Timing Analysis

[numbat]

I am getting a corrupted PCF file.  

ERROR:ConstraintSystem:300 - In file: p25_top.pcf(98210): Unexpected end of file '.\p25_top.pcf' during read.
ERROR:Par:51 - The .pcf file contains errors.  PAR cannot proceed.

Process "Place & Route" failed

From line 98210:

TIMEGRP P^Z<8f>^P = PIN
        "p25_i/clock_generator_0/clock_generator_0/MMCM0_INST/MMCM_ADV_inst_pins<2>";
TIMEGRP ðlØ^L = PIN
        "p25_i/clock_generator_0/clock_generator_0/MMCM0_INST/MMCM_ADV_inst_pins<2>";
TIMEGRP àòØ^L = PIN
        "p25_i/clock_generator_0/clock_generator_0/MMCM0_INST/MMCM_ADV_inst_pins<2>";

Many other TIMEGRP's in the file have garbage names.  The April update to 13.1 is installed.

Is there a fix available?

Thanks,

Stacey

[paullee3]

This is a known issue, CR has been filed, it is scheduled to be fixed in future release, meanwhile, please manually revise the pcf content to get around the issue.

[scollinson]

I am having the same problem as above but am unsure what exactly to change the offending line to. Below is a snippet of the offending section of the pcf:

PIN
        U_ila_pro_0/U0/I_YES_D.U_ILA/U_CAPSTOR/I_CASE1.I_YES_TB.U_TRACE_BUFFER/U_RAM/I_V5.U_CS_BRAM_CASCADE_V5/I_DEPTH_LTEQ_32K.U_SBRAM_0/I_B36KGT0.G_RAMB36[3].u_ramb36/U_RAMB36_pins<63>
        = BEL
        "U_ila_pro_0/U0/I_YES_D.U_ILA/U_CAPSTOR/I_CASE1.I_YES_TB.U_TRACE_BUFFER/U_RAM/I_V5.U_CS_BRAM_CASCADE_V5/I_DEPTH_LTEQ_32K.U_SBRAM_0/I_B36KGT0.G_RAMB36[3].u_ramb36/U_RAMB36"
        PINNAME CLKAU;
TIMEGRP Xw«K = BEL
        "U_ila_pro_0/U0/I_YES_D.U_ILA/U_G2_SQ.U_CAPCTRL/I_SRLT_NE_1.U_SCE/I_YESLUT6.U_SRLC16E"
        BEL
        "U_ila_pro_0/U0/I_YES_D.U_ILA/U_G2_SQ.U_CAPCTRL/I_SRLT_NE_1.U_WCE/I_YESLUT6.U_SRLC16E"

 What should this TIMEGRP line be changed to?

[rbod]

Same problem here.

How long will it be before the fix is available? Is there a better workaround because its impossible to guess what the corrupt line should be. At least give a hint on how to reconstruct the lines. Can I remove the offending lines?

[paullee3]

At this moment, Xilinx target this fix in 13.3.
Two possible ways to work around this problem:
1. Turn off the global_opt option before running MAP if this switch is ON
2. Edit the .pcf file after MAP and before PAR and correct the constraint syntax at the line specified in error message, using BEL info at the right side of the equal sign may point out what the original group name should be

[secondox]

could you indicate what and how much the loss would be  if global_opt is turned off?

[rcingham]

Loss of what?

[numbat]

I wasted huge amounts of time with 13.1 crashes and file corruption with global optimization enabled (not to mention wasted time actually watching global optimization run when it isn't crashing) and haven't noticed any area/speed/timing issues in my project with it disabled.

I've since upgraded to EDK 13.2 and haven't attempted to turn it back on since everything is running fairly smoothly without it.

[xsasser]

13.3 is now out. Has the pcf corruption problem been fixed? It not, when will it be?

[graces]

The issue hasn't been fixed in 13.3 yet. It's still under investigation.

In previous test cases which demonstrate this problem, it only occurs in Windows system. Same design in Linux writes out normal PCF.