A taste of an experimental QuickBMS feature

[aluigi]

just released.
let me know if now it's a bit faster.

[AlphaTwentyThree]

It's a bit faster, yes. But it's still really slow and I don't really know why. Could you take a look at my script? Here: viewtopic.php?f=13&t=4450&p=40773#p40773
I tried to pre-allocate the sizes with an array but all I get is an error message that the array is too big:

Code: Select all

for i = 0 < STREAMS # pre-allocate array
	PutArray 0 i 0
next i
get FSIZE asize
set OFFSET 0
DO
	set IDENT OFFSET
	math IDENT += 0x11
	goto IDENT
	get STREAM byte
	getArray TMP 0 STREAM
	math TMP += 0x930
	putArray 0 STREAM TMP
	math OFFSET += 0x930
WHILE OFFSET < FSIZE

I need this complete run-through because there is no way to tell where the layer parts are located. STREAMS has 4 as maximum value, so it should be just a 1x4 matrix.

[AlphaTwentyThree]

UPDATE: the long running time is definitely my script. Just tested the new QuickBMS with my deinterleave script and it really is a lot faster! Thanks for this update Luigi!

[aluigi]

it's enough that you specify only the latest value so the following:

Code: Select all

for i = 0 < STREAMS # pre-allocate array
   PutArray 0 i 0
next i

can be simply replaced by:

Code: Select all

PutArray 0 STREAMS 0

[AlphaTwentyThree]

Apparently there's something wrong with my script. But thanks for the hint.

[aluigi]

other improvements and fixes of the secure allocation feature.
it's too cool to abandon it even if there are some limitations that rarely can affect quickbms users.
more info here:
viewtopic.php?p=68373#p68373