 2006 by ebahapo
Post:
So far, two projects have added support for x86-64: SETI sends the x86 application and SIMAP, a true x86-64 application. Would you consider supporting the platform x86_64-pc-linux-gnu or x86_64-unknown-linux-gnu similarly?

Chess960 now supports AMD64 with a native Linux application too...

HashClash now supports AMD64 with a 32-bit Linux application...
8) Message boards : Feedback : AMD64 (Message 2194)
Posted 19 Aug 2006 by ebahapo
Post:
Martin,

I did the AMD64 port of the SETI Classic application. Right off the bat, it was about 15% faster. Later, as all AMD64 processors support SSE, I added some vectorized code which improved it by other 10%. Although it was too late to add it to SETI Classic, it became the basis for some experiments at SETI Beta before FFTW was used instead. But now that even GCC can generate vectorized code, taking advantage of SSE is even more attractive. And on AMD64 it's not necessary to go through hoops to go around the lack of SSE support.

More recent Linux distros and Windows also provide faster and more accurate math routines for AMD64.

Therefore, there's a potential performance improvement that can be achieved by just recompiling the application for AMD64.

BOINC 5.5 is said to consider CPU capabilities and if AMD64 support is one of them then even a 32-bit core client would be able to run AMD64 applications.

I have a 4P AMD64 system running both the 32-bit and the 64-bit core clients, with 2P each. As you can see, the one running the 64-bit Chess960 application is considerably faster than the other.

I wish I had SIMAP results too, but the AMD64 system is a production server whose kernel version is deemed too old by the AMD64 application. The advantage is that SIMAP WUs have a fixed amount of processing each, making comparisons straight-forward.
9) Message boards : Current tests : Forum code update. (Message 2188)
Posted 19 Aug 2006 by ebahapo
Post:
It seems that the rating of posts is kaput...

HTH
10) Message boards : Feedback : AMD64 (Message 2187)
Posted 19 Aug 2006 by ebahapo
Post:
So far, two projects have added support for x86-64: SETI sends the x86 application and SIMAP, a true x86-64 application. Would you consider supporting the platform x86_64-pc-linux-gnu or x86_64-unknown-linux-gnu similarly?

Chess960 now supports AMD64 with a native Linux application too...
11) Message boards : Feedback : AMD64 (Message 1913)
Posted 5 Aug 2006 by ebahapo
Post:
So far, two projects have added support for x86-64: SETI sends the x86 application and SIMAP, a true x86-64 application. Would you consider supporting the platform x86_64-pc-linux-gnu or x86_64-unknown-linux-gnu similarly?
12) Message boards : Current tests : WinXP 64bit/AMD64bit-Support? (Message 704)
Posted 27 Feb 2006 by ebahapo
Post:
I don't know the application at all, but the AMD64 ABIs use the SSE registers to pass arguments around. Moreover, many math transcendentals are implemented in routines which are faster and more precise than the equivalent x87 instructions.

HTH
13) Message boards : Feedback : AMD64 (Message 95)
Posted 17 Feb 2006 by ebahapo
Post:
Any plans to investigate a 64-bit application as S@H Beta has?

Thanks.






©2024 University of Washington
http://www.bakerlab.org