Post-simulation questions

[ilter]

post simulation survey questionnaire

Dear all,
I have some questions. I run post-simulation using hspice.
But I have some errors. Like below.

**error** the memory request exceeds the current available space.
# memory which has been allocated = 700231 kbytes
# memory in next memory request which fails = 25418 kbytes
Please continue by
Step 1: Inspect input file for possible errors in .TRAN
statement or any sources with repeating functions.
Step 2: If other users or jobs are running, wait until
they free up more memory space.
Step 3: Have your system manager enlarge the swap space
on this cpu

***** job aborted

I check my file. I use .tran 2n 5n 1n
I don't know what's wrong.
If you know, please tell me detail.

My circuit is dac, it is large.
If I want to run post-simulation, which tool is suitable?
Please tell me.
Thanks very much.

 

[winsonpku]

script calibre pex capacitance reduction

As i understand,you maybe have three choice:
!)first,replace a more good machine with the now machine
2)reduce the tran time and increse the incremental,but i think this maybe not easy,because the time must be long for some reason
3)use another simulation tool,i am not familiar with this,Hsim or nanosim maybe choince.

As a word, this problem only arouse for the bad machine or bad simulation tool,your hspce simulation file is not bad.

Dear all,
I have some questions. I run post-simulation using hspice.
But I have some errors. Like below.

**error** the memory request exceeds the current available space.
# memory which has been allocated = 700231 kbytes
# memory in next memory request which fails = 25418 kbytes
Please continue by
Step 1: Inspect input file for possible errors in .TRAN
statement or any sources with repeating functions.
Step 2: If other users or jobs are running, wait until
they free up more memory space.
Step 3: Have your system manager enlarge the swap space
on this cpu

***** job aborted

I check my file. I use .tran 2n 5n 1n
I don't know what's wrong.
If you know, please tell me detail.

My circuit is dac, it is large.
If I want to run post-simulation, which tool is suitable?
Please tell me.
Thanks very much.

 

[sunking]

reduce the file size.
this is to say, the small cap(such as 0.1f, it depond on the working frequency) and small res. omit.

 

[flushrat]

I have ever encountered such problem. Post-layout netlist is too big for hspice to simulate. Try to reduce its size or change to other simulation tools. I suggest you use hspice for small of critical circuit post-simulation and hsim for whole chip simulation.

 

[ilter]

Dear all,
Where can I download NanoSim and Hsim information or data?
I don't know how to use these tool. Thanks.

 

[jerryzhao]

I think reduce the file size ,that is a batter way.

I agree with sunking.

 

[Lantis]

Use perl to remove small parasitic cap (< 1e-15).
It will help reduce file size.

 

[ilter]

Use perl to remove small parasitic cap (< 1e-15).
It will help reduce file size.

Dear Lantis,
I think this is good way.
But I see my file.Like below.

.subckt PM_DAC3%XI0/XI0/XI1/XI8/net023 2 3
c0 4 0 0.142859f
c1 2 0 0.00187351f
r2 3 4 2.0925
r3 2 4 2.0925
.ends

.subckt PM_DAC3%XI0/XI1/XI0/XI5/net023 2 3
c0 4 0 0.146016f
c1 2 0 0.0173553f
r2 3 4 2.0925
r3 2 4 2.0925
.ends

.subckt PM_DAC3%XI0/XI0/XI0/XI5/net023 2 3
c0 4 0 0.133676f
r1 3 4 2.0925
r2 2 4 2.0925
.ends

There are many capacitors <1f .
I use ultra-editor.
How do I remote small parasitic cap quickly?
Thanks.

 

[Lantis]

Hi ilter,
"Perl" can handel it very easily.
You can try it by yourself to write a small Perl program.

Good luck.

 

[jcpu]

Dear Sir:

Often LPE tools has instructions for preventing
capacitance and resistance below certain threshold
to get generated.

