ibex
ibex copied to clipboard
[Confirmation] Regression and Coverage data metrics
Hi,
Below are the regression and coverage statistics for the most recent version of IBEX (ec32fb1a6460d8512f9192a84d5fc395502c13f5). Could you kindly confirm which is correct or not?
Regression data: 94.49% PASS 1285 PASSED, 75 FAILED
Coverage data:
Code Coverage:
Functional Coverage:
Note: I've read the vrefines files(core_ibex/waivers/*)
-- BR, Shailesh
As I can't see the column headings for those screenshots I cannot comment as to the accuracy of those numbers.
Our latest nightly regression (sadly results not currently publicly available though we are planning to make them public) reports
93.26% PASS 1287 PASSED, 93 FAILED
With coverage summary report:
Legend: Metric* means cumulative e.g. Block* means Cumulative Block Coverage
name Block* Covered Branch* Covered Statement* Covered Expression* Covered Toggle* Covered Statement* Covered Fsm* Covered Assertion* Covered CoverGroup* Covered
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
uvm_pkg n/a n/a n/a n/a n/a n/a n/a n/a 100.00% (3/3)
ibex_top 94.65% (12009/12688/1850) 90.55% (5498/6072/562) 94.62% (12407/13112/2294) 91.71% (4843/5281/4882) 98.55% (613/622/2402) 94.62% (12407/13112/2294) 100.00% (138/138) 98.08% (358/365/30) 41.54% (8189/19712)
name CoverGroup Average CoverGroup Covered
-----------------------------------------------------------------------------------
riscv_instr_pkg 92.69% 92.76% (10511/11331)
push_pull_agent_pkg 100.00% 100.00% (3/3)
core_ibex_fcov_if 92.41% 50.00% (1253/2506)
core_ibex_pmp_fcov_if.g_pmp_cgs_T 96.33% 91.86% (79/86)
core_ibex_pmp_fcov_if.g_pmp_cgs_T.g_pmp_region_fcov 93.18% 75.99% (345/454)
Note for the functional coverage the total bins covered is a little misleading as we have some very large crosses in there. We're still working on closing out coverage for these and there are likely large numbers of illegal bins we haven't had time to identify yet.
I have attached the full regression log so you can see the individual failures