ERROR (SPECTRE-7031): SST2 Error: No space left on device

 Dear All,

I was running a simulation of a PEXed circuit block.

The size of psf folder was exceeding 297GB !!  Why So large size ?

The PSS really took almost 2 days with APS enabled and at 39%, it gave following error:-

Could anybody please tell what is the cause of this error and how to avoid this error and coplete the PSS of big circuits.

\o     pss: time = 235.9 ns    (39.3 %), step = 1.025 ps     (171 u%)
\o
\o
\o
\o Error found by spectre at time = 235.888 ns during periodic steady state
\o         analysis `pss'.
\o     ERROR (SPECTRE-7031): SST2 Error: No space left on device.
\o
\o     ERROR (SPECTRE-7031): SST2 Error: No space left on device.
\o
\o
\o
\o Analysis `pss' was terminated prematurely due to an error.

The circuit simulation info is as below:-

 Circuit inventory:
\o               nodes 449447
\o
\o             bsim3v3 17914
\o
\o           capacitor 1397804
\o               diode 48    
\o
\o            inductor 1     
\o                port 5     
\o              res_va 12    
\o
\o            resistor 528771
\o
\o         transformer 22    
\o
\o                vcvs 9     
\o
\o             vsource 19    
\o
\o
\o
\o Notice from spectre during initial setup.
\o     APS Enabled.
\o
\o     Multithreading Enabled: 8 threads on system with 8 available processors.
\o         
\o
\o     3 notices suppressed.
\o
\o     1993 warnings suppressed.
\o
\o
\o
\o
\o Time for parsing: CPU = 389.862 s (6m  29.9s), elapsed = 392.837 s (6m
\o         32.8s).
\o
\o Time accumulated: CPU = 453.545 s (7m  33.5s), elapsed = 456.924 s (7m
\o         36.9s).
\o
\o Peak resident memory used = 3.19 Gbytes.
\o
\o
\o
\o
\o ******************
\o DC Analysis `dcOp'
\o ******************
\o
\o Finding DC approximate solution failed. Try again with try_fast_op set to no.
\o
\o Important parameter values:
\o     reltol = 1e-03
\o
\o     abstol(V) = 100 uV
\o     abstol(I) = 100 pA
\o
\o     temp = 27 C
\o
\o     tnom = 27 C
\o
\o     tempeffects = all
\o
\o     gmindc = 1 pS
\o     rabsshort = 1 mOhm
\o
\o Trying `homotopy = gmin'.
\o
\o
\o Notice from spectre during DC analysis `dcOp'.
\o     Bad pivoting is found during DC analysis. Option dc_pivot_check=yes is
\o
\o         recommended for possible improvement of convergence.
\o
\o
\o Convergence achieved in 80 iterations.
\o
\o Total time required for dc analysis `dcOp': CPU = 1.08634 ks (18m  6.3s),
\o         elapsed = 687.388 s (11m  27.4s).
\o
\o Time accumulated: CPU = 1.54125 ks (25m  41.2s), elapsed = 1.14572 ks (19m
\o         5.7s).
\o
\o Peak resident memory used = 4.49 Gbytes.
\o
\o
\o
\o dcOpInfo: writing operating point information to rawfile.
\o
\o
\o ***************************************************
\o Periodic Steady-State Analysis `pss': fund = 20 MHz
\o ***************************************************
\o
\o
\o
\o Notice from spectre during IC analysis, during periodic steady state analysis
\o         `pss'.
\o     Bad pivoting is found during DC analysis. Option dc_pivot_check=yes is
\o         recommended for possible improvement of convergence.
\o
\o
\o
\o =============================
\o `pss': time = (0 s -> 600 ns)
\o =============================
\o
\o Important parameter values in tstab integration:
\o
\o     start = 0 s
\o     outputstart = 0 s
\o     stop = 600 ns
\o     period = 50 ns
\o     step = 650 ps
\o
\o     maxstep = 2 ns
\o
\o     ic = all
\o
\o     useprevic = no
\o
\o     skipdc = no
\o
\o     reltol = 1e-03
\o
\o     abstol(V) = 100 uV
\o
\o     abstol(I) = 100 pA
\o     temp = 27 C
\o
\o     tnom = 27 C
\o
\o     tempeffects = all
\o
\o     method = gear2only
\o
\o     lteratio = 3.5
\o
\o     relref = sigglobal
\o
\o     cmin = 0 F
\o
\o     gmin = 1 pS
\o
\o     rabsshort = 1 mOhm