Add Decimation/Interpolation to node class. #141

drslebedev · 2023-07-27T15:34:31Z

A special tag PerformDecimationInterpolation is required for node be enable decimation/interpolation.
Add ports_status to simplify port limits checks.
Add qa_node.cpp for unit tests of node class.
The stride functionality will be implemented in a separate PR: Implement stride for node class. #142

New

┌──────────────────────────────benchmark:───────────────────────────────┬──────┬─#N──┬─CTX-SW─┬───CPU cache misses────┬───CPU branch misses───┬─<CPU-I>─┬──min───┬──mean──┬─stddev─┬─median─┬──max───┬─total time─┬─ops/s─┐
│ merged src->sink work                                                 │ PASS │ 10  │     0  │ 2.09k / 3.86k = 54.2% │   390  / 212k =  0.2% │   10.4  │  14 us │  14 us │ 329 ns │  14 us │  15 us │     143 us │  716M │
│ merged src->copy->sink                                                │ PASS │ 10  │     0  │  643  / 1.38k = 46.7% │  203  / 11.3k =  1.8% │    1.1  │   4 us │   5 us │ 484 ns │   5 us │   5 us │      45 us │  2.3G │
│ merged src->copy->sink work                                           │ PASS │ 10  │     0  │   276  / 726  = 38.0% │   151  / 210k =  0.1% │   10.3  │  10 us │  12 us │ 648 ns │  11 us │  13 us │     116 us │  880M │
│ merged src->copy^10->sink                                             │ PASS │ 10  │     0  │  635  / 1.37k = 46.4% │   219  / 110k =  0.2% │    7.7  │  10 us │  11 us │ 357 ns │  11 us │  11 us │     110 us │  929M │
│ merged src(N=1024)->b1(N≤128)->b2(N=1024)->b3(N=32...128)->sink       │ PASS │ 10  │     0  │  720  / 1.46k = 49.3% │  204  / 10.8k =  1.9% │    1.1  │   4 us │   4 us │ 608 ns │   4 us │   5 us │      44 us │  2.4G │
│ merged src->mult(2.0)->divide(2.0)->add(-1)->sink - float             │ PASS │ 10  │     0  │  652  / 1.48k = 44.0% │   230  / 110k =  0.2% │   10.7  │  12 us │  12 us │ 569 ns │  12 us │  14 us │     124 us │  829M │
│ merged src->mult(2.0)->divide(2.0)->add(-1)->sink - int               │ PASS │ 10  │     0  │  628  / 1.39k = 45.1% │   220  / 110k =  0.2% │   11.7  │  17 us │  17 us │ 406 ns │  17 us │  18 us │     171 us │  598M │
│ merged src->(mult(2.0)->div(2.0)->add(-1))^10->sink - float           │ PASS │ 10  │     0  │  727  / 1.41k = 51.6% │  3.39k / 113k =  3.0% │   38.7  │ 108 us │ 132 us │  30 us │ 109 us │ 177 us │       1 ms │ 77.7M │
│ merged src->(mult(2.0)->div(2.0)->add(-1))^10->sink - int             │ PASS │ 10  │     0  │  691  / 1.41k = 48.9% │   219  / 117k =  0.2% │   48.8  │ 158 us │ 181 us │  39 us │ 159 us │ 257 us │       2 ms │ 56.4M │
│ runtime   src->sink overhead                                          │ PASS │ 10  │     0  │  13.5k / 181k =  7.4% │  2.67k / 441k =  0.6% │   21.9  │  13 us │  51 us │ 110 us │  14 us │ 382 us │     509 us │  201M │
│ runtime   src->copy->sink                                             │ PASS │ 10  │     0  │  25.8k / 353k =  7.3% │  4.09k / 746k =  0.5% │   34.8  │  14 us │  67 us │ 157 us │  15 us │ 537 us │     668 us │  153M │
