| kkrunchy by Farbrausch [web] | ||||||||
|---|---|---|---|---|---|---|---|---|
|
|
|||||||
|
popularity : 74% |
|||||||
alltime top: #484 |
|
|||||||
|
||||||||
added on the 2006-08-24 20:54:35 by ryg  |
||||||||
popularity helper
comments
Respect.
yeah
I'm happy as a degraded consumer.
good move
i feel degraded o/
great packer
t3h shit!
utmost adoration.
Amazing! Just for fun I ran it on my old Goathack 64k intro. The version I released was packed with UPX and was 41.5kb. Using kkrunchy instead got it down to 32.5kb
degrade me beautiful.
thanks ryg
now go finish kkrieger! :P
Rulez!
top of the line!
danke schon ryg :)
Hooray!
that's what i was waiting for
crunch it!
rulez! thanx a lot ryg. :)
Reminds me of the good old times of powerpacker on amiga ;-)
p.s. Size does matter! 8)
What Maali said. :D
restekp
Doesn't run on my lobster statue.
because size matters.
Big thanks for releasing it! Never found any leaked versions anyway...
Nice to see that you finally got this into a releasable state. Here is my thumb for doing a CM-based 64k packer, so we dont have to :)
And fuckings to you for getting more attention and comments about this in less than a day than crinkler got in just over a year :)
And fuckings to you for getting more attention and comments about this in less than a day than crinkler got in just over a year :)
this tool has saved our asses more times over the past few years than anything else. and made more extra content possible than anything else. =)
megathumb for the tool, and also for all the help over the last few years with beta versions, info etc :)
megathumb for the tool, and also for all the help over the last few years with beta versions, info etc :)
Great tool for 64k intros.
Good
uristen \o/
Finally released officially? Yeah!
teh spirit.
the. packer. THE. PACKER.
ryg for prez
omfg
It's OK!
mentor, only 0.23alpha2 does context modelling, all previous versions are lz+arithmetic coding+the secret incredient :)
Great packer.
NICE!
NICE!
fuckin' awesome tool. just too bad intros suck :(
you did it!!! rulez
well done ryg!
Amazing packer! Thanks again Ryg for the versions you used to send me before releasing this one :)
IT'S AMAZING!! kkrunchy made me happy :-)
as magic said, size DOES matter =)
great tool
great tool
Jeez, this makes such a difference I can't believe it!
Tip of the day [by Vova4Age]:
You can pack trojans and viruses with this until antiviruses update their signatures with something like 'trojan.kkrunched.a' or 'bioskill.kkrunchy' :)))
You can pack trojans and viruses with this until antiviruses update their signatures with something like 'trojan.kkrunched.a' or 'bioskill.kkrunchy' :)))
alfsolft, don't start, had to contact 3 antivirus manifacturers this week because kkrieger got tagged as a trojan by their recent updates... resolved now, but it was DAMN annoying :)
ohne worte...
yes, finally!
I tried this on a couple of my (bloated, horribly oversized) programming projects for my astronomy class, and wow is it a good compressor (granted, even compressed my programs are rather large). The only thing I don't quite understand is what the difference between the two versions is.
variant A (i.e. 0.23alpha) uses a pretty basic LZ+arithmetic algorithm (no, not LZMA; worse pack ratio, but smaller depacker). it's slow to compress but decompresses a lot faster (very roughly 2MB/sec on my P4 2.4GHz for typical data).
A2 (0.23alpha2) is a context mixing based algorithm, akin to PAQ and crinkler (in fact I use the neural network mixer from PAQ7 with some simpler custom models). it's usually faster to pack than variant A, but takes just as long to depack (the algorithm is symmetric). i've spent quite some time (and sacrificed size too) to make average depack time reasonable for 64ks, but it's quite useless for anything much larger.
main reason for the version number strangeness is that both started out as experimental branches that i later deemed good enough to release :) - 0.23final (should I get around to make it) will probably support both variants in the same executable. and expose some other nifty features too.
A2 (0.23alpha2) is a context mixing based algorithm, akin to PAQ and crinkler (in fact I use the neural network mixer from PAQ7 with some simpler custom models). it's usually faster to pack than variant A, but takes just as long to depack (the algorithm is symmetric). i've spent quite some time (and sacrificed size too) to make average depack time reasonable for 64ks, but it's quite useless for anything much larger.
main reason for the version number strangeness is that both started out as experimental branches that i later deemed good enough to release :) - 0.23final (should I get around to make it) will probably support both variants in the same executable. and expose some other nifty features too.
