Galaxies can form in a sufficiently deep gravitational potential so that efficient gas cooling occurs. We estimate that such potential is provided by a halo of mass $M \gtsim M_{c} \approx 7.0 \times 10^{12} ~ (\Delta_{c}(z) (1+z)^{3})^{-1/2} \Msun$, where $\Delta_{c}(z)$ is the mean overdensity of spherically virialized objects formed at redshift $z$, and $M_{c} \approx 4.0 \times 10^{11} \Msun$ at $z = 0$. Based on this criterion, our galaxy samples are constructed from cosmology simulation data by using HiFOF to select subhalos in those FOF halos that are more massive than $M_{c}$. There are far more dark subhalos than galaxy-hosting subhalos. Several tests against observations have been performed to examine our galaxy samples, including the differential galaxy mass functions, the galaxy space density, the projected two point correlation functions (CF), the HODs, and the kinematic pair fractions. These tests show good agreements. Based on the consistency with observations, our galaxy sample is believed to correctly represent galaxies in real universe, and can be used to study other unexplored galaxy properties.

In this paper, we describe the architecture and performance of the GraCCA system, a Graphic-Card Cluster for Astrophysics simulations. It consists of 16 nodes, with each node equipped with 2 modern graphic cards, the NVIDIA GeForce 8800 GTX. This computing cluster provides a theoretical performance of 16.2 TFLOPS. To demonstrate its performance in astrophysics computation, we have implemented a parallel direct N-body simulation program with shared time-step algorithm in this system. Our system achieves a measured performance of 7.1 TFLOPS and a parallel efficiency of 90% for simulating a globular cluster of 1024K particles. In comparing with the GRAPE-6A cluster at RIT (Rochester Institute of Technology), the GraCCA system achieves a more than twice higher measured speed and an even higher performance-per-dollar ratio. Moreover, our system can handle up to 320M particles and can serve as a general-purpose computing cluster for a wide range of astrophysics problems.

The technique of weak-lensing aperture mass densitometry, so called the zeta-statistic, has recently been popular in actual observations for measurement of individual cluster mass. It has however been anticipated that the line-of-sight projection by foreground and background matter can adversely affect the cluster mass determination with not only substantial error dispersion but also a sizable positive systematic bias. Additionally, the finite number of background galaxies even at a reasonable observing depth can also introduce Poisson noise to the mass estimate. In this paper, we quantitatively investigate the degree of errors separately contributed by the two sources to the mass determination of those galaxy clusters with M_{200}>10^{14}M_{\odot}. We find that the aperture mass of zeta-statistic turns out to be a mass estimator of much reduced systematic bias, due to the cancellation by the positively biased local background mass sheet. However, the error dispersion of M_{200} arising from both projection effect and Poisson noise is found to be still sizable (40%-90%), even for the shear-selected, clean sample where multiple clusters located within a suitable projected aperture are removed. We also investigate how to remedy this large-error problem in weak lensing measurements, and propose a plausible alternative mass estimator, M(<theta_{1000}), an aperture mass measured within about half the virial radius. The aperture mass M(<theta_{1000}) is free of bias and has a substantially reduced error dispersion, 39% for the worst case of high-z, low-mass clusters, that can be smaller than the error dispersion of M_{200} as much as a factor 3.