blob: ed6ccf516200cfd23e733f8d4e308211fb60029c (
plain)
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
|
########################################################################
## The GRAS TODO list
########################################################################
* generate block task loops with fixed lengths
** need a lookup/jump table
** prove that compiler does unrolling
* handle forecast
** not 100% sure here with the input/output edge case sizes
** forecast + input buffer accumulate
* smarter forecast
** tell user number of inputs, user says output reqs
* add hooks to specify input reserve
** automatically calculate from output multiple and rel rate
* handle calculating noutputitems using ninputs as a constraint
* python wrapper for block will come from grextras
* track memcpys for block usage stats
* runtime history, input/output multiple change?
** resend the hint to the allocator and possibly re-allocate
** should only do this on topology commit?
* hooks for setting the "active framework"
** set this before creating blocks to give them an affinity
* put this todo into wiki format checklist
* more thread safety
** all element internals should be accessed via message
and not directly from block::<method>
* create an input and output config struct (per port)
** idea is, we dont need so many API calls for a port
** set/get input/output config
* want per-port token_pool for inputs and outputs
** if ports get removed, this will release subscribers
* token check for top block stop
** fixes qa UDP so we dont wait for inert message
|
