Optimzing map reading

[klusbert]

private void UpdateTileNums()

     {

         int index = 0;

 

         int Adr = client.Memory.ReadInt32(Adresses.Map.TileNumberArray);

         for (byte z = 0; z < 8; z++)

         {

             for (byte y = 0; y < 14; y++)

             {

                 for (byte x = 0; x < 18; x++)

                 {

                     TileNumberArray[z, y, x] = client.Memory.ReadInt32(Adr + index);

                     index += 4;

 

                 }

             }

         }

 

     }

     public IEnumerable<Tile> GetTilesSameFloor()

     {

         UpdateTileNums();

       

         Location Selfloc = client.PlayerLocation;

         int Index = 7 - Selfloc.Z;

         if (Index >= 0)

         {

             //nothing 

         }

         else

         {

             Index = 2;

         }

 

         for (byte y = 0; y < 14; y++)

         {

             for (byte x = 0; x < 18; x++)

             {

                 int tilenum =TileNumberArray[Index, y, x];

                 int adr = TileNumToMapAdress(tilenum);

                 Location loc = new Location(Selfloc.X + x - 8, Selfloc.Y + y -6, Selfloc.Z);

                 Tile t = new Tile(client, (uint)adr, (uint)tilenum, loc);

                 yield return t;

                

             }

         }

       

     }

This is just great, this code works like a charm now you dont have to convert tilenums to memloc and memloc to worldloc. I should work faster and it should be 100% accurate. Since tibiaAPIs mapreading is not always 100%.

http://i.imgur.com/LM9uTfq.jpg

[Blequi]

private void UpdateTileNums()

     {

         int index = 0;

 

         int Adr = client.Memory.ReadInt32(Adresses.Map.TileNumberArray);

         for (byte z = 0; z < 8; z++)

         {

             for (byte y = 0; y < 14; y++)

             {

                 for (byte x = 0; x < 18; x++)

                 {

                     TileNumberArray[z, y, x] = client.Memory.ReadInt32(Adr + index);

                     index += 4;

 

                 }

             }

         }

 

     }

Nice job, klusbert.
By the way, there's no need for this loop. You can accomplish the whole nested for loops in a single call to ReadProcessMemory and still get a speed boost.

[klusbert]

true, but I think this shows more what is going on. But Do you think I will gain much by caching both tilenum array and tiles array?

[Blequi]

I made this simple test using the Stopwatch class.

for each test, a single readprocessmemory call was ~ 2.75 MS faster than these multiple RPM calls. (just this for loop, not couning the tiles caching)

tested in a for from 0 to 1000 and it gave me 2.9 sec difference. Since it isn't a very huge difference, it's totally up to you. (test made in a single tibia client)

[Blood]

1) The code in the first post is to search player tile, but why? Since in this new structure the player tile is always the 116th.
2) What are the constants "Addresses.Map.MaxTilesPerFloor" and "Addresses.Map.TileNumberOffsetCenter" ?
3) Why in tibiaAPI "maxZ" is 8 since there are more floors than 8 ?
Thx

[pater]

For client 10.21 pointer for structure of "tile indexes" is at 0x5EA640 and it seems to be always located before map pointer.

[Blequi]

Did not you meant 0x5EA660 + base? Usually it's located at TilesPointer + 0x1C

[pater]

Yes, you are right Blequi, it is located at TilesPointer + 0x1C. TilesPointer - 0x4 is max tiles which is always 2016. My mistake, sorry.

[normik]

Sorry for bumping, but this method is nice(Y). Now is there any way to get tile objects straight from this method?

[FrankJScott]

Please try Google before asking about High Rated Product Guide 89dedbb