Tuesday, May 29, 2007


Roelofs et al. derive parallaxes for a set of AM CVn binaries and predict LISA GW signals.

Dietrich et al. describe BLOX, a program to select galaxy clusters based on optical, X-ray and weak lensing shear.

Percival et al. measure the baryon oscillation signal at z=0.2 and z=0.35 using SDSS and 2dFGRS to constrain w=-1.004 +/- 0.089 for a flat model including the supernova and CMB constraints.

Wallin et al. place limits on deviations from Newtonian gravity in the outer Solar System through observations of the orbits of TNOs.

Markevitch points out that X-ray emission from clusters of galaxies depends on the He abundance in their cores which is unknown (and may be enhanced due to sedimentation). Combining SZ measurements with distances determined using other techniques can be used to estimate the He abundance.

Ruiter et al. calculate the LISA foreground due to mass-transferring binaries and find that it becomes significant at frequencies above ~7 mHz.

Bickel et al. develop methods for searching for periodicity in photon event times with application in the gamma-ray regime.

Prix & Whelan apply F-statistic methods developed for LIGO to the MLDC binary challenges.

srcut and sresc seg fault in xspec v12

Heather Matheson at U.Manitoba reported that srcut seg faults in v12. This turned out to be due to an error in the C++ to Fortran conversion for the power-law routine that is called by srcut and sresc.

Tuesday, May 22, 2007

another seg fault in compps

I had a slight error in the dynamic memory allocation for the compps model which caused a potential seg fault if the response had < 88 energy bins. This is patched in both v11 and v12.

Thursday, May 17, 2007


Jee et al. use weak-lensing analysis of Cl0024+17 and find a ring of dark matter which they interpret as the result of a high-velocity line-of-sight collision.

Tsalmantza et al. present the first results from a program to generate a library of synthetic galaxy spectra to be used to classify the several million unresolved galaxies whose spectra will be acquired by Gaia.

Wu & Wang calculate Fe K line shapes from relativistic accreting disks including the effects of self-shadowing.

Tornatore et al. use hydro simulations to investigate chemical evolution in clusters of galaxies.

Trimble et al. summarize astrophysics in 2006.

Padmanabhan reviews dark energy cosmology, argues that only a cosmological constant makes sense and that the problem cannot be satisfactorily solved at our current level of (mis)understanding.

Fiore et al. combine optical, IR, and X-ray observations of the Chandra Deep Field South to identify the Compton-thick AGN that have so far been missing from deep Chandra observations.

Kelly describes a Bayesian approach using Gaussian mixtures for linear regression in the presence of measurement errors, non-detections, and selection effects.

Eckert et al. analyze a 1.1Ms Integral ISGRI observation of the Coma cluster to find an extended hard X-ray excess in the direction of the infalling cluster A1376.

Sanders & Fabian continue their analysis of the very deep Chandra observation of the Perseus cluster finding a) the ripples interpreted as sound waves carry a significant fraction of the energy lost to radiative cooling b) the distribution of metals peaks at 40 kpc and is clumpy on scales of 5 kpc c) there is no evidence for thermal emission from the radio bubbles.

Repin et al. give analytic approximations for the relativistic Fe K line shape from accretion disks. A table is available.

Friday, May 04, 2007


Jeremy Sanders provides a faster version of the gsmooth model which precalculates an array of erf values then interpolates on them.

Thursday, May 03, 2007


Stanek proposes a an alternative history version of astro-ph for failed telescope and other proposals. Probably should have come out on Apr 1.

Feroz & Hobson describe an efficient sampling method for multimodel and/or degenerate distributions when calculating Bayesian evidence.

Hallman et al. use numerical simulations to show that using beta-model fits to X-ray data in the analysis of SZE observations leads to systematic biases and suggest a simple correction.

Miller reviews relativistic X-ray lines from the inner regions of accretion disks.

Magliocchetti & Bruggen combine X-ray data from the REFLEX and NORAS samples and radio data from the NVSS to investigate the effects of radio sources on the cluster L_X-T relation.

Kitaura & Ensslin survey Bayesian reconstruction schemes for large-scale structure.

Kerkyacharian et al. present a new estimation method for the inverse problem building on SVD decomposition.

Wednesday, May 02, 2007

building with gfortran

Successfully built headas on my Mac using the gfortran from the installer supplied at gcc.gnu.org. This places the gfortran library files in /usr/local/gfortran/lib so configure has to be altered so that F77LIBS4C in hmakerc includes this directory. If the gfortran plug-in for Xcode is used then the gfortran library files end up in /usr/local/lib which is in the link path anyway so need not be specified in F77LIBS4C. Also had to update cctools to avoid an undefined dso_handle. Had to add DHAVE_ETIME=1 to DEFS in hmakerc to ensure that etime is not defined again in xanlib/sysdep/sysc.c.

Not directly connected to gfortran but we discovered cod.dhf was not being built - this requires running hlpdhf on cod.hlp which means ensuring ia/doc is built ahead of qdp.

Tuesday, May 01, 2007

udmget : Too many blocks error

Both Roderick Johnstone and Ewan O'Sullivan have reported getting "Too many blocks" errors when using xspec v11 with complicated models. MAXBLK is set to 100 in udmget.c and there is no good reason why this should be so small. I've changed this to 1000 for the next release which should solve these problems.

Note that Bryan has updated xspec/src/tcl/Makefile so that it will rebuild libxantcl* when changes are made in the libxanlib* library. I really should be building libxantcl under the xanlib tree rather than the somewhat bizarre way it is done at present under the xspec tree.