Hello, I did a short blast comparison using multiple cores to see how they behave. I published it in my blog and thought it would be a good idea to share it here.
http://voorloopnul.com/blog/how-to-correctly-speed-up-blast-using-num_threads/
Hello, I did a short blast comparison using multiple cores to see how they behave. I published it in my blog and thought it would be a good idea to share it here.
http://voorloopnul.com/blog/how-to-correctly-speed-up-blast-using-num_threads/
Benchmarking on VPS servers is tricky since everything is virtualized: CPU, harddrive etc.
Also the size of the database and size of a typical query may also be a factor. Each of these could turn a blast lookup from being CPU bound to IO bound and could substantially alter the results of the test.
Hello Istvan,
Most of your concerns are stated in the "Disclaimer" section, since the article is not intended to be a "One size fits all".
But Indeed, benchmarking on a VPS is really tricky, I had to switch one of the VMs because it was giving me absurd results, I believe the physical server hosting this particular VM was with high load.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
You should modify y-axis (start from 0), for me it wasn't clear.
Edit:
Some time ago I asked similar question on http://unix.stackexchange.com/ -- got really informative answers.
Why using more threads makes it slower than using less threads