 

[ilter]

Hi ilter,
"Perl" can handel it very easily.
You can try it by yourself to write a small Perl program.

Good luck.

Dear all,
I try use matlab. My code is like below.

[name,node1,node2,data]=textread('cap.dat','%s%s%s%s');
[r_n, c_n]=size(data);
for n=r_n:-1:1
data{n}=strrep(data{n},'f','e-15');
data{n}=str2num(data{n});
if( data{n} <= 10^-15 )
name=[];
node1=[];
node2=[];
data=[];
end
end
[r, c]=size(data);
for n=r:-1:1
data{n}=strrep(data{n},'e-15','f');
data{n}=num2str(data{n});
end
%for i=1:r_n;
strcat(name,node1,node2,data);
Ans=[name node1 node2 data];


But I doing my post-simulation.
My file like below.
.subckt PM_DAC3%XI0/XI1/XI1/XI8/net023 2 3 4
c12 4 0 0.0120904f
r13 3 4 2.0925
r14 2 4 2.0925
.ends

.subckt PM_DAC3%XI0/XI0/XI1/XI8/net023 2 3 4
c11 4 0 0.0123106f
r12 3 4 2.0925
r13 2 4 2.0925
.ends

.subckt PM_DAC3%XI0/XI1/XI0/XI5/net023 2 3 4
c13 4 0 0.0120904f
r14 3 4 2.0925
r15 2 4 2.0925
.ends

.subckt PM_DAC3%XI0/XI0/XI0/XI5/net023 2 3 4
c7 4 0 0.0150778f
r8 3 4 2.0925
r9 2 4 2.0925
.ends

.subckt PM_DAC3%noxref_X3728/X1/X12/6 2 3 4
c11 4 0 3.15555e-20
r12 3 4 0.952364
r13 2 4 0.952364
.ends

.subckt PM_DAC3%noxref_X3728/X1299/X12/6 2 3 4
c13 4 0 7.9591e-21
r14 3 4 0.952364
r15 2 4 0.952364
.ends

.subckt PM_DAC3%noxref_X3728/X0/X8/6 2 3 4
c9 4 0 3.15555e-20
r10 3 4 0.952364
r11 2 4 0.952364
.ends

.subckt PM_DAC3%noxref_X3728/X8867/X8/6 2 3 4
c10 4 0 4.01134e-20
r11 3 4 0.952364
r12 2 4 0.952364
.ends

