Joris van Rantwijk
2343a70c7d
* Evaluate sine against non-phase-adjusted reference which is a little
...
more pessimistic but also a little more fair.
2016-04-18 22:21:35 +02:00
Joris van Rantwijk
5e02c373de
* Implement serial port interface in test design.
2016-04-18 21:14:31 +02:00
Joris van Rantwijk
5b72e7f3f3
* Start work on synthesizable tester for sin/cos core.
2016-04-16 22:56:42 +02:00
Joris van Rantwijk
cf4e802edf
* Add comment in simulation Makefile.
2016-04-16 22:18:41 +02:00
Joris van Rantwijk
5892955445
* Reorganize comments at top of wrappers for sin/cos generator.
2016-04-16 22:16:20 +02:00
Joris van Rantwijk
8796491f37
* Further reduction of multiplier width.
...
At this point it very slightly affects output quality.
This change makes it possible to implement up to 24-bit sine generators
using just 18x18-bit multipliers (i.e. Spartan-6).
2016-04-16 09:10:27 +02:00
Joris van Rantwijk
7ce5b82ff0
* eval_sine_quality.py: Add Python program to evaluate sine waveform.
2016-04-16 09:09:28 +02:00
Joris van Rantwijk
aa8f0217d3
* Full (all-input) test bench for sincos_gen_d18_p20.
2016-04-14 23:56:19 +02:00
Joris van Rantwijk
d2a948f34e
* Fix mistake in 2nd order Taylor correction.
...
* Fix mistake in testbench for 24-bit sine generator.
2016-04-14 23:14:58 +02:00
Joris van Rantwijk
15518ce6f7
* Add test-bench for superficial check of 24-bit sine generator.
...
* Minor improvement in test-bench for 18-bit sine generator.
2016-04-14 23:06:21 +02:00
Joris van Rantwijk
42046dc818
* Add wrapper for 24-bit sine generator.
2016-04-14 22:53:40 +02:00
Joris van Rantwijk
9873f91e8e
* Further reduction of multiplier width.
...
* Add more comments.
2016-04-14 22:49:32 +02:00
Joris van Rantwijk
a3d8939440
* Add comments.
2016-04-14 00:42:58 +02:00
Joris van Rantwijk
2dbc6d44db
* Reduce width of sin/cos input to multiplier by 1 bit.
...
* This has no significant effect on accuracy, but reduces multiplier width.
2016-04-14 00:05:52 +02:00
Joris van Rantwijk
9f4bb7f9b0
* Increase precision of delta_phase term by 1 bit.
...
* This improves accuracy of 1st order Taylor variant to less than 1.0 lsb peak deviation.
* Add comments.
2016-04-13 23:32:02 +02:00
Joris van Rantwijk
12b896c2df
Fix mistake in Taylor correction.
...
This improves accuracy, but there is probably room for further improvement by avoiding accumulation of rounding errors.
2016-04-11 23:21:25 +02:00
Joris van Rantwijk
7de76c6696
Reduce amplitude to avoid numeric overflow.
2016-04-11 21:28:17 +02:00
Joris van Rantwijk
0685db9d92
Document latency of sincos core.
2016-04-10 01:27:24 +02:00
Joris van Rantwijk
e8fdb3cff1
Full (all-input) test bench for sincos_gen_d18_p20.
2016-04-10 01:26:05 +02:00
Joris van Rantwijk
7703971a4d
Add Makefile for simulation with GHDL.
2016-03-24 23:55:32 +01:00
Joris van Rantwijk
e078fffd3e
Test bench for sincos_gen_d18_p20.
2016-03-24 23:37:00 +01:00
Joris van Rantwijk
8090886ab0
VHDL wrapper for sin/cos function with 18-bit sin/cos, 20-bit phase.
2016-03-24 23:34:20 +01:00
Joris van Rantwijk
59ec0505ab
Main VHDL file of sin/cos function core.
2016-03-24 23:32:59 +01:00
Joris van Rantwijk