similar to: [LLVMdev] Statistics for 'switch' instruction.

Hi Stepan, These are quite interesting numbers. Thanks for running this experiment! Do you have some numbers on the difference of memory consumption and compile time (at -O0 and -O2) with and without case ranges? I think we need to see those numbers before we can decide wether it's interesting to commit to case ranges or not. Thanks, Nuno Quoting Stepan Dyatkovskiy <stpworld at

Hi, You cannot call removeModule on a JIT with no modules: jitstate will be 0 and therefore we have a null-pointer exception. The function returns a boolean for success/failure, however, so you would expect to be able to call it and get false back. Should we be checking for jitstate != 0 before accessing the variable? - if (jitstate->getModule() == M) { + if (jitstate &&

Hello, I managed to write my own pass =) and now I want it to be executed by the LLVM-JIT but I don't see the right way to do it. The EngineBuilder only allows to set an opt level and as far as I can see the ExecutionEngine does not provide access to the jitstate which holds its passmanager. Could anyone give me a hint what I have to do to add my pass to the passmanager of the jit? Thank

Hi all, I can't find a way to add a MachineFunctionPass to the common codegen passes (LLVMTargetMachine::addPassesToEmitMachineCode) while JITting (the pass manager is associated with the jitstate of the JIT and I can't access it because it's private). Have I missed something? Or adding a MachineFunctionPass to codegen requires to change the

hi list, I've got a situation where I'm constantly adding and removing Modules from an ExecutionEngine. When removing a module and the jitstate is freed, clean up of the StructLayoutMap ends up trying to access a Type instance was was already freed by the removal of another Module. In trying to debug this, it basically comes down to a Type instance serving as key in

Hi Andreas, you can run any set of passes you like on your module before handing it to the JIT. Ciao, Duncan. > I managed to write my own pass =) and now I want it to be executed by the LLVM-JIT but I don't see the right way to do it. > The EngineBuilder only allows to set an opt level and as far as I can see the ExecutionEngine does not provide access to the jitstate which holds its

Anybody else having problems accessing this site? Haven''t had much luck over the past couple of days. -- Posted via http://www.ruby-forum.com/.

Hi, I have Installed GSU 6.4 under wine it works but at dead slow space. No Crashes are observed. Save Load functions OK 2d Gfx save OK No functionality glitches are observed on Limited Test Conducted. Kubuntu, wine 1.1.2, nvidia ge4 440 GO Latest Driver from nvidia. OpenGL works fine as far as Linux desktop is concerned. When I run Sketchup the CPU utilization is almost 100%. On same machine

Evan, So, if I understand you correctly, the design you have in mind is to: create a PassManager, pass it to the JIT on construction, and modify runJITOnFunction to run the second PassManager on the Function being jit'd before running the codegen PassManager. Thanks. Tom ----- Original Message ----- From: "Evan Cheng" <evan.cheng at apple.com> To: "LLVM Developers Mailing

Evan, My overall goal is to support dynamic optimization in LLVM. In order to do so, I must gather profiling information at runtime, then recompile the profiled functions. Currently, I'm just adding and removing calls into my profiler in a custom pass. What is the advantage of giving the JIT a second profile manager over my current implementation? Thanks. Tom ----- Original Message -----

----- "Evan Cheng" <evan.cheng at apple.com> wrote: > On Sep 16, 2008, at 8:44 AM, Thomas B. Jablin wrote: > > > Evan, > > So, if I understand you correctly, the design you have in mind is > > to: create a PassManager, pass it to the JIT on construction, and > > modify runJITOnFunction to run the second PassManager on the > > Function

Hi Bert, I am sorry to bother you on weekend. I am still struggling on defining a correct function. I first defined the function RFS (see below), then run it by provide the two argument. m52.2cluster <-RFS(inputfile =allinfo_m52, N=2 ) I do not get error message, but no figure displays on screen. I do not know what is going on. Can you help me a little more on this issue? Thank you,

Hi Richard, Thank you so much!! I understand the problem now, I assign a name to the "ggsurvplot" object and then add print(fig) at bottom of function definition, now figure gets printed on screen. Ding # function to generate RFS curves RFS <- function( inputfile, N ) { cluster<- survfit(Surv(RFS_days2, OV_Had_a_Recurrence_CODE) ~ clusters, data =

FAQ 7.22 You must print a ggplot object, for example with print(m52.2cluster) For the FAQ, run the line system.file("../../doc/FAQ") in R on your computer. Open up the resulting filepath in your favorite editor and scroll down to 7.22 On Sun, Jan 14, 2018 at 4:21 PM, Ding, Yuan Chun <ycding at coh.org> wrote: > Hi Bert, > > I am sorry to bother you on weekend. >

That is certainly OK, but you can also just use print(ggsurvplot(...)) as your final statement. out <- RFS( ...) would then return the ggsurvplot object *and* graph it. Any good R tutorial or a web search will provide more details on function returns, which you might find useful. Cheers, Bert Bert Gunter "The trouble with having an open mind is that people keep coming along and

Will MultiMaster dsync work fine or will be big problems ? I configured dovecot to dsync from serv1 to serv2 to serv3 and repeat to serv1 to serv2 to serv3 and repeat... When email received to serv2, it will be synced to serv3, and serv3 will sync it to serv1, and serv1 will sync it to serv2 and will stop here because emails matched. What problems can be with this kind of configuration ?

Thank you, your suggestion is simpler and logically better. I had impression that the last object in a function gets returned, so I did not add the print function at the bottom line of the function definition. Returning an object and graph the object are different process, I am a beginner for writing R function and need to find a good guide source about writing R functions. If you know a good

My external node classifier returns some arrays in the list of parameters. Example output (names have been changed to protect the innocent): $ ./node_classifier a.b.com --- %YAML:1.0 "classes": ["class1", "class2", "class3"] "parameters": "hostname": "a" "name": "a" "domain":

HI R users, I construct dendrogram tree and cut the tree into different clusters, then generate survival curves by the following three functions. All variables are included in an inputfile. Can you help me to consolidate the following three function into one functions? I thought about using if else function, but not sure how to do it. Thank you, Ding # function to generate RFS RFS2cluster

R experts, I have three columns. c1 and c2 are numeric variables whereas c3 are the clusters classes (nominal variable, 10 different: cluster1, cluster2, cluster3, cluster4, cluster5 ....). I'd like to plot c1 against c2 (easy!) in a 2D plot and put different color depending to the cluster class automatically regardless the number of clusters. Could anyone give a hand? Josep Maria, matrix