.subckt PM_DAC3%XI0/XI1/XI0/net29 4 9 16 19 21 39 40 42 43 67 68 70 71 83 95 96
+ 97 98 101 105 106 121 123 125 126 129 131 132 135 138 140 143 144 146 148 151
+ 154 159 163
c184 163 0 0.0472231f
c185 159 0 0.153621f
c186 154 0 0.00387965f
c187 151 0 0.225288f
c188 148 0 0.170554f
c189 146 0 0.00634714f
c190 144 0 0.0709585f
c191 143 0 0.00700669f
c192 140 0 0.0218542f
c193 138 0 0.00509041f
c194 135 0 0.0104909f
c195 131 0 0.0262096f
c196 129 0 0.0439879f
c197 125 0 0.182381f
c198 123 0 0.0694815f
c199 121 0 0.0457971f
c200 106 0 0.030374f
c201 105 0 0.105931f
c202 101 0 0.0294233f
c203 98 0 0.147245f
c204 97 0 0.0739036f
c205 96 0 0.0545689f
c206 95 0 0.0850038f
c207 83 0 0.0380526f
c208 71 0 0.053631f
c209 70 0 0.0641982f
c210 68 0 0.0592931f
c211 67 0 0.0730853f
c212 43 0 0.0651903f
c213 42 0 0.0722673f
c214 40 0 0.0412032f
c215 39 0 0.0900002f
c216 21 0 0.0149729f
c217 19 0 0.00919115f
c218 16 0 0.0530341f
r219 159 163 6.4
r220 154 156 6.4
r221 154 159 1.05505
r222 148 163 0.156
r223 146 148 6.4
r224 146 151 1.0725
r225 144 163 0.237805
r226 144 145 0.00433426
r227 142 148 0.544098
r228 142 143 0.00366914
r229 138 140 0.0664773
r230 134 145 0.0384835
r231 134 135 0.0874059
r232 131 145 0.0384835
r233 131 132 0.121359
r234 128 156 0.0217904
r235 128 129 0.220235
r236 125 156 0.0217904
r237 125 126 0.642353
r238 124 143 0.0394103
r239 123 143 0.0394103
r240 122 140 0.0195
r241 122 123 0.222182
r242 121 124 0.189091
r243 108 109 7.8
r244 106 109 0.0409894
r245 105 106 0.0435808
r246 103 105 0.449342
r247 103 104 7.8
r248 151 101 6.4
r249 101 102 7.8
r250 98 104 3.63879
r251 96 102 3.63879
r252 96 97 91.2889
r253 95 108 3.63879
r254 95 97 57.0556
r255 98 83 42.1333
r256 76 126 7.8
r257 129 74 7.8
r258 70 74 3.63879
r259 70 71 28.9667
r260 67 76 3.63879
r261 67 68 34.2333
r262 48 135 7.8
r263 46 132 7.8
r264 42 48 3.63879
r265 42 43 33.3556
r266 39 46 3.63879
r267 39 40 18.4333
r268 121 21 0.0195
r269 19 21 0.039
r270 15 19 0.0195
r271 15 16 0.0413636
r272 7 9 2.86803
r273 7 138 11
r274 2 4 2.69913
r275 2 16 5
.ends

There is one question.

.subckt PM_DAC3%XI0/XI0/XI1/XI8/net023 2 3 4

.subckt PM_DAC3%XI0/XI1/XI0/net29 4 9 16 19 21 39 40 42 43 67 68 70 71 83 95 96
+ 97 98 101 105 106 121 123 125 126 129 131 132 135 138 140 143 144 146 148 151
+ 154 159 163

This is a abnormal thing.
So my code can only deal small subckt and can't deal with big circuit.
Thanks very much.

 

[lovseed]

Dear Sir:

Often LPE tools has instructions for preventing
capacitance and resistance below certain threshold
to get generated.

I think this is a better and quick method to get a smaller netlist.

If you use script to delete small resistors and caps, you may have some nodes dangling or unconnected. You scripts may destroy the connection of nodes,devices,etc.

The RC extraction tool,however, can control how much parasitics to be extracted without losing important resistor or caps even though they are small.

Just go through the runset for RC extraction to raise the threshold value for resistor and caps.

 

[ilter]

Dear Sir:

Often LPE tools has instructions for preventing
capacitance and resistance below certain threshold
to get generated.

Dear lovseed,
I use calibre_pex but I don't see any choices for killing below threshold capacitor. If you know , please tell me where. Thanks very much.



I think this is a better and quick method to get a smaller netlist.

If you use script to delete small resistors and caps, you may have some nodes dangling or unconnected. You scripts may destroy the connection of nodes,devices,etc.

The RC extraction tool,however, can control how much parasitics to be extracted without losing important resistor or caps even though they are small.

Just go through the runset for RC extraction to raise the threshold value for resistor and caps.

 

[lovseed]

you may need to look into the manual. This option is usually defined in the runset file.

 

[yen]

Dear Sir:

Often LPE tools has instructions for preventing
capacitance and resistance below certain threshold
to get generated.

Dear all,
I use calibre-pex but I don't see any choices for killing that I don't need capacitors. Thanks.

 

[lovseed]

maybe no direct threshold for resistor or cap values, but there are reduction algorithms can be turned on and you can specify threshold to be reduced.

Also no coupling caps between nodes (but just shunt caps to ground), no paracitics on power nets will reduce much RC.