I've recently had to re-setup a misplaced configuration of CUDA on our Tesla based Ubuntu server.
Given we have an Ubuntu 10.04.4 LTS system I've downloaded the latest stable Cuda Toolkit. That is the release 4.1.
Then a colleague invited me to install SOAP3 and the installation process went smoothly.
Then I left the machine to my colleague. So he started preparing the working enviroment and the let me know that:
works ok, the second command, instead, as an issue.
Indeed in /usr/local/cuda/lib there is only libcudart.so.4
SOAP3 website explicitly says: "SOAP3 was developed under the 64-bit linux platform and the CUDA Driver version 3.2."
So I hoped making a simbolic link would work but this gave a worst result.
Tried to write down an email to the SOAP3 staff address provided but their mailserver gave me an error
Googling has given me nothing to work on.
Do you have any ideas other than removing CUDA 4.1 and try to get working the 3.2 release ?
TIA
Given we have an Ubuntu 10.04.4 LTS system I've downloaded the latest stable Cuda Toolkit. That is the release 4.1.
Then a colleague invited me to install SOAP3 and the installation process went smoothly.
Then I left the machine to my colleague. So he started preparing the working enviroment and the let me know that:
Code:
$ /opt/soap3_0.01beta/2bwt-builder /biodata/arab/TAIR9_chr_all.fas
Code:
$ /opt/soap3_0.01beta/BGS-Build TAIR9_chr_all.fas.index /opt/soap3_0.01beta/BGS-Build: error while loading shared libraries: libcudart.so.3: cannot open shared object file: No such file or directory
SOAP3 website explicitly says: "SOAP3 was developed under the 64-bit linux platform and the CUDA Driver version 3.2."
So I hoped making a simbolic link would work but this gave a worst result.
Tried to write down an email to the SOAP3 staff address provided but their mailserver gave me an error
Failure to deliver the message to the recipient or group
Restrict delivery to this address can not pass your e-mail.
Restrict delivery to this address can not pass your e-mail.
Googling has given me nothing to work on.
Do you have any ideas other than removing CUDA 4.1 and try to get working the 3.2 release ?
TIA
Comment