• Non ci sono risultati.

Questa tesi ha avuto come principale scopo progettare un framework per l’implementazione di algoritmi di simulazione dinamica su complessi simpli- ciali su Graphical Processing Unit (GPU). La peculiarit´a dell’architettura delle GPU, pensata per velocizzare il processo di visualizzazione, pone molti ostacoli alla realizzazione di algoritmi e strutture dati General Purpose, sia per l’impiego di memoria specializzata, nei formati, nei metodi di allocazione e lettura e nelle dimensioni, per contenere immagini, sia per il paradigma di programmazione offerto. Questi tesi ha dimostrato come, utilizzando i com- plessi simpliciali come struttura di base per la rappresentazione geometrica sia possibile implementare algoritmi computazionalmente pesanti, quali quel- li usati per la modellazione real-time di oggetti deformabili, su GPU. I test sperimentali effettuati hanno dimostrato che l’utilizzo della GPU in questo contesto porta un effettivo guadagno in termini di tempo di calcolo rispetto ad una tradizionale implementazione su CPU.

Appendice A

Screenshot del simulatore

Figura A.1: Deformazione di un cilindro tramite l’applicativo sviluppato parallelamente alla trattazione in esame

APPENDICE A. SCREENSHOT DEL SIMULATORE 136

APPENDICE A. SCREENSHOT DEL SIMULATORE 137

Figura A.3: La rete usata per i test sui 1-complessi simpliciali in due dimensioni

Appendice B

Corrispondenza con la Nvidia

B.1

Bug nel compilatore

---Original Message---

From: Guido Ranzuglia [mailto:inopportuno@yahoo.com] Sent: Wednesday, February 09, 2005 11:35 AM To: GLSL-Support Subject: OGLSL linking -> Exception during compilation [supplement]

sorry the other mail doesn’t contain two little files in the shader directory...

The problematic shader is rungekuttafg.glsl.

euler.glsl is complementary at the rungekuttafg.glsl and can use to test how the application works.

APPENDICE B. CORRISPONDENZA CON LA NVIDIA 139 vtlookup.glsl and fglookup.glsl are the shader that implement the texture’s lookup in the vertex program...

Sorry....

Guido Ranzuglia

_______________________________________________

Hi Guido,

thanks for the report.

Exception messages in the logfile indicate a problem in the compiler and we need to know these to fix them.

I wasn’t really able to find a simple central download location of the VCG on SourceForge, but that was not necessary to reproduce the problem this time. (In the future please submit full projects. You never know if third party code we download from somewhere will be identical to what you tested. There shouldn’t be a problem with bigger attachments as long as they aren’t named .zip)

Nevermind, the good news is that I could reproduce the problem with the 71.80 drivers and the problem does not exist in a newer version I tried.

APPENDICE B. CORRISPONDENZA CON LA NVIDIA 140 That means the problem has been fixed internally and our QA is going to check if the next release already has the fix.

Detlef Roettger

---

Senior Systems Software Engineer Workstation Software Engineering NVIDIA GmbH, Germany

Hi Guido,

no problem, it said this:

>>

we’ve regressed the problem the next imminent driver release does not contain the fix. A driver from the next branch (versions 75+) has the necessary changes.

At this time I don’t know when that driver is going to be publicly available. <<

Sorry that I can’t help more, the problem was in the internal interface between driver and compiler and that was overhauled from release 70 to 75 drivers because of problems like you reported. There is no plan to backport this because rel70 had code freeze.

APPENDICE B. CORRISPONDENZA CON LA NVIDIA 141 Detlef

_____________________________________________

---Original Message---

From: Guido Ranzuglia [mailto:inopportuno@yahoo.com] Sent: Friday, January 14, 2005 5:38 AM To: Keith Galocy Subject: Re: Probably problem with glCopyTexSubImage2D

This is a little visual studio’s project with wich you may be able to reproduce my problem. I have already included it in the bug report but probably it isn’t arrived to you...

Thanks

Guido Ranzuglia

_____________________________________________

B.2

Bug nel driver

Guido,

We have confirmed the issue and cause of the problem. We’ll get it fixed asap.

