Understanding Nvidia GT200 GPU and CUDA implementation microbenchmarks!

Hi,
are you just wanting to know GPU GT200 hardware architecture see:

GeForce GTX 200 Technical Brief
todo: poster benchmark gt200 pi

NVIDIA GT200 GPU and Architecture Analysis (beyond3d)

NVIDIA's GT200: Inside a Parallel Processor By: David Kanter

Read PTX Nvidia spec..
Search for decuda for cubin assemly and diassembly from/to PTX..

Understanding both hardware and CUDA using microbenchmarks:
"Benchmarking GPUs to tune dense linear algebra" SC08

slides:

This Volkovs page has matmul and other blas codes, factorization (LAPACK) codes and FFT codes:


Also "Micro-benchmarking the GT200 GPU" by
Misel-Myrto Papadopoulou Maryam Sadooghi-Alvandi Henry Wong

For testing:
TODO:search code CUDA Forums of instruction throughput code

And for having addc and other integer functionality from nvcc see patched nvcc (now with sources):
http://www.mpi-inf.mpg.de/~emeliyan/cuda-compiler/

Of course emulators:
1.GPUOcelot paper
search: http://code.google.com/p/gpuocelot/wiki/References
2.Barra "Barra, a Parallel Functional GPGPU Simulator"
search Publications: http://gpgpu.univ-perp.fr/index.php/Barra
3. GPGPU-sim paper: http://www.ece.ubc.ca/~aamodt/papers/gpgpusim.ispass09.pdf
more info:http://www.ece.ubc.ca/~aamodt/gpgpu-sim/

of course
GT300 hardware info
and
graphics info (?)