[joaquimandrade]

About set_pev/entity_set the results say that it is basically the same thing.

(The same for pev/entity_get)

[ot_207]

Originally Posted by joaquimandrade About set_pev/entity_set the results say that it is basically the same thing.

Ok then Emp` is right!

Originally Posted by Emp` I think there will need to be more testing on the actual fakemeta natives vs the engine natives. As for now the only argument has been fakemeta conversions (stocks that emulate the engine natives), but http://forums.alliedmods.net/showthread.php?t=65427 just puts it to fakemeta, there is not any fakemeta conversions so you can't completely say that fakemeta is slower in that situation.

[joaquimandrade]

Originally Posted by ot_207 Ok then Emp` is right!

He's right about what? Using fakemeta is never faster, even not being engine always clearly faster.

Basically:

All this years everybody (Except Hawk) thought that fakemeta way was faster. Fakemeta way is never faster and can be much slower.

[ot_207]

Originally Posted by joaquimandrade His right about what? Using fakemeta is never faster, even not being engine always clearly faster.

He said about natives functions like pev,set_pev,create_tr2,get_pdata_int are the same. He told us that stocks that emulate functions of engine use a lot of cpu. For example fakemeta util stocks use too much cpu usage if some of those stocks can be found in other modules than it is important to use those modules instead of the stocks.
Example:
Insead of fm_set_rendering it is best to use set_rendering from fun.

Anyway the test shows a slight difference.

[joaquimandrade]

Originally Posted by ot_207 He said about natives functions like pev,set_pev,create_tr2,get_pdata_int are the same. He told us that stocks that emulate functions of engine use a lot of cpu. For example fakemeta util stocks use too much cpu usage if some of those stocks can be found in other modules than it is important to use those modules instead of the stocks. Example: Insead of fm_set_rendering it is best to use set_rendering from fun. Anyway the test shows a slight difference.

No. He just said that more tests should be made. Here they are and fakemeta is never faster.

[ot_207]

Originally Posted by joaquimandrade No. He just said that more tests should be made. Here they are and fakemeta is never faster.

Oh, didn't read right.