SVT: an online silicon vertex tracker for the CDF upgrade Page: 5 of 6
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
DDRESS #xy AD
_j ME L I==
-- - -- - - -- NO - -
- - - - MATCH,
- - - - - - kxxx ,
Fig. 2. Associative Memory working principle (left) and a AM8 prototype module (right)
inside a single road, thus increasing the processing time in the TFs. A simulation shows that
with a 250 pm superstrip size and assuming an average of 3 real tracks per event, the average
number of roads per event is three and a total of about 350K patterns is enough to cover all
tracks with Pt > 2 GeV/c. To fullfill these requirements, each of the AM banks must have a
32K pattern capacity.
2.2 Linearized Track Fitting
Each TF processor receives a road in the form of a road address and a list of 'hit' coordinates.
It must then process all the possible combinations of six coordinates (one for each of the 4
SVXII layers used plus a 0 and Pt from the XFT track candidate). A combination correspond-
ing to a real track must satisfy three geometrical constraints, which define a 3D hypersurface
in the 6D 'hit' coordinate space. It has been shown  that it is possible to approximate the
constraint hypersurface to an hyperplane, thus reducing the problem of track fitting and track
parameters extraction to the performance of a small number of scalar products. The small
number of parameters needed to define the constraint plane are calculated at the beginning of
a run. Simulations show that a single set of parameters can be used without a significant loss
of resolution for all the roads in a single wedge.
To perform the actual calculations two options have been considered. The first uses a commer-
cial DSP to perform the scalar products, and custom logic for road distribution and intercon-
nection with the rest of the system. The second, which is the current default, uses a specialized
hardware based on lookup tables which are directly addressed by the hit coordinates .
3 Hardware implementation
The architecture of the system is data-driven. Various tasks are performed by different func-
tional blocks, which exchange data via high-speed point-to-point links. Each block starts work-
ing upon receiving data and outputs results as they are ready, without any syncronization or
handshaking other than the data flow.
Here’s what’s next.
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
Bardi, A.; Belforte, S. & Berryhill, J. SVT: an online silicon vertex tracker for the CDF upgrade, article, July 1, 1997; Batavia, Illinois. (digital.library.unt.edu/ark:/67531/metadc690535/m1/5/: accessed February 19, 2019), University of North Texas Libraries, Digital Library, digital.library.unt.edu; crediting UNT Libraries Government Documents Department.