Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Calculate Delay from .Lib file

32 views
Skip to first unread message

Ali Arabi

unread,
Aug 28, 2006, 7:46:18 PM8/28/06
to
Dear Friends
How can we calculate Delay of each cell, in Artisan .LIB file?
we haven¡¦t any formula to calculate delay from lookup table timing
of
cell.

t typical=t intrinsic + (k load *C load)

In lib file for example in 4 different states we have like this lookup
table , and in PDF file that explain cell we have this table

AND2(XL)
Intrinsic Delay Kload(ns/pF)
A->Y up 0.0841 6.274
A->Y down 0.1006 3.4200
B->Y up 0.0923 6.2745
B->Y down 0.1126 3.4246

timing() {
related_pin : "A";
timing_sense : positive_unate;
cell_rise(delay_template_7x7) {
index_1 ("0.03, 0.1, 0.4, 0.9, 1.5, 2.2, 3");
index_2 ("0.000175, 0.0105, 0.01925, 0.042, 0.0735, 0.1155,
0.15575");
values ( \
"0.080607, 0.150106, 0.205288, 0.347674, 0.544107, 0.805701,
1.056285", \
"0.092210, 0.161200, 0.216270, 0.358643, 0.555134, 0.816778,
1.067389", \
"0.112858, 0.182922, 0.237694, 0.379680, 0.576040, 0.837712,
1.088363", \
"0.111214, 0.184757, 0.239252, 0.380505, 0.576823, 0.838233,
1.088781", \
"0.090919, 0.168979, 0.224332, 0.366194, 0.561995, 0.823386,
1.073944", \
"0.056146, 0.138642, 0.195301, 0.338845, 0.535911, 0.797161,
1.047540", \
"0.008458, 0.095244, 0.153515, 0.298830, 0.498188, 0.760983,
1.011433");
}
rise_transition(delay_template_7x7) {
index_1 ("0.03, 0.1, 0.4, 0.9, 1.5, 2.2, 3");
index_2 ("0.000175, 0.0105, 0.01925, 0.042, 0.0735, 0.1155,
0.15575");
values ( \
"0.042321, 0.151954, 0.250093, 0.507235, 0.863538, 1.338583,
1.793845", \
"0.042829, 0.152011, 0.250100, 0.507260, 0.863535, 1.338584,
1.793845", \
"0.048219, 0.154139, 0.251460, 0.507370, 0.863538, 1.338584,
1.793840", \
"0.057045, 0.159433, 0.253617, 0.508984, 0.864706, 1.338869,
1.793844", \
"0.065497, 0.168310, 0.260115, 0.512225, 0.865917, 1.340145,
1.794842", \
"0.074833, 0.178598, 0.268261, 0.520047, 0.870877, 1.341815,
1.795897", \
"0.084218, 0.189854, 0.277514, 0.529231, 0.881000, 1.348090,
1.798778");
}
cell_fall(delay_template_7x7) {
index_1 ("0.03, 0.1, 0.4, 0.9, 1.5, 2.2, 3");
index_2 ("0.000175, 0.0105, 0.01925, 0.042, 0.0735, 0.1155,
0.15575");
values ( \
"0.087344, 0.134763, 0.166237, 0.241997, 0.344630, 0.481154,
0.611946", \
"0.105535, 0.152875, 0.184365, 0.260149, 0.362789, 0.499316,
0.630109", \
"0.178071, 0.228516, 0.260849, 0.336985, 0.439534, 0.575983,
0.706742", \
"0.268932, 0.324313, 0.358480, 0.436243, 0.539164, 0.675576,
0.806192", \
"0.363057, 0.423664, 0.460117, 0.540220, 0.643491, 0.779892,
0.910664", \
"0.462961, 0.528881, 0.567910, 0.651651, 0.756655, 0.893396,
1.024021", \
"0.569940, 0.640736, 0.682483, 0.770481, 0.878383, 1.016849,
1.147930");
}
fall_transition(delay_template_7x7) {
index_1 ("0.03, 0.1, 0.4, 0.9, 1.5, 2.2, 3");
index_2 ("0.000175, 0.0105, 0.01925, 0.042, 0.0735, 0.1155,
0.15575");
values ( \
"0.033109, 0.088035, 0.131503, 0.250086, 0.420233, 0.649117,
0.868716", \
"0.033668, 0.088061, 0.131613, 0.250088, 0.420232, 0.649120,
0.868715", \
"0.041242, 0.093877, 0.135425, 0.251087, 0.420372, 0.649123,
0.868719", \
"0.051760, 0.104890, 0.144704, 0.256058, 0.422661, 0.649901,
0.868827", \
"0.063366, 0.117463, 0.156189, 0.262657, 0.425283, 0.651777,
0.870387", \
"0.074639, 0.131292, 0.170108, 0.273798, 0.431797, 0.654301,
0.871924", \
"0.086110, 0.145777, 0.185386, 0.288006, 0.443365, 0.661865,
0.876196");
}
}

I will be very appreciating if you can help me in this information.

Sincerely yours,
Ali Arabi

mk

unread,
Aug 30, 2006, 12:17:39 AM8/30/06
to
On 28 Aug 2006 16:46:18 -0700, "Ali Arabi" <ali....@gmail.com>
wrote:

>Dear Friends
>How can we calculate Delay of each cell, in Artisan .LIB file?
>we haven・t any formula to calculate delay from lookup table timing
>of cell.

You need to look at the definition of delay_template_7x7. It's
probably a two dimensional table with input indeces of input net
transition & total output capacitance. The indeces are rowxcolumn like
a regular matrix and they define breakpoints in the lookup table. You
find your two indeces between which your values lie (ie two indeces in
input net transition between which your input transition value is etc)
then I think you do linear interpolation to find your own value
although this is not very clear from the liberty documentation.

hnu...@gmail.com

unread,
Mar 28, 2014, 2:16:44 PM3/28/14
to
在 2006年8月28日星期一UTC-4下午7时46分18秒,Ali Arabi写道:
Are you resolved this problem?
0 new messages