Blitting speed

Is blitting a 64x64 pixel area onto the screen quicker than blitting
four 32x32 pixel areas onto the screen?–
I will not yell “she’s dead” at roll call

6AD6 865A BF6E 76BB 1FC2 | www.piku.org.uk/public-key.asc
E4C4 DEEA 7D08 D511 E149 | www.piku.org.uk wnzrf at cvxh.bet.hx (rot13’d)

Is blitting a 64x64 pixel area onto the screen quicker than blitting
four 32x32 pixel areas onto the screen?

Yes.

See ya,
-Sam Lantinga, Software Engineer, Blizzard Entertainment

Is blitting a 64x64 pixel area onto the screen quicker than blitting
four 32x32 pixel areas onto the screen?

Yes.
second question, on which os :slight_smile:

Is blitting a 64x64 pixel area onto the screen quicker than blitting
four 32x32 pixel areas onto the screen?

Yes.
and why is it ?

From what I know, it’s only going to be marginaly faster, and on any OS.

If someone more knowledgeable would correct me I’d be happy, but I think
it’s the time it takes to call/setup a function. The core components of the
two take the same time, but register data switching (don’t know a technical
term for it) and such would take some [usually small and constant] time per
call.

-Jim> ----- Original Message -----

From: gj_22@wp.pl (Grzegorz Jaskiewicz)
To:
Sent: Saturday, May 25, 2002 6:43 PM
Subject: Re: [SDL] blitting speed

Is blitting a 64x64 pixel area onto the screen quicker than blitting
four 32x32 pixel areas onto the screen?

Yes.
second question, on which os :slight_smile:


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl

| > > Is blitting a 64x64 pixel area onto the screen quicker than blitting
| > > four 32x32 pixel areas onto the screen?
| >
| > Yes.
| and why is it ?

function call overheads probably. I had a feeling it might be quicker,
but I wanted to check before I rewrote a load of code.On Sun, May 26, 2002 at 12:42:48AM +0200, Grzegorz Jaskiewicz wrote:


The Good Humor Man can only be pushed so far

6AD6 865A BF6E 76BB 1FC2 | www.piku.org.uk/public-key.asc
E4C4 DEEA 7D08 D511 E149 | www.piku.org.uk wnzrf at cvxh.bet.hx (rot13’d)

From what I know, it’s only going to be marginaly faster, and on any OS.

If someone more knowledgeable would correct me I’d be happy, but I think
it’s the time it takes to call/setup a function. The core components of
the
two take the same time, but register data switching (don’t know a
technical
term for it) and such would take some [usually small and constant] time
per
call.
Still i don’t understeand, processor is doing that 4 times instead of one.
Logicaly means, slower not faster !

ok, now i see. it’s late (in austria :slight_smile: ) and … forget that :>> On Sun, May 26, 2002 at 12:42:48AM +0200, Grzegorz Jaskiewicz wrote:

| > > Is blitting a 64x64 pixel area onto the screen quicker than blitting
| > > four 32x32 pixel areas onto the screen?
| >
| > Yes.
| and why is it ?

function call overheads probably. I had a feeling it might be quicker,
but I wanted to check before I rewrote a load of code.

function call overheads probably. I had a feeling it might be quicker,
but I wanted to check before I rewrote a load of code.

ok, now i see. it’s late (in austria :slight_smile: ) and … forget that :>

Not just function call overhead, but basic blit setup, etc.

On X11, four calls to SDL_UpdateRect() means four separate conversations
with the X server (You can reduce this by passing four rectangles to
SDL_UpdateRects()–plural–but that’s still going through a slower
algorithm to accomplish the same thing).

We’re talking about milliseconds here, but they can add up.

–ryan.