aprender-cbtop 0.69.1

Compute Block Top - Real-time load testing and hardware monitoring TUI
Documentation 
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
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213

use super::*;

#[test]
fn test_create_cache() {
    let cache = PagedKvCache::new(1024, 16, 32, 128);
    assert_eq!(cache.total_blocks(), 1024);
    assert_eq!(cache.block_size(), 16);
    assert_eq!(cache.free_block_count(), 1024);
    assert_eq!(cache.used_block_count(), 0);
    assert!((cache.utilization() - 0.0).abs() < f64::EPSILON);
}

#[test]
fn test_allocate_sequence() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    let seq_id = SeqId(1);
    cache.allocate(seq_id, 48).unwrap(); // 3 blocks needed (48/16)

    assert_eq!(cache.used_block_count(), 3);
    assert_eq!(cache.num_sequences(), 1);

    let seq = cache.get_sequence(seq_id).unwrap();
    assert_eq!(seq.num_tokens, 48);
    assert_eq!(seq.num_blocks(), 3);
}

#[test]
fn test_append_tokens() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    let seq_id = SeqId(1);
    cache.allocate(seq_id, 8).unwrap(); // 1 block (8 tokens, block_size=16)
    assert_eq!(cache.used_block_count(), 1);

    cache.append(seq_id, 4).unwrap(); // Still fits in 1 block (12 tokens)
    assert_eq!(cache.used_block_count(), 1);

    cache.append(seq_id, 8).unwrap(); // Needs 2nd block (20 tokens > 16)
    assert_eq!(cache.used_block_count(), 2);
}

#[test]
fn test_free_sequence() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    let seq_id = SeqId(1);
    cache.allocate(seq_id, 48).unwrap();
    assert_eq!(cache.used_block_count(), 3);

    cache.free(seq_id).unwrap();
    assert_eq!(cache.used_block_count(), 0);
    assert_eq!(cache.num_sequences(), 0);
}

#[test]
fn test_fork_sequence() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    let src = SeqId(1);
    let dst = SeqId(2);

    cache.allocate(src, 48).unwrap();
    cache.fork(src, dst).unwrap();

    // Both sequences share blocks
    assert_eq!(cache.num_sequences(), 2);
    // No new blocks allocated (COW)
    assert_eq!(cache.used_block_count(), 3);
    assert_eq!(cache.stats().total_forks, 1);
}

#[test]
fn test_out_of_memory() {
    let mut cache = PagedKvCache::new(10, 16, 32, 128);

    let result = cache.allocate(SeqId(1), 200); // Needs 13 blocks, only 10 available
    assert!(matches!(result, Err(PagedKvError::OutOfMemory { .. })));
}

#[test]
fn test_sequence_not_found() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    let result = cache.free(SeqId(999));
    assert!(matches!(result, Err(PagedKvError::SequenceNotFound(_))));
}

#[test]
fn test_utilization() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    cache.allocate(SeqId(1), 160).unwrap(); // 10 blocks
    let util = cache.utilization();
    assert!((util - 0.1).abs() < 0.001); // 10/100 = 10%

    cache.allocate(SeqId(2), 480).unwrap(); // 30 blocks
    let util = cache.utilization();
    assert!((util - 0.4).abs() < 0.001); // 40/100 = 40%
}

#[test]
fn test_lru_eviction() {
    let mut cache =
        PagedKvCache::new(100, 16, 32, 128).with_eviction_strategy(EvictionStrategy::LRU);

    cache.allocate(SeqId(1), 16).unwrap();
    std::thread::sleep(std::time::Duration::from_millis(10));
    cache.allocate(SeqId(2), 16).unwrap();
    std::thread::sleep(std::time::Duration::from_millis(10));
    cache.allocate(SeqId(3), 16).unwrap();

    // Seq 1 should be evicted (oldest)
    let target = cache.select_eviction_target();
    assert_eq!(target, Some(SeqId(1)));
}

#[test]
fn test_longest_first_eviction() {
    let mut cache =
        PagedKvCache::new(100, 16, 32, 128).with_eviction_strategy(EvictionStrategy::LongestFirst);

    cache.allocate(SeqId(1), 16).unwrap();
    cache.allocate(SeqId(2), 64).unwrap(); // Longest
    cache.allocate(SeqId(3), 32).unwrap();

    // Seq 2 should be evicted (longest)
    let target = cache.select_eviction_target();
    assert_eq!(target, Some(SeqId(2)));
}

#[test]
fn test_evict_to_threshold() {
    let mut cache = PagedKvCache::new(10, 16, 32, 128);

    // Fill to 80%
    for i in 0..8 {
        cache.allocate(SeqId(i), 16).unwrap();
    }
    assert!((cache.utilization() - 0.8).abs() < 0.01);

    // Evict to 50%
    let evicted = cache.evict_to_threshold(0.5).unwrap();
    assert!(evicted.len() >= 3); // Need to evict at least 3 to go from 8 to 5
    assert!(cache.utilization() <= 0.5);
}

#[test]
fn test_memory_calculation() {
    // block_size=16, num_heads=32, head_dim=128
    // block_memory = 2 * 16 * 32 * 128 * 2 = 262144 bytes = 256KB
    let cache = PagedKvCache::new(100, 16, 32, 128);
    assert_eq!(cache.block_memory_bytes(), 262144);
    assert_eq!(cache.total_memory_bytes(), 26214400); // 25MB
}

#[test]
fn test_stats_tracking() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    cache.allocate(SeqId(1), 32).unwrap();
    cache.allocate(SeqId(2), 16).unwrap();
    cache.fork(SeqId(1), SeqId(3)).unwrap();
    cache.free(SeqId(2)).unwrap();

    let stats = cache.stats();
    assert_eq!(stats.total_allocations, 3);
    assert_eq!(stats.total_frees, 1);
    assert_eq!(stats.total_forks, 1);
}

#[test]
fn test_eviction_strategy_display() {
    assert_eq!(format!("{}", EvictionStrategy::LRU), "LRU");
    assert_eq!(format!("{}", EvictionStrategy::LFU), "LFU");
    assert_eq!(
        format!(
            "{}",
            EvictionStrategy::StreamingLLM {
                sink_tokens: 4,
                window_tokens: 512
            }
        ),
        "StreamingLLM(sink=4, window=512)"
    );
}

#[test]
fn test_duplicate_sequence() {
    let mut cache = PagedKvCache::new(100, 16, 32, 128);

    cache.allocate(SeqId(1), 16).unwrap();
    let result = cache.allocate(SeqId(1), 16);
    assert!(matches!(result, Err(PagedKvError::InvalidOperation(_))));
}

#[test]
fn test_blocks_needed_calculation() {
    let cache = PagedKvCache::new(100, 16, 32, 128);

    // blocks_needed is private, test through allocate behavior
    let mut cache2 = cache;
    cache2.allocate(SeqId(1), 1).unwrap();
    assert_eq!(cache2.used_block_count(), 1);

    let mut cache3 = PagedKvCache::new(100, 16, 32, 128);
    cache3.allocate(SeqId(1), 16).unwrap();
    assert_eq!(cache3.used_block_count(), 1);

    let mut cache4 = PagedKvCache::new(100, 16, 32, 128);
    cache4.allocate(SeqId(1), 17).unwrap();
    assert_eq!(cache4.used_block_count(), 2);
}