Physics Stack Exchange is a question and answer site for active researchers, academics and students of physics. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I ran into this while looking at measures of humongous amounts of data. How does the information (data) collected in an inverse femtobarn exposure compare to a gigabyte of data ?

share|cite|improve this question
up vote 7 down vote accepted

I think what you're getting at is not some kind of mathematically rigorous equivalence, but more what it means for a particle physics experiment like ATLAS to collect 1 inverse femtobarn of data. And actually this is computable quire easily.

The design frequency of the LHC is 40Mhz (which corresponds to 25ns bunch spacing, but now i is at 50ns). But since most events are uninteresting background all modern experiments have a system called a "Trigger" which only records events which pass some rough requirements which would render them interesting (maybe a high-momentum electron or jet).

ATLAS is routinely recording at 300Hz (a $10^5$ reduction of rate from the initial collision rate). That is 300 events per second. The size of an event in terms of storage space varies from experiment to experiment and depends on the software it uses, but for ATLAS it is something of the order of 1.5MB/event.

Currently the LHC runs at peak luminosities of 12600 $\mu b^{-1}/s$ (microbarn per second), this decreases over time since the beam intensities decrease, let's just run with 1000$\mu b^{-1}/s$. An inverse femtobarn is $10^9\mu b^{-1}$

so we have:

$$\frac{300 \text{ Events}}{s}\frac{1.5 \text{ MB}}{\text{ Event}}\frac{s}{1000\mu\text{b}^{-1}} \approx 0.5 \frac{\text{MB}}{\mu\text{b}^{-1}}$$

so for $10^9 \mu b^{-1}$ we have


so 500 TB of data

PS: this is just an back-of-the-envelope calculation of course. The rates are constantly changing and the luminosities as well. So collecting 1/fb of data in a low lumi setting requires much more data (since one would still max out the bandwidth of 300Hz recording) than in high lumi settings (where one is still bound by the 300Hz boundary, so the trigger would have to do a tighter selection)

share|cite|improve this answer
This is a perfectly good approach for answering the question as it relates to any particular detector (possibly dependent on trigger configurations), but be aware that the answer can vary by orders of magnitude between detectors. – dmckee Jul 7 '11 at 21:27
It not only varies between detectors, but also between different representations of the data. You can leave out all "raw" information like measured voltages, and just save reconstructed particle tracks etc.. Or you skim your "dataset" and only include events which fired a particular trigger etc., but you still have the same int. luminosity. So while it is a practical measure of data (information collected about nature), it is not really a measure of bytes. – jdm Jul 8 '11 at 8:03
I just checked, a skim I am using (this is the DØ experiment) containing only events with at least one electron (passing a certain preselection), in a not-to-raw format, weighs about 133 GB/fb-1. This contains all information I need for my particular analysis, and it is a size I can reasonably work with (~1.4 TB for the whole dataset). Others will use different parts of the dataset, of course. – jdm Jul 8 '11 at 8:14

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.