Meriken's Tripcode Engine No. 10

(Continued from )

This is a support thread for Meriken's Tripcode Engine.

3.2.11 (June 3, 2016)
Add Pascal (GTX 1070/1080) support. Add AVX/AVX2 support for POSIX systems. Improve build script.

"Meriken's Tripcode Engine" is a cross-platform application
designed to generate custom/vanity tripcodes at maximum speed.
It is arguably the fastest and most powerful program of its kind.
(This program actually runs faster than the original mty or
MTY_CL on the same hardware!) It makes effective use of
available computing power of CPUs and GPUs, and the user
can specify flexible regex patterns for tripcodes. It features highly
optimized, extensively parallelized implementations of bitslice
DES and SHA-1 for OpenCL, AMD GCN, NVIDIA CUDA,
and Intel SSE2/AVX/AVX2.

"Meriken's Tripcode Engine" is free software, licensed
under GPLv3 and available for free download at the following link:

github.com/meriken/merikens-tripcode-engine-v3

Other urls found in this thread:

tripcode.ygch.net/yggdrasil/
nmav.gnutls.org/2011/12/self-modifying-code-using-gcc.html
stackoverflow.com/questions/31734263/write-x86-asm-functions-portably-win-linux-osx-without-a-build-depend-on-yasm
darkside.com.au/bitslice/
csrc.nist.gov/publications/fips/fips46-3/fips46-3.pdf
twitter.com/SFWRedditGifs

A few serious bugs in the POSIX version have been ironed out. The app should be pretty stable both on Windows and Linux.

Actually, the porting to Linux went surprisingly well. I'm pretty sure I'll be using using Ubuntu for my next tripcode rig.

The next step is to update my distributed tripcode generation service:

tripcode.ygch.net/yggdrasil/

I want to add an English interface, auctioning, and trading at the least.

Can you please stop posting?

Is there any performance difference between Windows and Linux?

There shouldn't be any difference except for a slight speed reduction (approx. 20%) with CPU for Linux. I just didn't feel like porting the self-morphing assembly code. It should be negligible given the speed of GPUs, though.

This doesn't look all that hard... I will work on it when I have time.

nmav.gnutls.org/2011/12/self-modifying-code-using-gcc.html
stackoverflow.com/questions/31734263/write-x86-asm-functions-portably-win-linux-osx-without-a-build-depend-on-yasm

I never understood.... what exactly does this do that is new or in any way unique?

yes

how did you get the algorithm though?

Adding some helpflags to the executable would benice, or a man page.

>putting all this effort into generating a tripcode when 99% of people will just immediatly filter you for using a trip

Four things:

(1) Self-modifying code for OpenCL/GCN to reduce overhead for referencing arrays, particularly the DES expansion function,
(2) Maximizing register/instruction cache usage and occupancy for bitslice DES using GCN assembler,
(3) Preparing 4096 kernels corresponding to 4096 salts at the compile time, and
(4) Developing new S-boxes for the NVIDIA Maxwell architecture.

I was the first one to successfully implement (1) and (2), and the credits for (3) and (4) go to the wonderfully capable DeepLearningJohnDoe.

I started with John the Ripper's bitslice DES implementation 4 years ago and has been modifying it heavily since then.

"have been"

I will make sure to take care of that.

These links should help you if you want to know more about the algorithm:

darkside.com.au/bitslice/
csrc.nist.gov/publications/fips/fips46-3/fips46-3.pdf

You know, I developed this program originally for 2ch.net, and the culture there is less aversive to use of tripcodes. Oh well.

You ported it to Linux?

I did. The app should be fully functional on Linux now. The porting wasn't that easy with threading and IPC.

What parts were difficult?
Roughly how long did it take you?
I haven't found much documentation, but what's the general algorithm for generating a tripcode, if someone wanted to do something like this, but in a different language?

The most difficult part was that so-called "cross-platform" libraries such as std and Boost weren't all that dependable, especially with threading and IPC. Boost.Process 0.5 was particularly difficuly in that regard.

It took me about a month to port the whole thing to Linux. Much less than I thought, and the outcome was surprisingly good.

Why did you decide to make A POSIX port?

I ported it just because people here at Sup Forums asked for it. "The more, the merrier," you know.

>Please note that NVIDIA-optimized versions (-enable-cuda-des-multiple-kernels-mode) take extremely long time to build.
How long?

The general algorithm has three parts: (1) random key generation, (2) bitslice DES and (3) matching.

Basically you generate trillions of random tripcodes to see if there is anything that matches specified patterns.

Well, thank you! It's appreciated.

On my PC, it takes 4 to 8 hours depending on the platform.

No shit. What does it do in those hours?

Does anyone has the link for the tripcode engine that was optimized for just one gpu model?

>このページを表示するには、広告のブロックを解除してください。
>Please disable the ad blocker to browse this page.

Japanese is a cute language. Cute.

I either sleep or go to work :) I only build that version once a month or so. It's very painful.

>using any tripcode but a secure tripcode
How'd your first year of CS go?