APPENDICE B. CORRISPONDENZA CON LA NVIDIA 142 Thanks for your assistance, Keith

Bibliografia

[CD04] Henrik Wann Jensen Craig Donner, Faster gpu computations using adaptive refinement, Proceedings of ACM SIGGRAPH, Technical Sketches, 2004.

[DHH05] Tor Dokken, Trond R. Hagen, and Jon M. Hjelmervik, The gpu as a high performance computational resource, SCCG ’05: Procee- dings of the 21st spring conference on Computer graphics (New York, NY, USA), ACM Press, 2005, pp. 21–26.

[FHWZ04] Randy Fernando, Mark Harris, Matthias Wloka, and Cyril Zel- ler, Programming graphics hardware, Bello, The Eurographics Association, 2004.

[FSH04] K. Fatahalian, J. Sugerman, and P. Hanrahan, Understanding the efficiency of gpu algorithms for matrix-matrix multiplication, TITOl, 2004.

[GEW05] Joachim Georgii, Florian Echtler, and R¨udiger Westermann, In- teractive simulation of deformable bodies on gpus, Simulation and Visualisation 2005, 2005.

[HBSL03] Mark J. Harris, William V. Baxter, Thorsten Scheuermann, and Anselmo Lastra, Simulation of cloud dynamics on gra-

BIBLIOGRAFIA 144 phics hardware, HWWS ’03: Proceedings of the ACM SIGGRA- PH/EUROGRAPHICS conference on Graphics hardware (Aire- la-Ville, Switzerland, Switzerland), Eurographics Association, 2003, pp. 92–101.

[JKW04] P. Kipfer J. Krueger, T. Schiwietz and R. Westermann, Nume- rical simulations on pc graphics hardware, ParSim 2004 (Special Session of EuroPVM/MPI 2004), 2004.

[KF05] Emmett Kilgariff and Randima Fernando, The geforce6 series gpu architercture, ch. 30, pp. 471–491, Addison Wesley, 2005. [KKW05] Polina Kondratieva, Jens Kr¨uger, and R¨udiger Westermann,

The application of gpu particle tracing to diffusion tensor field visualization, Proceedings IEEE Visualization 2005, 2005. [KRD+03] Ujval J. Kapasi, Scott Rixner, William J. Dally, Brucek Khailany,

Jung Ho Ahn, Peter Mattson, and John D. Owens, Programmable stream processors, IEEE Computer (2003), 54–62.

[KSW04] Peter Kipfer, Mark Segal, and R¨udiger Westermann, Uberflow: a gpu-based particle engine, HWWS ’04: Proceedings of the ACM SIGGRAPH/EUROGRAPHICS conference on Graphics hardware (New York, NY, USA), ACM Press, 2004, pp. 115–122. [LBM+05] Aaron Lefohn, Ian Buck, Patrick McCormick, John D. Owens, Tim Purcell, and Robert Strzodka, Gpgpu : General- purpose computation on graphics processors, Tutorial in IEEE Visualization 2005, 2005.

[Moo65] Gordon Moore, Cramming more components onto integrated circuits, Electronics (1965), no. 38, 114–117.

BIBLIOGRAFIA 145 [OLG+05] John D. Owens, David Luebke, Naga Govindaraju, Mark Harris, Jens Krger, Aaron E. Lefohn, Timothy, and J. Purcell, A sur- vey of general-purpose computation on graphics hardware, The Eurographics Association, 2005.

[Owe02] John D. Owens, Computer graphics on a stream architecture, Ph.D. thesis, Stanford University, November 2002.

[Owe05] John Owens, Streaming architectures and tecnology trends, ch. 29, pp. 457–470, Addison Wesley, 2005.

[PBMH02] Timothy J. Purcell, Ian Buck, William R. Mark, and Pat Han- rahan, Ray tracing on programmable graphics hardware, ACM Transactions on Graphics 21 (2002), no. 3, 703–712, ISSN 0730-0301 (Proceedings of ACM SIGGRAPH 2002).

Documenti correlati