greg/sparsemap
1
0
Fork 0
Code Issues Pull requests 1 Projects Releases Packages Wiki Activity

WIP

Browse source
This commit is contained in:
Gregory Burd 2024-07-27 13:16:37 -04:00
parent bf3840e183
commit 40951e37e6
2 changed files with 15 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
README.md
View file

@ -12,7 +12,7 @@ The "best" case happens when large consecutive sequences of the bits are
either set ("1") or not set ("0"). If your numbers are consecutive 64 bit
integers then sparsemap can compress up to 16kb in just 8 bytes.
## How does it work?
## How does it work? TODO... refine...
On the lowest level a bitmap contains a number of chunks. Each chunk has a
starting offset (`uint32_t`), a descriptor (the first `sm_bitvec_t`), and may

19
sparsemap.c
View file

@ -1429,7 +1429,7 @@ bidx_clear(sparsemap_t *map, sparsemap_idx_t idx)
/* Set the chunk flags to all ones, ... */
pivot_chunk.m_data[0] = ~(__sm_bitvec_t)0;
/* ... set the flag for the position containing the index to mixed ... */
SM_CHUNK_SET_FLAGS(pivot_chunk.m_data[0], aligned_idx / SM_BITS_PER_VECTOR, SM_PAYLOAD_MIXED);
SM_CHUNK_SET_FLAGS(pivot_chunk.m_data[0], idx / SM_BITS_PER_VECTOR, SM_PAYLOAD_MIXED);
/* ... and clear only the bit at that index in this chunk. */
pivot_chunk.m_data[1] = ~(__sm_bitvec_t)0 & ~((__sm_bitvec_t)1 << (idx % SM_BITS_PER_VECTOR));
__sm_when_diag({
@ -1815,14 +1815,23 @@ bidx_set(sparsemap_t *map, sparsemap_idx_t idx)
if (SM_IS_CHUNK_RLE(&adj)) {
/* Does it align with this full sparse chunk? */
if (start + SM_CHUNK_MAX_CAPACITY == adj_start) {
/* The stars have aligned, combine them! */
// TODO
fprintf(stdout, "whee");
size_t adj_length = __sm_chunk_rle_get_length(&adj);
if (adj_length + SM_CHUNK_MAX_CAPACITY < SM_CHUNK_RLE_MAX_CAPACITY) {
/* The stars have aligned, transform to RLE and combine them! */
// __sm_when_diag({ fprintf(stdout, "\n%s\n", QCC_showChunk(p, 0)); } );
// __sm_when_diag({ fprintf(stdout, "\n%s\n", QCC_showChunk(adj_p, 0)); } );
SM_CHUNK_SET_RLE(&chunk);
__sm_chunk_rle_set_length(&chunk, __sm_chunk_rle_get_length(&adj));
__sm_chunk_rle_set_capacity(&chunk, __sm_chunk_rle_get_capacity(&adj));
// __sm_when_diag({ fprintf(stdout, "\n%s\n", QCC_showChunk(p, 0)); } );
__sm_remove_data(map, adj_offset, SM_SIZEOF_OVERHEAD + sizeof(__sm_bitvec_t));
__sm_set_chunk_count(map, __sm_get_chunk_count(map) - 1);
}
}
} else {
/* Is this adjacent sparse chunk also all ones? */
if (adj.m_data[0] == ~(__sm_bitvec_t)0 && start + SM_CHUNK_MAX_CAPACITY == adj_start) {
/* The stars have aligned, combine them! */
/* The stars have aligned, transform to RLE and combine them! */
// TODO
fprintf(stdout, "whee");
}