The best! Thanks for releasing it. :)
Ryg - I hear you about the virus thing... MuCruncher went through the same thing, and I had to jump through hoops to get them to remove it. Kaspersky decided that anything comrpessed with Mu was a virus.
Bastiches.
Bastiches.
ryg: The interesting thing is, I don't recall ANY size difference between the files compressed with a and a2. I would have expected a slight difference given the different algorithms (but then again I know nothing of compression algorithms). I may have to make a more careful study sometime...
deathy, i guess the .exe dropping played a big part in that; for kkrieger at least, after we noticed it is was mainly a matter of writing a few mails. no bitching no nothing, and it got removed on the next update in all cases :)
crusader: well, that can ofcourse happen, but it's unlikely :)
crusader: well, that can ofcourse happen, but it's unlikely :)
ryg - Actually, it wasn't the droppings... it was that they couldn't figure out how to depack them. I'd think if they noticed the droppings, they'd have been able to decompress them. heheheheh
They wanted more than an e-mail from me, they wanted me to write them a depacker.
They wanted more than an e-mail from me, they wanted me to write them a depacker.
this thumb goes for a better future in a 64k world.
Great stuff, I borrowed couple of ideas from this one to my own PE-packer. (My packer uses aPLib though, so it gets nowhere close to packing ratios of this one).
Kudos.
Kudos.
Nothing more to say; its just fantastic. Thx!
By the way, Va2 makes a difference of 6kb in our last intro when compared to the kkcrunchy I was using before, and 5 kb when compared to kkcrunchy Va [so, > 16 kb when compared to upx, hehe]. Its also much MUCH faster to compress. Amazing!!
guess i should copy over the "sufficiently advanced technology" tagline i already use on the kkapture homepage, eh? :)
I used Kkrunchy for my GLrez intro (and credited Ryg into the scrolltext) and I must admit it's one of the best tool ever made for intro !
Ryg! Your Kkrunchy made my intro 20kb :D
\o/
1337!
Ryg! Your Kkrunchy made my intro 20kb :D
\o/
1337!
nice one ryg!
Hot.
Rulz!
[quote]Instead of releasing the tools, it would have been better to write tutorials for how to make tools of one's own. This would encourage learning and competition. By releasing their tools to the masses, Farbrausch put themselves above the scene and degrade demo sceners to consumers.[quote]
Finally, something better than upx.
MadenMann! Wubbish! I have "pre-visualised" two music videos in this software (I.E. Done them in Werkkzueg - then rendered them in 3DSMax). It's not for beginnerz at all.
However, going back to your statement, why should "they" not get a chance to do something cool? You can't do this on a Playstation can ya?
And I bet none of said "consumers" will win Assembly'07 or '08 or '09, or......
However, going back to your statement, why should "they" not get a chance to do something cool? You can't do this on a Playstation can ya?
And I bet none of said "consumers" will win Assembly'07 or '08 or '09, or......
hello arblickspule, congratulations! you just made a fool out of yourself! have fun clicking the link in madenmann's post and *reading that entire thread*... :-)
just vote
HAVOC!!! Fair play. Hands up, I admit it. I was a bit drunk and had two (acutally four) browser windows up and got a bit confused and one thing lead to another and......
Anyway, I apologise to all concerned, and endorse V2 as a fellow Buzz user myself (when I'm not making a drunken twat out of myself that is!). Have fun, live well, and get hammered before writing post'z - It seems to everyone laugh.
;)
Anyway, I apologise to all concerned, and endorse V2 as a fellow Buzz user myself (when I'm not making a drunken twat out of myself that is!). Have fun, live well, and get hammered before writing post'z - It seems to everyone laugh.
;)
"It seems to everyone laugh"!!! Not even drunk this time!!
Argh!!!!
Argh!!!!
Looks like more and more AV softwares see Kkrunchy's signature as a virus.
"Warning: This message has had one or more attachments removed (outline.zip). Please read the "VirusWarning.txt" attachment(s) for more information."
Not to mention the guys who complained on heise.de's Breakpoint thread about a trojan.
I'm wondering if it was such a good idea to make it available to the public, even if there was already a leaked version because of some idiots that should not have been trusted.
How about making a new version spread for friends / eleet only? ;)
(I haven't tried yet but I guess hexediting the "FARBRAUSCH" bytes in the head might fool an AV or two)
End of ranting :)
"Warning: This message has had one or more attachments removed (outline.zip). Please read the "VirusWarning.txt" attachment(s) for more information."
Not to mention the guys who complained on heise.de's Breakpoint thread about a trojan.
