English
Language : 

EP4SE360F35I4 Datasheet, PDF (50/432 Pages) Altera Corporation – This section provides a complete overview of all features relating to the Stratix IV device family, which is the most architecturlly advanced
2–14
Chapter 2: Logic Array Blocks and Adaptive Logic Modules in Stratix IV Devices
Adaptive Logic Modules
Shared Arithmetic Mode
In shared arithmetic mode, the ALM can implement a three-input add within the
ALM. In this mode, the ALM is configured with 4 four-input LUTs. Each LUT either
computes the sum of three inputs or the carry of three inputs. The output of the carry
computation is fed to the next adder (either to adder1 in the same ALM or to adder0 of
the next ALM in the LAB) using a dedicated connection called the shared arithmetic
chain. This shared arithmetic chain can significantly improve the performance of an
adder tree by reducing the number of summation stages required to implement an
adder tree. Figure 2–11 shows the ALM using this feature.
Figure 2–11. ALM in Shared Arithmetic Mode
shared_arith_in
carry_in
datae0
datac
datab
dataa
4-Input
LUT
4-Input
LUT
labclk
DQ
reg0
To general or
local routing
To general or
local routing
datad
datae1
4-Input
LUT
4-Input
LUT
DQ
reg1
To general or
local routing
To general or
local routing
carry_out
shared_arith_out
You can find adder trees in many different applications. For example, the summation
of the partial products in a logic-based multiplier can be implemented in a tree
structure. Another example is a correlator function that can use a large adder tree to
sum filtered data samples in a given time frame to recover or de-spread data that was
transmitted using spread-spectrum technology.
Shared Arithmetic Chain
The shared arithmetic chain available in enhanced arithmetic mode allows the ALM
to implement a three-input add. This significantly reduces the resources necessary to
implement large adder trees or correlator functions.
Shared arithmetic chains can begin in either the first or sixth ALM in the LAB. The
Quartus II Compiler creates shared arithmetic chains longer than 20 (10 ALMs in
arithmetic or shared arithmetic mode) by linking LABs together automatically. For
enhanced fitting, a long shared arithmetic chain runs vertically, allowing fast
horizontal connections to the TriMatrix memory and DSP blocks. A shared arithmetic
chain can continue as far as a full column.
Stratix IV Device Handbook
Volume 1
February 2011 Altera Corporation