| bio | website | |
|---|---|---|
| location | ||
| age | ||
| visits | member for | 10 months |
| seen | Apr 23 at 13:36 | |
| stats | profile views | 47 |
|
Mar 21 |
comment |
How to have Mathematica find asymptotics with correct asumptions So why does Evaluate[Gamma[-2]/Gamma[-1]] in Mathematica give complex infinity? How could I make Mathematica spit out a complex number instead? |
|
Mar 21 |
comment |
How to have Mathematica find asymptotics with correct asumptions @whuber thanks. I was looking for $x\to\infty$ limit rather than $x\to 0$. Could you give me some more info about what your first line is doing here? as for your second line are you suggesting that this ratio is infact quite well defined? |
|
Mar 21 |
asked | How to have Mathematica find asymptotics with correct asumptions |
|
Mar 10 |
revised |
Launching remote slave kernels on HPC causes slow down added 921 characters in body |
|
Mar 10 |
comment |
Launching remote slave kernels on HPC causes slow down You are probably right. What a shame if this is the case, I could have lived with it being a little bit slower, but not this much slower. |
|
Mar 10 |
comment |
Launching remote slave kernels on HPC causes slow down @Ajasja I haven't tried it, but it can't hurt so I will give that a go next, thanks. |
|
Mar 9 |
comment |
Mathematica Parellelization on HPC @OleksandrR. I think you are right MathLink does seem to be dying for me with 190 Kernels. Despite getting there in the end it now takes vastly longer than when I had 19 single MathKernels with 10 local slaves each. |
|
Mar 9 |
revised |
Launching remote slave kernels on HPC causes slow down added 104 characters in body |
|
Mar 9 |
asked | Launching remote slave kernels on HPC causes slow down |
|
Mar 6 |
comment |
Efficient way to utilise Parallel features to make use of many cores @Szabolcs thanks, this is going to save me a lot of time. |
|
Mar 6 |
accepted | Efficient way to utilise Parallel features to make use of many cores |
|
Mar 6 |
comment |
Efficient way to utilise Parallel features to make use of many cores the matrix of pairs and ParallelMap might just be the way to go, thanks. |
|
Mar 6 |
comment |
Efficient way to utilise Parallel features to make use of many cores @Szabolcs Yep, that is the problem. Is there no way around this then? |
|
Mar 5 |
comment |
Mathematica Parellelization on HPC @Szabolcs Hi, I finally got this working. I don't know much about Package writing, so have just placed my needed Kernel Launch commands in job.nb along with parsing commands for hostfile. If I try to use your package as a template then add it to my notebook with Needs things no longer work for some reason. I don't know much about package writing so I am prob doing something stupid there. If you like however I can make a package from my notebook and upload to bitbucket and perhaps you could take a look so we could publish it there for others using PBSpro in the future? |
|
Mar 5 |
asked | Efficient way to utilise Parallel features to make use of many cores |
|
Feb 28 |
comment |
Mathematica Parellelization on HPC @Szabolcs totally baffled as to why it's giving bash: math: command not found when I am passing the absolute path to math everywhere I can in CRC.m. Also even from the terminal simply doing ssh node055 /cm/shared/apps/Mathematica_8.0.4/bin/math gives the familiar Input[] but then times out after 15 secs. |
|
Feb 27 |
comment |
Mathematica Parellelization on HPC I defined ssh="/usr/bin/ssh" already in place of your rsh` def. |
|
Feb 27 |
comment |
Mathematica Parellelization on HPC I do think I understand it reasonably well now, although I am a little rusty on some Mathematica syntax to do with string joins etc. How is; SubKernels`RemoteKernels`RemoteMachine[
host,
ssh <> " -x -f -n `1` \"" <> "/cm/shared/apps/Mathematica_8.0.4/bin/math" <>
" -mathlink -linkmode Connect `4` -linkname '`2`' -subkernel -noinit\"",
16]? On these HPC math is not an alias and full path is needed- the error I now get is bash: math: command not found
KernelObject::rdead: Subkernel connected through remote[node053] appears dead. (with or without abs path in above) |
|
Feb 27 |
comment |
Mathematica Parellelization on HPCParallelTable::nopar: No parallel kernels available; proceeding with sequential evaluation. is what I get now. With my hostfile in the above format, I changed things in CRC.m to be like hosts = Import[hostfile, "List"];
$ConfiguredKernels = Join[$ConfiguredKernels, CRCKernel @@@ Rest[hosts]] and changed CRCKernel to CRCKernel[host_String]:=
SubKernels`RemoteKernels`RemoteMachine[
host,
ssh -x -f <> " -n `1` \"" <> math <>
" -mathlink -linkmode Connect `4` -linkname '`2`' -subkernel -noinit\"",
16] as well as the original definition. |
|
Feb 27 |
comment |
Mathematica Parellelization on HPC Ah, it just contains one column with node053
node055,..... the number of cores is always 16 though, so I can put that in manually by just removing the second arg of CRC. |