I'm wondering if it was such a good idea to make it available to the public, even if there was already a leaked version because of some idiots that should not have been trusted.
How about making a new version spread for friends / eleet only? ;)
(I haven't tried yet but I guess hexediting the "FARBRAUSCH" bytes in the head might fool an AV or two)
End of ranting :)
in the headER :)
Respect.
Hmm, i thought i had thumbed it up...ok...here it comes.
needs no comment - just brilliant ^^
Holy smokes. Up to now I've been using upx with optimized settings, and I just tested out kkrunchy... Saved me an extra 10k (41k vs 51k)!!! 0.o
I also tried it on some old stuff I did, and the improvement is amazing over the other packers I've been using. Guess what I'll be using now, this stuff is absolutely awesome :-)
I also tried it on some old stuff I did, and the improvement is amazing over the other packers I've been using. Guess what I'll be using now, this stuff is absolutely awesome :-)
There is a new version of ANDpakk2, called apk2_v0_18, and it beat kkrunchy's pack ratio on all files, that i tried to pack. Who knew, maybe new version of kkrunchy is out before ASSEMBLY and we get great intros packed with kkrunchy, not apk2.
current eval versions are here: http://www.farbrausch.de/~fg/kkrunchy/test/ ("neu" means "new", but there are dates too).
apk2 0.18 vs current kkrunchy0.23a3 (yes, i don't increase the version number because these are not "official" releases yet):
*: not the released version, but a pre- or post-party version (depending on the production)
i stil use for testing to get meaningful comparisions with previous results.
**: kkrunchy uses a 512byte-aligned file layout that is compatible with windows 9x,
apk2 doesn't; lines marked like this mean that the size with kkrunchy is the size
of the apk2-packed file rounded up to the next 512b boundary. all relevant size
limits are multiples of 512 bytes, so this shouldn't make a difference in practice.
i didn't check whether the apk2'ed exes worked (i know they work with kkrunchy because these are a random sample of my regular test cases). apk2 sizes include the progress bar because it's on by default; feel free to subtract whatever that costs from the apk2 results if you want to, but remember to take the 512b alignment thing into account.
so i guess you'll just have to try both out. i'm pretty happy though that the two have quite similar compression ratios, but kkrunchy (de)packing is consistently faster by a factor of 1.7x or more :)
apk2 0.18 vs current kkrunchy0.23a3 (yes, i don't increase the version number because these are not "official" releases yet):
Code:
name packer size time comparision
che guevara* apk2 62560 3.06s >
kkr 56320 1.84s
fr-019 apk2 54432 3.11s >
kkr 51200 1.70s
debris* apk2 179744 12.80s <
kkr 181248 6.81s
fresnel apk2 52608 2.30s =
kkr 52736** 1.39s
mem. from the mcp* apk2 63424 4.67s =
kkr 63488** 2.54s
project genesis* apk2 54800 9.00s >
kkr 53760 4.10s
southside works apk2 55984 3.92s >
kkr 55808 2.15s
kings..playground apk2 57664 2.72s =
kkr 57856** 1.57s
kkrieger* apk2 90960 4.58s >
kkr 90624 2.67s
memento apk2 44608 7.89s <
kkr 47616 2.90s
unrel. v2m musicdisk apk2 67216 31.38s <
kkr 70144 10.12s
*: not the released version, but a pre- or post-party version (depending on the production)
i stil use for testing to get meaningful comparisions with previous results.
**: kkrunchy uses a 512byte-aligned file layout that is compatible with windows 9x,
apk2 doesn't; lines marked like this mean that the size with kkrunchy is the size
of the apk2-packed file rounded up to the next 512b boundary. all relevant size
limits are multiples of 512 bytes, so this shouldn't make a difference in practice.
i didn't check whether the apk2'ed exes worked (i know they work with kkrunchy because these are a random sample of my regular test cases). apk2 sizes include the progress bar because it's on by default; feel free to subtract whatever that costs from the apk2 results if you want to, but remember to take the 512b alignment thing into account.
so i guess you'll just have to try both out. i'm pretty happy though that the two have quite similar compression ratios, but kkrunchy (de)packing is consistently faster by a factor of 1.7x or more :)
woops, just noticed that i didn't take the current dev version, but the last semi-stable one before that. i put the current one online too (http://www.farbrausch.de/~fg/kkrunchy/test/kkrunchy_023a3_neuneu.exe).
addendum to the table above with the actual current version :):
short version: it's slightly faster and compresses slightly better (yay).
addendum to the table above with the actual current version :):
Code:
che guevara kkrn 55808 1.82s
fr-019 kkrn 51200 1.67s
debris kkrn 179200 6.67s
fresnel kkrn 52736 1.40s
mem. from the mcp kkrn 62976 2.51s
project genesis kkrn 52736 3.89s
southside works kkrn 55296 2.07s
kings..playground kkrn 57344 1.60s
kkrieger kkrn 90112 2.65s
memento kkrn 47104 2.95s
unrel. v2m musicdisk kkrn 69632 10.06sshort version: it's slightly faster and compresses slightly better (yay).
