/
TODO
76 lines (58 loc) · 3.33 KB
/
TODO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
*********************************************************
* Try to make the configurations as automatic as possible
*********************************************************
1. Use two metrics for comparison (DONE)
: Light-checking mode: Pin discrepancy mode
- quickly check if there is any interface-level discrepancy
- Simply checks the sign of each coefficient in linear regression with simple linear model (no interaction, no high-order)
: Deep-checking mode
- Allow interaction and poly-nomials
- check if the max residual error is less than the tolerance of the output
- Adaptively changes the number of test vectors
: Try to fit the golden model by looking at the response residuals and see if the rate of change of response residuals as you exercise more test vectors
- Provide pin brief pin discrepancy with simple linear regression (no polynomial, no interaction terms)
- Get rid of wire section and let the tool automatically declare wires by looking up some cross-reference file (DONE)
**************
* Small tasks
**************
- Create a new, integrated package distribution script (DONE)
- Create a simple licensing scheme for evaluation purpose (DONE)
- Support Synopsys CustomSim for AMS simulation
- Create more model/test templates. In particular, for RF systems
- Explicitly declare Analog Input and Analog Control Input in port specification, such that the tool automatically generate a test to extract a transfer function from each analog input to analog output.
- Add runcommand file
: e.g. bus_delimiter = <>
- Adjust offset term of revised model to that of golden model for comparison if there is not any predictor variable. (DISCARDED)
- Filter out insignificant predictors which embrace 0.0 in their confidence interval (DONE)
**************
* After V1.0
**************
V1.01
-----
- Get rid of wire section and let the tool automatically declare wires by looking up some cross-reference file (DONE)
- Get rid of 'dut' field in a test configuration (DONE)
- Change 'time' to 'timeend' for simulation time (DONE)
- Get rid of 'timeprec' in [simulation] section and use 'timeunit' for 'timeprec' as well. (DONE)
- Change keywords for regression section and fields in a test configuration (DONE)
V1.02
-----
- Resurrect gain tolerance in a report (DONE)
- change the keyword, 'timeend', in simulation section to 'trantime' (DONE)
- Add copyright notifications to samples and examples (DONE)
V1.05
----
- Get rid of "min_sample" and "no_of_grid" in mProbo. Let users specify "max_sample" instead. mProbo will terminate the checking before fully committing "max_sample" sampling. (DONE)
- Simplify regression options (DONE)
- Make equivalence criterion simple (e.g. remove gain tolerance spec.) (DONE)
- Revise vector generation (DONE)
- Update manual
- Revise report generation
********************
1. dump.v: when te is too large, "iter" can be negative integer (DONE)
2. Minimum number of test vector calculation. When interaction terms are considered, the # of vectors should be larger than the terms. In particular, it needs many vectors when there are quantized analog ports.
3. In cached mode, run simulations if simulation directory does not exist. (DONE)
4. txline.v has a bug
5. interim checking result (DONE)
6. add "port_reference file" argument for mProbo (DONE)
********************
1. Allow to configure ignorant digital mode