│ runtime   src->copy^10->sink                                          │ PASS │ 10  │     0  │  114k / 1.91M =  6.0% │ 19.6k / 3.55M =  0.6% │    153  │  30 us │ 403 us │   1 ms │  37 us │   4 ms │       4 ms │ 25.4M │
│ runtime   src(N=1024)->b1(N≤128)->b2(N=1024)->b3(N=32...128)->sink    │ PASS │ 10  │     0  │  46.5k / 679k =  6.8% │ 7.15k / 1.46M =  0.5% │   67.0  │  21 us │ 165 us │ 413 us │  28 us │   1 ms │       2 ms │ 62.0M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - float             │ PASS │ 10  │     0  │  46.6k / 720k =  6.5% │ 8.25k / 1.39M =  0.6% │   62.7  │  20 us │ 156 us │ 395 us │  22 us │   1 ms │       2 ms │ 65.7M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - int               │ PASS │ 10  │     0  │  45.5k / 715k =  6.4% │ 8.23k / 1.40M =  0.6% │   64.1  │  19 us │ 157 us │ 403 us │  23 us │   1 ms │       2 ms │ 65.2M │
│ runtime   src->(mult(2.0)->div(2.0)->add(-1))^10->sink - float        │ PASS │ 10  │     0  │  315k / 5.61M =  5.6% │ 50.6k / 10.2M =  0.5% │    448  │ 121 us │   1 ms │   3 ms │ 131 us │  10 ms │      11 ms │  9.6M │
│ runtime   src->(mult(2.0)->div(2.0)->add(-1))^10->sink - int          │ PASS │ 10  │     0  │  292k / 5.61M =  5.2% │ 50.7k / 10.2M =  0.5% │    457  │ 129 us │   1 ms │   3 ms │ 136 us │   9 ms │      10 ms │  9.8M │
│ runtime   src->mult(2.0)->mult(0.5)->add(-1)->sink (SIMD)             │ PASS │ 10  │     0  │  46.2k / 706k =  6.6% │ 7.80k / 1.39M =  0.6% │   61.7  │  17 us │ 160 us │ 418 us │  21 us │   1 ms │       2 ms │ 64.0M │
│ runtime   src->(mult(2.0)->mult(0.5)->add(-1))^10->sink (SIMD)        │ PASS │ 10  │     0  │  270k / 5.43M =  5.0% │ 49.6k / 9.95M =  0.5% │    419  │  80 us │   1 ms │   3 ms │  87 us │  10 ms │      10 ms │  10.M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - float single      │ PASS │ 10  │    18  │ 1.20M / 6.32M = 18.9% │  215k / 10.7M =  2.0% │    549  │   2 ms │   2 ms │ 835 us │   2 ms │   4 ms │      20 ms │  5.1M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - int single        │ PASS │ 10  │    24  │ 1.16M / 6.25M = 18.5% │  195k / 10.6M =  1.8% │    543  │   2 ms │   2 ms │ 724 us │   2 ms │   4 ms │      19 ms │  5.5M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - float bulk        │ PASS │ 10  │     0  │  40.3k / 705k =  5.7% │ 8.02k / 1.42M =  0.6% │   63.0  │  20 us │ 150 us │ 368 us │  27 us │   1 ms │       1 ms │ 68.5M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - int bulk          │ PASS │ 10  │     0  │  37.9k / 704k =  5.4% │ 7.83k / 1.51M =  0.5% │   69.4  │  32 us │ 154 us │ 353 us │  36 us │   1 ms │       2 ms │ 66.6M │
└───────────────────────────────────────────────────────────────────────┴──────┴─────┴────────┴───────────────────────┴───────────────────────┴─────────┴────────┴────────┴────────┴────────┴────────┴────────────┴───────┘

