vBulletin Search Engine Optimization
| |||||||
| Register | FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Hi, I work for a group that primarily uses our servers for statistical analysis. We mostly use SAS, Stata, Matlab, et al. We're trying to plan for upgrades to our existing older servers (v880s, 440s) and I am wondering if anyone out there has any experience with the T2000 series as replacements for systems that pull this kind of duty. If so, how does the T2000 compare in performance? tia, -y (woof) |
| |||
| dogdogdog <young.sul@gmail.com> wrote: > If so, how does the T2000 compare in performance? Is floating point performance important? If it is, wait for the new SPARC chips with a FP unit per core rather than one shared between all the cores. -- Robert A Heinlein: Theology is never any help; it is searching in a dark cellar at midnight for a black cat that isn't there. |
| |||
| Words to the wise, dogdogdog <young.sul@gmail.com> wrote: >Hi, >I work for a group that primarily uses our servers for statistical >analysis. We mostly use SAS, Stata, Matlab, et al. > >We're trying to plan for upgrades to our existing older servers >(v880s, 440s) and I am wondering if anyone out there has any >experience with the T2000 series as replacements for systems that pull >this kind of duty. > >If so, how does the T2000 compare in performance? The T2000 is not meant to server as number cruncher. Its more of a lightweight machine, suitable for lots of concurrent sessions on an Apache or Tomcat Server. We have a couple of those at work and they serve our purpose just fine, but for your line of work, I am not sure. -- Claus Dragon <clauskick@mpsahotmail.com> =(UDIC)= d++ e++ T-- K1!2!3!456!7!S a29 "Coffee is a mocker. So, I am going to mock." - Me, lately. |
| |||
| dogdogdog <young.sul@gmail.com> wrote: > Hi, > I work for a group that primarily uses our servers for statistical > analysis. We mostly use SAS, Stata, Matlab, et al. > > We're trying to plan for upgrades to our existing older servers > (v880s, 440s) and I am wondering if anyone out there has any > experience with the T2000 series as replacements for systems that pull > this kind of duty. > > If so, how does the T2000 compare in performance? > We have a T1000 which is used as a database server and a web server. For these purposes, it works well. For your work, involving floating point calculations, the T1000/T2000 is a poor choice, and your V880 will do better. I have run MATLAB on our T1000, but it was just awful. It goes much better on my 2x750 MHz 280R. It is possible that the new cool threads machines would do better, as they have more than just the single floating point processor that the T1000 suffers from. Try and Buy is your friend. Pick a couple of likely machines and give them a go. Setting up SAS, Stata, Matlab, et al is not too hard, so it shouldn't take much effort to do some proper testing. -- Dr Tristram J. Scott Energy Consultant |
| |||
| > Is floating point performance important? If it is, wait for the new SPARC > chips with a FP unit per core rather than one shared between all the > cores. > > -- > Robert A Heinlein: Theology is never any help; it is searching in a dark > cellar at midnight for a black cat that isn't there. And how about T5220 as number cruncher? Heinz |
| |||
| Oh! Good point. The T5220 has "eight floating point units," according to Sun's technical specs. Have any of you out there used this model for number crunching? On Mar 17, 2:20 pm, "Heinz Müller" <onkelhe...@mscologne.de> wrote: > > Is floating point performance important? If it is, wait for the new SPARC > > chips with a FP unit per core rather than one shared between all the > > cores. > > > -- > > Robert A Heinlein: Theology is never any help; it is searching in a dark > > cellar at midnight for a black cat that isn't there. > > And how about T5220 as number cruncher? > > Heinz |
| |||||
| Hi, I worked on V440 and Solaris T2000 on telecom as a Performance Test Engineer, I saw great improvement on this machine (T2000). Thanks, Yacoob I Quote:
|