similar to: [LLVMdev] Performance degradation when repeatedly exchanging JITted functions

On 3/6/12 5:28 PM, Joerg Sonnenberger wrote: > Advantage of using the latest address: one saved jmp per call. Per newly JITted call ;) > Advantage of using the initial address: easier G/C of intermediate > versions, less things to keep track of. I still think both versions require larger changes. When using the latest address, you have to keep track of all JITted functions per

On 3/6/12 6:48 PM, James Molloy wrote: >> I don't think that a patch implementing any of those approaches would be >> accepted, that's why I am tending towards implementing it outside of LLVM. > > Why not? If they make LLVM better and aren't hacks, why wouldn't they be accepted? Okay, that motivated me to work on the patch again. I think I found a compromise of

On Tue, Mar 06, 2012 at 04:29:28PM +0100, Clemens Hammacher wrote: > I think a solution would be to always call a function through it's > stub, so that there is a single location to update when the function > is exchanged. This would mean that there is always exactly one level > of indirection, which is worse for programs that don't exchange > functions at runtime, but is

On Tue, Mar 06, 2012 at 04:09:36PM +0000, James Molloy wrote: > Surely you need to patch *all* functions, not just the initial? Depends on whether you always link to the original address or not. If you use link with the latest address, you have to patch all versions to point to the latest, otherwise you can just patch the first. Advantage of using the latest address: one saved jmp per call.

Surely you need to patch *all* functions, not just the initial? The point is with the current solution no matter which version of the function another function is linked to, it will hit a sled of JMPs and eventually end up at the newest. If you only patched the first, that sled wouldn't work. So you'd have to patch all instances. That still shouldn't be too hard. Cheers, James

> I don't think that a patch implementing any of those approaches would be > accepted, that's why I am tending towards implementing it outside of LLVM. Why not? If they make LLVM better and aren't hacks, why wouldn't they be accepted? ________________________________________ From: llvmdev-bounces at cs.uiuc.edu [llvmdev-bounces at cs.uiuc.edu] On Behalf Of Clemens Hammacher

Hi All, Another nls related problem (for background, I'm migrating a complicated modelling package from S-plus to R). Below I've reduced this to the minimum necessary to demonstrate my problem (I think); the real situation is more complicated. Two similar selfStart functions, ssA and ssB. The 'initial' function for ssB modifies its arguments a little and then calls getInital

Hi Everyone, I'd omitted the non-optional 'parameters' argument to selfStart. Making this change to SSbatch gives the same (successful) result from the two calls to nls. SSbatch<-selfStart( model=function(Batch, Coeffs) { Coeffs[Batch] } ,initial=function(mCall, data, LHS) { # Estimate coefficients as mean of each batch xy <- sortedXyData(mCall[["Batch"]],

Last week I posted a question concerning the mCall function, which is used to create self-starting functions and is described in the book by Pinheiro, J.C. and Bates, D.M. (Mixed-effects models in S and S-PLUS). On page 345 one finds the following call: xy<-sortedXyData(mCall[["x"]], LHS,data) It is necessary to replace the "x" in the call to mCall by the actual

Hello, I'm trying to do some non-linear regression with 2 cell types and 4 tissue type treatments using selfStart models Following Ritz and Streibig (2009), I wrote the following routines: ##Selfstart expDecayAndConstantInflowModel <- function(Tb0, time, aL, aN, T0){ exp(-time*aL)*(T0*aL+(-1+exp(time * aL))*Tb0 * aN)/aL } expDecayAndConstantInflowModelInit <- function(mCall, LHS,

Hi, I have great problems with my work in R. I look for to model the growth of fish. I have "Longitudinal data", a serie of repeated measures for each individual. Using the corresponding packages "nlme" in R. I treat to fit to the data different growth functions, wich were entered by me. Unfortunately for no it was arrived at the convergence, several error messages appeared. I

Following on my posting of this morning, concerning a problem that I am having constructing a self-starting function for use with nls (and eventually with nlsList and nlme), the following is the self-starting function called NRhyperbola: > NRhyperbola function (Irr,theta,Am,alpha,Rd) { # Am is the maximum gross photosynthetic rate # Rd is the dark resiration rate (positive value) #

Hi Everyone, I seek guidance to avoid wasting a lot of time and doing things badly. Several times I've solved my problems, only to find that my solutions were clumsy and not robust. (see "nested" getInitial calls; variable scoping problems: Solved?? http://finzi.psych.upenn.edu/R/Rhelp02a/archive/139943.html for one truly horrible approach). I'm sure that I'm not the

On 05/30/2016 09:07 AM, Dominique Ramaekers wrote: >> root@host_a:~# virsh migrate --verbose --p2p --copy-storage-all --persistent -- >> change-protection --abort-on-error --undefinesource --live domain >> qemu+ssh://root@host_b/system --xml domain.ceph.xml > > Weird: The domain should be persistent Well, the domain is persistent. But the changes i did to domain.ceph.xml

> -----Oorspronkelijk bericht----- > Van: libvirt-users-bounces@redhat.com [mailto:libvirt-users- > bounces@redhat.com] Namens Björn Lässig > Verzonden: vrijdag 27 mei 2016 10:10 > Aan: libvirt-users@redhat.com > Onderwerp: [libvirt-users] migrate local storage to ceph | exchanging the > storage system > > TLDR: Why is virsh migrate --persistent --live domain >

Hello. Following pages 342-347 of Pinheiro & Bates, I am trying to write a self-starting nonlinear function (a non-rectagular hyperbola) to be used in nonlinear least squares regression (and eventually for a mixed model). When I use the getInitial function for my self-starting function I get the following error message: > getInitial(photo~NRhyperbola(Irr,theta,Am,alpha,Rd),dat) Error

Hi, I have two zoo objects (time series of same frequency) say A and B. I need to exchange A's dates for B's. Is there and easy way? I managed to do it by converting A to a vector and pasting it on a zoo vector (all ones) with B's dates but I wonder if there is an easier way by some "zoo" command I have overlooked... Thanks in advance, Costas

Hi, is there any way to exchange (A LOT OF) data between SPPS and R/Splus, i.e., I want to get data from SPSS into R. It is a huge data frame with many columns being factors, so manual conversion is out of question (and there are more data to come). Exporting ASCII from SPSS seems to loose all level names, and the portable SPSS format doesn't look too nice. I've searched S-news and

Hi Rick Thanks to Dieter Menne I did manage to solve the problem of imposing bounds on the parameter space duirng an nlsList fit. He suggested using optim to optimize the parameters prior to each fit. This worked well for me as I had a customized selfStart function that then optimized the parameters for each individual separately. if you rewrote your selfStart function as: powrDpltInit <-

Did somebody got to put 1.000 printing queues in a Linux+Samba+Cups production server without degradation? Thanks for any reply, Bruno Gomes Pessanha