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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139

use std::num::{NonZeroU32, NonZeroU64};

use serde::{Deserialize, Serialize};
use serde_state::DeserializeState;

use necsim_core_bond::PositiveF64;
use necsim_partitioning_core::partition::{Partition, PartitionSize};

use necsim_impls_no_std::parallelisation::independent::{DedupCache, EventSlice, RelativeCapacity};

#[derive(Clone, Debug, Serialize, Deserialize)]
#[serde(deny_unknown_fields)]
pub struct MonolithicParallelismMode {
    pub event_slice: EventSlice,
}

#[derive(Clone, Debug, Serialize, Deserialize)]
#[serde(deny_unknown_fields)]
pub struct IsolatedParallelismMode {
    pub event_slice: EventSlice,
    pub partition: Partition,
}

#[derive(Clone, Debug, Serialize, Deserialize)]
pub enum ParallelismMode {
    Monolithic(MonolithicParallelismMode),
    IsolatedIndividuals(IsolatedParallelismMode),
    IsolatedLandscape(IsolatedParallelismMode),
}

impl<'de> DeserializeState<'de, PartitionSize> for ParallelismMode {
    fn deserialize_state<D>(
        partition_size: &mut PartitionSize,
        deserializer: D,
    ) -> Result<Self, D::Error>
    where
        D: serde::de::Deserializer<'de>,
    {
        use serde::de::Error;

        let parallelism_mode = ParallelismMode::deserialize(deserializer)?;

        match parallelism_mode {
            ParallelismMode::Monolithic(..)
            | ParallelismMode::IsolatedIndividuals(..)
            | ParallelismMode::IsolatedLandscape(..)
                if !partition_size.is_monolithic() =>
            {
                Err(D::Error::custom(format!(
                    "parallelism_mode {parallelism_mode:?} is incompatible with non-monolithic \
                     partitioning."
                )))
            },
            partition_mode => Ok(partition_mode),
        }
    }
}

#[derive(Clone, Debug, Serialize)]
#[allow(clippy::module_name_repetitions)]
pub struct CudaArguments {
    pub device: u32,
    pub ptx_jit: bool,
    pub delta_t: PositiveF64,
    pub block_size: NonZeroU32,
    pub grid_size: NonZeroU32,
    pub step_slice: NonZeroU64,
    pub dedup_cache: DedupCache,
    pub parallelism_mode: ParallelismMode,
}

impl<'de> DeserializeState<'de, PartitionSize> for CudaArguments {
    fn deserialize_state<D>(
        partition_size: &mut PartitionSize,
        deserializer: D,
    ) -> Result<Self, D::Error>
    where
        D: serde::de::Deserializer<'de>,
    {
        let raw = CudaArgumentsRaw::deserialize_state(partition_size, deserializer)?;

        let parallelism_mode = if let Some(parallelism_mode) = raw.parallelism_mode {
            parallelism_mode
        } else if !partition_size.is_monolithic() {
            return Err(serde::de::Error::custom(
                "The CUDA algorithm is (currently) incompatible with non-monolithic partitioning.",
            ));
        } else {
            ParallelismMode::Monolithic(MonolithicParallelismMode {
                event_slice: EventSlice::Relative(RelativeCapacity {
                    factor: PositiveF64::new(20.0_f64).unwrap(),
                }),
            })
        };

        Ok(CudaArguments {
            device: raw.device,
            ptx_jit: raw.ptx_jit,
            delta_t: raw.delta_t,
            block_size: raw.block_size,
            grid_size: raw.grid_size,
            step_slice: raw.step_slice,
            dedup_cache: raw.dedup_cache,
            parallelism_mode,
        })
    }
}

#[derive(Debug, DeserializeState)]
#[serde(default, deny_unknown_fields)]
#[serde(deserialize_state = "PartitionSize")]
pub struct CudaArgumentsRaw {
    pub device: u32,
    pub ptx_jit: bool,
    pub delta_t: PositiveF64,
    pub block_size: NonZeroU32,
    pub grid_size: NonZeroU32,
    pub step_slice: NonZeroU64,
    pub dedup_cache: DedupCache,
    #[serde(deserialize_state)]
    pub parallelism_mode: Option<ParallelismMode>,
}

impl Default for CudaArgumentsRaw {
    fn default() -> Self {
        Self {
            device: 0_u32,
            ptx_jit: true,
            delta_t: PositiveF64::new(3.0_f64).unwrap(),
            block_size: NonZeroU32::new(64_u32).unwrap(),
            grid_size: NonZeroU32::new(64_u32).unwrap(),
            step_slice: NonZeroU64::new(150_u64).unwrap(),
            dedup_cache: DedupCache::Relative(RelativeCapacity {
                factor: PositiveF64::new(0.1_f64).unwrap(),
            }),
            parallelism_mode: None,
        }
    }
}