main

┌──────────────────────────────benchmark:───────────────────────────────┬──────┬─#N──┬─CTX-SW─┬───CPU cache misses────┬───CPU branch misses───┬─<CPU-I>─┬──min───┬──mean──┬─stddev─┬─median─┬──max───┬─total time─┬─ops/s─┐
│ merged src->sink work                                                 │ PASS │ 10  │     0  │ 1.67k / 3.88k = 43.1% │   457  / 213k =  0.2% │   12.4  │  13 us │  13 us │ 334 ns │  14 us │  14 us │     134 us │  762M │
│ merged src->copy->sink                                                │ PASS │ 10  │     0  │  653  / 1.43k = 45.7% │  193  / 11.0k =  1.8% │    1.1  │   5 us │   6 us │ 404 ns │   6 us │   6 us │      55 us │  1.8G │
│ merged src->copy->sink work                                           │ PASS │ 10  │     0  │   259  / 759  = 34.1% │   139  / 209k =  0.1% │   12.2  │  11 us │  11 us │ 300 ns │  11 us │  12 us │     112 us │  913M │
│ merged src->copy^10->sink                                             │ PASS │ 10  │     0  │  674  / 1.46k = 46.1% │   237  / 110k =  0.2% │    7.7  │  10 us │  11 us │ 557 ns │  11 us │  12 us │     110 us │  933M │
│ merged src(N=1024)->b1(N≤128)->b2(N=1024)->b3(N=32...128)->sink       │ PASS │ 10  │     0  │  705  / 1.51k = 46.7% │  206  / 11.1k =  1.9% │    1.1  │   4 us │   5 us │ 530 ns │   5 us │   5 us │      46 us │  2.2G │
│ merged src->mult(2.0)->divide(2.0)->add(-1)->sink - float             │ PASS │ 10  │     0  │  655  / 1.47k = 44.6% │   213  / 110k =  0.2% │   10.7  │  11 us │  12 us │ 850 ns │  12 us │  14 us │     120 us │  852M │
│ merged src->mult(2.0)->divide(2.0)->add(-1)->sink - int               │ PASS │ 10  │     0  │  698  / 1.47k = 47.6% │   202  / 110k =  0.2% │   11.7  │  16 us │  17 us │ 390 ns │  17 us │  18 us │     170 us │  603M │
│ merged src->(mult(2.0)->div(2.0)->add(-1))^10->sink - float           │ PASS │ 10  │     0  │  745  / 1.49k = 50.0% │  3.39k / 114k =  3.0% │   38.8  │ 108 us │ 131 us │  30 us │ 109 us │ 177 us │       1 ms │ 78.3M │
│ merged src->(mult(2.0)->div(2.0)->add(-1))^10->sink - int             │ PASS │ 10  │     0  │  744  / 1.44k = 51.7% │   226  / 117k =  0.2% │   48.8  │ 159 us │ 171 us │  17 us │ 161 us │ 200 us │       2 ms │ 59.8M │
│ runtime   src->sink overhead                                          │ PASS │ 10  │     0  │  15.8k / 181k =  8.7% │  2.65k / 440k =  0.6% │   21.8  │  13 us │  42 us │  86 us │  14 us │ 301 us │     423 us │  242M │
│ runtime   src->copy->sink                                             │ PASS │ 10  │     0  │  28.4k / 353k =  8.0% │  4.29k / 746k =  0.6% │   34.7  │  13 us │  83 us │ 201 us │  16 us │ 687 us │     825 us │  124M │
│ runtime   src->copy^10->sink                                          │ PASS │ 10  │     0  │  118k / 1.93M =  6.1% │ 18.9k / 3.55M =  0.5% │    153  │  33 us │ 431 us │   1 ms │  39 us │   4 ms │       4 ms │ 23.8M │
│ runtime   src(N=1024)->b1(N≤128)->b2(N=1024)->b3(N=32...128)->sink    │ PASS │ 10  │     0  │  45.2k / 673k =  6.7% │ 7.35k / 1.45M =  0.5% │   66.3  │  21 us │ 173 us │ 431 us │  31 us │   1 ms │       2 ms │ 59.2M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - float             │ PASS │ 10  │     0  │  46.5k / 712k =  6.5% │ 7.87k / 1.38M =  0.6% │   62.0  │  19 us │ 158 us │ 402 us │  24 us │   1 ms │       2 ms │ 64.9M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - int               │ PASS │ 10  │     0  │  36.5k / 709k =  5.2% │ 7.84k / 1.38M =  0.6% │   63.0  │  21 us │ 210 us │ 548 us │  29 us │   2 ms │       2 ms │ 48.7M │
│ runtime   src->(mult(2.0)->div(2.0)->add(-1))^10->sink - float        │ PASS │ 10  │     0  │  305k / 5.61M =  5.4% │ 50.5k / 10.0M =  0.5% │    435  │ 107 us │   1 ms │   3 ms │ 117 us │   9 ms │      10 ms │  9.9M │
│ runtime   src->(mult(2.0)->div(2.0)->add(-1))^10->sink - int          │ PASS │ 10  │     0  │  300k / 5.57M =  5.4% │ 50.2k / 10.0M =  0.5% │    446  │ 120 us │   1 ms │   3 ms │ 125 us │   9 ms │      11 ms │  9.7M │
│ runtime   src->mult(2.0)->mult(0.5)->add(-1)->sink (SIMD)             │ PASS │ 10  │     0  │  42.6k / 706k =  6.0% │ 6.95k / 1.36M =  0.5% │   60.1  │  23 us │ 182 us │ 476 us │  24 us │   2 ms │       2 ms │ 56.2M │
│ runtime   src->(mult(2.0)->mult(0.5)->add(-1))^10->sink (SIMD)        │ PASS │ 10  │     0  │  284k / 5.42M =  5.2% │ 48.8k / 9.94M =  0.5% │    419  │  76 us │   1 ms │   3 ms │  80 us │  10 ms │      10 ms │  9.8M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - float single      │ PASS │ 10  │    25  │ 1.13M / 6.17M = 18.3% │  207k / 10.5M =  2.0% │    538  │   2 ms │   2 ms │ 798 us │   2 ms │   4 ms │      20 ms │  5.2M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - int single        │ PASS │ 10  │    26  │ 1.15M / 6.15M = 18.6% │  204k / 10.5M =  1.9% │    540  │   2 ms │   2 ms │ 775 us │   2 ms │   4 ms │      21 ms │  5.0M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - float bulk        │ PASS │ 10  │     0  │  39.5k / 742k =  5.3% │ 8.01k / 1.45M =  0.6% │   64.6  │  23 us │ 164 us │ 408 us │  29 us │   1 ms │       2 ms │ 62.5M │
│ runtime   src->mult(2.0)->div(2.0)->add(-1)->sink - int bulk          │ PASS │ 10  │     1  │  40.6k / 748k =  5.4% │ 8.49k / 1.63M =  0.5% │   77.1  │  52 us │ 198 us │ 428 us │  56 us │   1 ms │       2 ms │ 51.7M │
└───────────────────────────────────────────────────────────────────────┴──────┴─────┴────────┴───────────────────────┴───────────────────────┴─────────┴────────┴────────┴────────┴────────┴────────┴────────────┴───────┘

