edg-l/ddnet
1
0
Fork 0
mirror of https://github.com/ddnet/ddnet.git synced 2024-11-10 10:08:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge #5715

Browse source 
5715: Use djb2 for snapshot item hashlist r=def- a=Robyt3

The previous hash function was heavily biased towards the hash buckets 64-79, making those buckets overflow faster, which results in snapshot CRC errors and lags.

Using the djb2 hash yields an almost even distribution over the entire range of values.

Source for djb2: http://www.cse.yorku.ca/~oz/hash.html
(we incidentally use the same implementation for `str_quickhash`)

Sample which compares the usage of hash buckets in the test map from #5454 with the old and new hash function: [SnapshotHash.csv](https://github.com/ddnet/ddnet/files/9285148/SnapshotHash.csv) (this table also differentiates between the source and target of the snapshot delta)

Closes #4379.

## Checklist

- [X] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test (especially base/) or added coverage to integration test
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)


Co-authored-by: Robert Müller <robytemueller@gmail.com>
This commit is contained in:
bors[bot] 2022-08-08 22:38:22 +00:00 committed by GitHub
parent 3273781d3f 243ab7c2f0
commit dbe6d8cb29
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 20 additions and 10 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

30
src/engine/shared/snapshot.cpp
View file

@ -144,18 +144,28 @@ bool CSnapshot::IsValid(size_t ActualSize) const
// CSnapshotDelta
struct CItemList
{
int m_Num;
int m_aKeys[64];
int m_aIndex[64];
};
enum
{
HASHLIST_SIZE = 256,
HASHLIST_BUCKET_SIZE = 64,
};
struct CItemList
{
int m_Num;
int m_aKeys[HASHLIST_BUCKET_SIZE];
int m_aIndex[HASHLIST_BUCKET_SIZE];
};
inline size_t CalcHashID(int Key)
{
// djb2 (http://www.cse.yorku.ca/~oz/hash.html)
unsigned Hash = 5381;
for(unsigned Shift = 0; Shift < sizeof(int); Shift++)
Hash = ((Hash << 5) + Hash) + ((Key >> (Shift * 8)) & 0xFF);
return Hash % HASHLIST_SIZE;
}
static void GenerateHash(CItemList *pHashlist, CSnapshot *pSnapshot)
{
for(int i = 0; i < HASHLIST_SIZE; i++)
@ -164,8 +174,8 @@ static void GenerateHash(CItemList *pHashlist, CSnapshot *pSnapshot)
for(int i = 0; i < pSnapshot->NumItems(); i++)
{
int Key = pSnapshot->GetItem(i)->Key();
int HashID = ((Key >> 12) & 0xf0) | (Key & 0xf);
if(pHashlist[HashID].m_Num != 64)
size_t HashID = CalcHashID(Key);
if(pHashlist[HashID].m_Num < HASHLIST_BUCKET_SIZE)
{
pHashlist[HashID].m_aIndex[pHashlist[HashID].m_Num] = i;
pHashlist[HashID].m_aKeys[pHashlist[HashID].m_Num] = Key;
@ -176,7 +186,7 @@ static void GenerateHash(CItemList *pHashlist, CSnapshot *pSnapshot)
static int GetItemIndexHashed(int Key, const CItemList *pHashlist)
{
int HashID = ((Key >> 12) & 0xf0) | (Key & 0xf);
size_t HashID = CalcHashID(Key);
for(int i = 0; i < pHashlist[HashID].m_Num; i++)
{
if(pHashlist[HashID].m_aKeys[i] == Key)