Electronic Components Datasheet Search
  English  ▼

X  

ORLI10G Datasheet(PDF) 5 Page - Agere Systems

Part # ORLI10G
Description  Quad 2.5 Gbits/s 10 Gbits/s, and 12.5 Gbits/s Line Interface FPSC
PDF  72 Pages
Scroll/Zoom Zoom In 100%  Zoom Out
Manufacturer  AGERE [Agere Systems]
Direct Link  
Logo AGERE - Agere Systems

ORLI10G Datasheet(HTML) 5 Page - Agere Systems

  ORLI10G Datasheet HTML 1Page - Agere Systems ORLI10G Datasheet HTML 2Page - Agere Systems ORLI10G Datasheet HTML 3Page - Agere Systems ORLI10G Datasheet HTML 4Page - Agere Systems ORLI10G Datasheet HTML 5Page - Agere Systems ORLI10G Datasheet HTML 6Page - Agere Systems ORLI10G Datasheet HTML 7Page - Agere Systems ORLI10G Datasheet HTML 8Page - Agere Systems ORLI10G Datasheet HTML 9Page - Agere Systems Next Button
Zoom Inzoom in Zoom Outzoom out
 5 / 72 page
background image
Agere Systems Inc.
5
Data Sheet
October 2001
10 Gbits/s, and 12.5 Gbits/s Line Interface FPSC
ORCA ORLI10G Quad 2.5 Gbits/s
Programmable Features (continued)
s
Enhanced twin-quad programmable function unit
(PFU):
— Eight 16-bit look-up tables (LUTs) per PFU.
— Nine user registers per PFU, one following each
LUT, and organized to allow two nibbles to act
independently, plus one extra for arithmetic opera-
tions.
— New register control in each PFU has two inde-
pendent programmable clocks, clock enables,
local set/reset, and data selects.
— New LUT structure allows flexible combinations of
LUT4, LUT5, new LUT6, 4
→ 1 MUX, new
8
→ 1 MUX, and ripple mode arithmetic functions
in the same PFU.
— 32 x 4 RAM per PFU, configurable as single- or
dual-port. Create large, fast RAM/ROM blocks
(128 x 8 in only eight PFUs) using the SLIC
decoders as bank drivers.
— Soft-wired LUTs (SWL) allow fast cascading of up
to three levels of LUT logic in a single PFU
through fast internal routing which reduces routing
congestion and improves speed.
— Flexible fast access to PFU inputs from routing.
— Fast-carry logic and routing to all four adjacent
PFUs for nibble-wide, byte-wide, or longer arith-
metic functions, with the option to register the PFU
carry-out.
s
Abundant high-speed buffered and nonbuffered
routing resources provide 2x average speed
improvements over previous architectures.
s
Hierarchical routing optimized for both local and
global routing with dedicated routing resources. This
results in faster routing times with predictable and
efficient performance.
s
SLIC provides eight 3-stable buffers, up to a 10-bit
decoder, and PAL™-like and-or-invert (AOI) in each
programmable logic cell.
s
New 200 MHz embedded quad-port RAM blocks,
two read ports, two write ports, and two sets of byte
lane enables. Each embedded RAM block can be
configured as:
— 1—512 x 18 (quad-port, two read/two write) with
optional built-in arbitration.
— 1—256 x 36 (dual-port, one read/one write).
— 1—1k x 9 (dual-port, one read/one write).
— 2—512 x 9 (dual-port, one read/one write for
each).
— 2 RAMs with arbitrary number of words whose
sum is 512 or less by 18 (dual-port, one read/one
write).
— Supports joining of RAM blocks.
— Two 16 x 8-bit content addressable memory
(CAM) support.
— FIFO 512 x 18, 256 x 36, 1k x 9, or dual 512 x 9.
— Constant multiply (8 x 16 or 16 x 8).
— Dual variable multiply (8 x 8).
s
Embedded 32-bit internal system bus plus 4-bit
parity interconnects FPGA logic, microprocessor
interface (MPI), embedded RAM blocks, and
embedded standard cell blocks with 100 MHz bus
performance. Included are built-in system registers
that act as the control and status center for the
device.
s
Built-in testability:
— Full boundary scan (IEEE 1149.1 and draft 1149.2
JTAG) for the programmable I/Os only.
— Programming and readback through boundary-
scan port compliant to IEEE Draft 1532:D1.7.
— TS_ALL testability function to 3-state all I/O pins.
— New temperature-sensing diode.
s
Improved built-in clock management with
programmable phase-locked loops (PPLLs) provides
optimum clock modification and conditioning for
phase, frequency, and duty cycle from 20 MHz up to
420 MHz. Multiplication of input frequency up to 64x
and division of input frequency down to 1/64x
possible.
s
New cycle stealing capability allows a typical 15% to
40% internal speed improvement after final place
and route. This feature also enables compliance with
many setup/hold and clock to out I/O specifications
and may provide reduced ground bounce for output
buses by allowing flexible delays of switching output
buffers.



Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72


Datasheet Download

Go To PDF Page


Link URL



Does ALLDATASHEET help your business so far?  [ DONATE ] 

About Alldatasheet   |   Advertisement   |   Contact us   |   Privacy Policy   |   Link to Datasheet    |   Link Exchange   |   Manufacturer List
All Rights Reserved©Alldatasheet.com


Mirror Sites
English : Alldatasheet.com  |   English : Alldatasheet.net  |   Chinese : Alldatasheetcn.com  |   German : Alldatasheetde.com  |   Japanese : Alldatasheet.jp
Russian : Alldatasheetru.com  |   Korean : Alldatasheet.co.kr  |   Spanish : Alldatasheet.es  |   French : Alldatasheet.fr  |   Italian : Alldatasheetit.com
Portuguese : Alldatasheetpt.com  |   Polish : Alldatasheet.pl  |   Vietnamese : Alldatasheet.vn
Indian : Alldatasheet.in  |   Mexican : Alldatasheet.com.mx  |   British : Alldatasheet.co.uk  |   New Zealand : Alldatasheet.co.nz
Family Site : ic2ic.com  |   icmetro.com