pr-explainer-bot · 2023-07-27T15:35:50Z

Pull Request Review

Hey there! 👋 I've summarized the results of the previous tasks for you. Let's get started with the review!

Changes

Added a new struct PerformDecimationInterpolation in annotated.hpp.
Added a new operator= function in Annotated struct in annotated.hpp.
Added a new operator std::string_view function in Annotated struct in annotated.hpp.
Added a new unwrap_if_wrapped_t type trait in Annotated struct in annotated.hpp.
Added a new typelist specialization in SupportedTypes struct in annotated.hpp.
Added a new operator<< function in Annotated struct in annotated.hpp.
Added a new samples_to_next_tag function in port.hpp.
Added a new struct ports_status_t to track the status of input and output ports.
Added a member variable ports_status of type ports_status_t to the node class.

Suggestions

In annotated.hpp, line 62-63: Consider adding a brief description for the PerformDecimationInterpolation struct.
In annotated.hpp, line 65-66: Consider adding a brief description for the BlockingIO struct.
In annotated.hpp, line 133-134: Consider adding a brief description for the Annotated struct.
In annotated.hpp, line 189-190: Consider adding a brief description for the unwrap_if_wrapped type trait.
In port.hpp, line 605-606: Consider adding a brief description for the publish_tag function.
In port.hpp, line 608-609: Consider adding a brief description for the samples_to_next_tag function.
In the update_ports_status() function (lines 311-366), you can simplify the logic for updating ps.in_samples by using std::clamp instead of multiple if statements. Here's an example:
```
ps.in_samples = std::clamp(ps.in_available, ps.in_min_samples, ps.in_max_samples);
```

Bugs

In port.hpp, line 605-606: The publish_tag function may have a potential bug. The tag_offset parameter is not used in the function body.
In the update_ports_status() function (lines 311-366), there is a potential bug in the calculation of ps.in_at_least_one_tag_available. The line ps.in_at_least_one_tag_available = ps.in_at_least_one_port_has_data | (port.tagReader().available() > 0); should use the logical OR operator (||) instead of the bitwise OR operator (|). Here's the corrected line:
```
ps.in_at_least_one_tag_available = ps.in_at_least_one_port_has_data || (port.tagReader().available() > 0);
```