I really wonder how a mashup between "southside exploration" and "eastside works" looks like ;)
Great tool though!
Great tool though!
![[self-ban]](https://content.pouet.net/avatars/intro.gif)
er. southside exploration. sorry. i only have the filenames to go by ("southside_unpacked.exe") and obviouslyly misremembered :)
oh yeah.. the 023a3_neuneu version did shave of 5000 bytes more then the 023a version on my current project. That means I can add more effects!! Thanks, Ryg! :)
why were you using 023a in the first place if depack time is not an issue? 023a2 is publicly available since the initial public release and typically compresses a lot better :)
Very awesome little tool. :)
AVG killed my precious demo collection :(
most recent test version (asm07 special :) at http://www.farbrausch.de/~fg/kkrunchy/test/kkrunchy_023a4_asm07.exe. somewhat slower than previous versions, due to some tweaks that increase run time (and memory usage) but increase compression, as that seems to be what people want. whatever :)
Respect.
I really love that packer - but my dumb AV software does says "Infected with: Packer.Krunchy.B" ... uhhh ohhh ... wtf??! ...
Funny virus name IMHO ;)
I'm using BitDefender Antivirus 10 ... -.-
Funny virus name IMHO ;)
I'm using BitDefender Antivirus 10 ... -.-
of course I meant
"virus" name
... everybody knows that these EXEs are right - it's just the signature - but how do software developers from AV products come to that "conclusion" ????!
"virus" name
... everybody knows that these EXEs are right - it's just the signature - but how do software developers from AV products come to that "conclusion" ????!
complain to them via mail. the more complaints arrive at their doorsteps, the better.
Are the output .exe's compatible with Vista? I'm coding a demo and my dad can't run it on Vista but it's fine with XP SP2.
Great ppacker though :)
Great ppacker though :)
works with the exes i checked, but the easiest way is to find out for sure is to just give him an uncompressed exe and check that too (don't have access to any vista machines atm).
Useful! Not only for demos.
kkrunch me baby one more time.
How does this compare to kkrunchy? The don't seem to mention kkrunchy in their compression comparisons and I have no Windows machine to test it on.
standard de-facto in pc-scene, best packer even I've seen
Smallest Joke ever! kkrunchy.exe --> in kkrunchy.zip ;)
Thanks too :)
:) :)
The mother of a packers, at least until crinkler came out (which is targeted more for the 4k audience).
you killed the 64k scene, bastards!
Great tool!
cout << "Crunched bytes, powerful than ever." ;
i've just released the code for disfilter, the x86 code preprocessing filter used by kkrunchy, and IMHO the most interesting part of the whole package (it's certainly the most unique part about it). i've been meaning to do this for quite some time, but the code was badly in need of cleanup and i just never got started on that until last weekend. :)
get it here if you're interested.
get it here if you're interested.
magnificent :)
Although it rulez. Ad Aware blocks it process (it does not allow me to compress stuff while running).
And deletes compressed stuff :/
And deletes compressed stuff :/
I guess I'm forced to thumb it up: It helped us so much!
Thanks a million!
Thanks a million!
"ERROR: files with exports or tls data are not supported"
im bad with compilers and not sure what that means. anyone have an idea on what settings i need to do in VS2008 to fix this?
im bad with compilers and not sure what that means. anyone have an idea on what settings i need to do in VS2008 to fix this?
Mandatory thumb. 64k would be a pretty dry category without it.
button: Eg. don't use Delphi/Borland C++ to make demos, they generate TLS callbacks. Also DLLs won't pack because they contain exports. Same goes for exported functions in EXEs.
that said, mandatory thumb for inspiring me to make my own personal PE packer/protector!
that said, mandatory thumb for inspiring me to make my own personal PE packer/protector!
.
mother
submit changes
if this prod is a fake, some info is false or the download link is broken,
do not post about it in the comments, it will get lost.
instead, click here !
link is to the website. i'll update it when newer versions become available :) - and yeah, it's not really a demotool, but i guess kkrunchy is more demotool than all other packers out there with the exception of crinkler :)