I think secure tripcodes are fundamentally lacking in that they are not transferable to other online forums. We can always use more secure 12 character tripcodes on online forums in Japan, but I suppose Sup Forums is different.

I just hope the app is working as expected on Linux. Please let me know if there are any problems.

I know the NVIDIA-optimized version is getting ridiculously huge...
I will do something about that.

Sorry you haven't been receiving much feedback, but it's working fine here on my end

Great! I suppose this is one of those situations where you only hear from people when there are problems.

It's faggots like you that promote tripfagging, literally end yourself.

this

fucking kill yourself with your reddit cancer namefagging aids

I figured out how to create such binaries without significantly modifying the existing code base. I will include them in the next version.

you are the fucking best
thankyou

You are welcome :) I'm building Windows binaries of the next version right now. I hope they would be ready before this thread ends.

s-so h-how would I go to crack a tripcode found on this board? do I just copy&paste the trip like "HSA !!HBro02/b3bh" in the patterns.txt under #regex and then spend $1000 on amazon ec2 to get it done fast?

please kill yourself. you're a worthless faggot.

just hide threads if your autism gets triggered this easily.

That's a secure tripcode, so you cannot crack it.

Vanity tripcodes are fun.
I do only use them in select threads when I do contribute something useful.

I stumbled upon that one with your program, didn't even ask for it. I think it automatically adds these to your online DB, right?

I tried an EC2 setup for this. It doesn't work unless you cluster and at that point you might as well just buy a good GPU for all the cash you're burning. The GPUs they have there are slow as molasses.

Drastui dimitry

Just kill yourself you disgusting vile brazilian piece of shit and stop shitposting on Sup Forums.

You are just a filthy rodent. No one likes you filthy brazilian nigger rodents.

Even with tripcodes that are not "secure" ones, you need to spend way more than $1000. The last time I checked, g2.8xlarge instances cost $2.6/h, so $1000 only gives you 16 days. Now g2.8xlarge instances seem to feature 2 Kepler-based GRID K520s, which aren't all that suitable for tripcode generation. At the kind of speed you get with Kepler-based GPUs, it would take forever to crack a tripcode. I'm talking about decades if not centuries here. Hope this helps you to put things into perspective.

I really like them, too. Hoping to get another one like !EEEEEEEEEE.

>didn't even ask for it

That sounds a little strange... Were you using Meriken's Tripcode Generator instead of Meriken's Tripcode Engine? Did you get it through Yggdrasil?

Yeah, I'm using MTG Free. I'm only running it occasionally though, I don't need to crunch for trips right now, and I don't want to kill my GPU prematurely.

I used to run MTE (or was it Tripcode Explorer? I don't remember) on a pre-production server that had to be decommissioned at work back in 2013.

I just told them it was a compiler when I was asked about it.

>2.1.1
Whoops, might need to update.

I was wondering if it was you. That doesn't explain why you got !ssSsssssss, though. Strange, strange...

I forgot to mention that I developed the algorithms for random key generation and regex tripcode matching by myself. I hope to write about them sooner than later.

Thanks Meriken!

According to the download page for MTG, you said:
>無料版では分散トリップ検索サービス「Meriken's Tripcode Yggdrasil」への参加が必須です。一部の特殊トリップが自動的にYggdrasilに回収され、検索速度が最大で15%ほど低下するので注意してください。

「一部の特殊トリップ」← Isn't that what you meant by it?

Also
>検索速度が最大で15%ほど低下するので注意してください。
Is this true?
Should I switch to MTE instead then?

MTE is definitely faster than MTG Free, but you are helping others with MTG Free. It's entirely up to you.

Nice tripcode! Glad to hear the app worked for you.

did it work

Good! The Windows binaries for the next version are almost ready. I will definitely switch to gmake the next time for parallel compilation...

Ah yeah, I've got one more question.
Are tripcodes found for e.g. me by other people via Yggdrasil made public to them?

No, they are all private.

how long will it take my quad 980 tis to crack your trip

Yes.

Literally just end yourself you flaming faggot. No one likes your shit, stop samefagging your shit thread and fuck off back to 2chan and reddit you shit stain.

All I can say is that you could try it if you have $10,000+ to burn and are willing to wait for years.
Besides, if you really had 4 980 Ti's, this app would give you a pretty good estimate.

fuck you, faggot.

gtfo with your reddit tier namefagging.

>do I fit in yet?
tripfags != namefags dumb false flagger

What makes secure tripcodes uncrackable?

salt but unsecure tripcodes are virtually uncrackable as well. You really have to dedicate weeks of non-stop cracking on a typical setup to crack a tripcode.

bump

I just created a manpage. The build process is taking forever, though...

With GPU tripcode generation it takes days.

>You need the following tools to build Meriken's Tripcode Engine.
>Visual Studio 2015 Community

fuck outta here

>look mom i'm being a hater on the internet

You guys are very optimistic... That wouldn't be possible even with four 980 Ti's.

Oops... I mean and

bump

It looks like this thread won't survive until the build ends. Oh well. I will make sure to upload the next version tonight. I will come back when I upgrade my distributed tripcode generation service.

go fuck yourself