Improvements

In annotated.hpp, line 133-134, the description() function can be refactored for better readability:
```
static constexpr std::string_view description() noexcept {
    return description_;
}
```
In the update_ports_status() function (lines 311-366), you can improve readability by using range-based for loops instead of meta::tuple_for_each. Here's an example:
```
for (auto& port : input_ports(&self())) {
    // Your code here
}
```

Rating

Overall code rating: 8/10

Criteria:

Readability: 8/10
Performance: N/A
Security: N/A

That's it for the review! Let me know if you need any further assistance. Good luck with your pull request! 🚀

* A special tag PerformDecimationInterpolation is required for node be enable decimation/interpolation. * Add ports_status to simplify port limits checks. * Add qa_node.cpp for unit tests of node class.

…r, denominator, stride.

sonarcloud · 2023-08-01T12:45:51Z

SonarCloud Quality Gate failed.

0 Bugs
0 Vulnerabilities
0 Security Hotspots
29 Code Smells

62.4% Coverage
0.0% Duplication

Catch issues before they fail your Quality Gate with our IDE extension SonarLint

wirew0rm · 2023-08-01T13:21:41Z

Force merging due to coverage requirements. We'll probably have to add some rules to make it possible to actually reach the coverage goal, there are e.g. some fmt constructs that generate branches that are not easily reachable.

drslebedev linked an issue Jul 27, 2023 that may be closed by this pull request

interpolator/decimator blocks fair-acc/opendigitizer#67

Closed

drslebedev temporarily deployed to configure coverage July 27, 2023 15:34 — with GitHub Actions Inactive

drslebedev had a problem deploying to configure coverage July 27, 2023 15:34 — with GitHub Actions Failure

drslebedev temporarily deployed to configure coverage July 27, 2023 15:34 — with GitHub Actions Inactive

drslebedev had a problem deploying to configure coverage July 27, 2023 15:34 — with GitHub Actions Failure

drslebedev force-pushed the intdec branch from 5db1a26 to f837fba Compare July 27, 2023 19:18

drslebedev temporarily deployed to configure coverage July 27, 2023 19:18 — with GitHub Actions Inactive

drslebedev had a problem deploying to configure coverage July 27, 2023 19:18 — with GitHub Actions Failure

drslebedev temporarily deployed to configure coverage July 27, 2023 19:18 — with GitHub Actions Inactive

drslebedev had a problem deploying to configure coverage July 27, 2023 19:18 — with GitHub Actions Failure

drslebedev had a problem deploying to configure coverage August 1, 2023 11:45 — with GitHub Actions Failure

drslebedev temporarily deployed to configure coverage August 1, 2023 11:45 — with GitHub Actions Inactive

drslebedev added 3 commits August 1, 2023 14:11

Add Decimation/Interpolation to node class.

2d160e3

* A special tag PerformDecimationInterpolation is required for node be enable decimation/interpolation. * Add ports_status to simplify port limits checks. * Add qa_node.cpp for unit tests of node class.

Update test.grc.expected file by adding new node parameters: numerato…

fc85aaf

…r, denominator, stride.

Add ill-defined parameters check.

8c5c140

drslebedev force-pushed the intdec branch from 265c934 to 8c5c140 Compare August 1, 2023 12:11

drslebedev temporarily deployed to configure coverage August 1, 2023 12:12 — with GitHub Actions Inactive

wirew0rm merged commit 691415a into main Aug 1, 2023

wirew0rm deleted the intdec branch August 1, 2023 13:21

wirew0rm mentioned this pull request Aug 14, 2023

[Xpt] graph: refactor Block::work() function #81

Closed

5 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add Decimation/Interpolation to node class. #141

Add Decimation/Interpolation to node class. #141

drslebedev commented Jul 27, 2023 •

edited

Loading

pr-explainer-bot bot commented Jul 27, 2023

sonarcloud bot commented Aug 1, 2023

wirew0rm commented Aug 1, 2023

Add Decimation/Interpolation to node class. #141

Add Decimation/Interpolation to node class. #141

Conversation

drslebedev commented Jul 27, 2023 • edited Loading

New

main

pr-explainer-bot bot commented Jul 27, 2023

Pull Request Review

Changes

Suggestions

Bugs

Improvements

Rating

sonarcloud bot commented Aug 1, 2023

wirew0rm commented Aug 1, 2023

drslebedev commented Jul 27, 2023 •

edited

Loading