Closed vijayank88 closed 10 months ago
Do you have any macros? We have seen such cases when there is an unreasonably small max_transition
Yes, sram integrated.
How did you generate the .lib for the sram?
It is coming with part of gf180 pdk.
Careful with keywords in PR descriptions: this is not yet fixed in OpenLane.
@donn @vijayank88 : What are the next steps needed for this to be fixed in OpenLane ? Do we need to update OpenLane to use the latest OR code?
The update is the OL team's responsibility (@donn )
I got similar problem in my design. The "[STEP 15] Running Global Routing Resizer Design Optimizations" stage takes forever, while all previous steps take a bunch of minutes. Here's the tail of the log:
[INFO GRT-0096] Final congestion report:
Layer Resource Demand Usage (%) Max H / Max V / Total Overflow
---------------------------------------------------------------------------------------
Metal1 0 0 0.00% 0 / 0 / 0
Metal2 1144386 129375 11.31% 0 / 0 / 0
Metal3 1144522 151323 13.22% 0 / 0 / 0
Metal4 1119052 41853 3.74% 0 / 0 / 0
---------------------------------------------------------------------------------------
Total 3407960 322551 9.46% 0 / 0 / 0
[INFO GRT-0018] Total wirelength: 3697520 um
[INFO GRT-0014] Routed nets: 39950
[INFO]: Setting RC values...
[INFO RSZ-0058] Using max wire length 9189um.
I just updated OpenLane and everything. I am trying to harden my design to GF180MCUD PDK. The configuration file is:
{
"DESIGN_NAME": "CORTEZ1",
"DESIGN_IS_CORE": 0,
"VERILOG_FILES": [
"dir::../../verilog/rtl/all_modules.v",
"dir::../../verilog/rtl/defines.v",
"dir::../../verilog/rtl/CORTEZ1.v"
],
"CLOCK_PERIOD": 20,
"CLOCK_PORT": "wb_clk_i",
"CLOCK_NET": "NETWORK_TOP_WRAPPER_i.CLK",
"FP_SIZING": "absolute",
"DIE_AREA": "0 0 3000 3000",
"FP_PIN_ORDER_CFG": "dir::pin_order.cfg",
"MAX_TRANSITION_CONSTRAINT": 10.0,
"MAX_FANOUT_CONSTRAINT": 16,
"PL_RESIZER_TIMING_OPTIMIZATIONS": 1,
"PL_RESIZER_SETUP_SLACK_MARGIN": 0.5,
"PL_RESIZER_HOLD_SLACK_MARGIN": 0.5,
"PL_RESIZER_ALLOW_SETUP_VIOS": 1,
"GLB_RESIZER_TIMING_OPTIMIZATIONS": 1,
"GLB_RESIZER_SETUP_SLACK_MARGIN": 0.5,
"GLB_RESIZER_HOLD_SLACK_MARGIN": 0.5,
"GLB_RESIZER_ALLOW_SETUP_VIOS": 1,
"MAGIC_DEF_LABELS": 0,
"SYNTH_BUFFERING": 0,
"RUN_HEURISTIC_DIODE_INSERTION": 1,
"HEURISTIC_ANTENNA_THRESHOLD": 110,
"GRT_REPAIR_ANTENNAS": 1,
"VDD_NETS": [
"vccd1"
],
"GND_NETS": [
"vssd1"
],
"IO_SYNC": 0,
"BASE_SDC_FILE": "dir::base_cortez.sdc",
"RUN_CVC": 1,
"PDK": "gf180mcuD",
"STD_CELL_LIBRARY": "gf180mcu_fd_sc_mcu7t5v0",
"FP_CORE_UTIL": 40,
"RT_MAX_LAYER": "Metal4",
"PL_TARGET_DENSITY": 0.45,
"SYNTH_STRATEGY": "DELAY 4"
}
Any idea?
If disabling it is out of the question, you may want file a new issue
Description
I am trying to harden this GFMPW design : https://github.com/gregdavill/gf180-mpw0-serv/tree/main/openlane/serv_0 with latest OpenLane, but resizer got struck and consume all RAM and get killed at the end.
Expected Behavior
Complete resizer the stage.
Environment report
Reproduction material
issue_reproducible.zip
Relevant log output