NVIDIA Pascal: a Great Password Cracking Tool

July 26th, 2016 by Oleg Afonin

During the last several years, progress on the CPU performance front has seemingly stopped. Granted, last-generation CPUs are cool, silent and power-efficient. Anecdotal evidence: my new laptop (a brand new Macbook) is about as fast as the Dell ultrabook it replaced. The problem? I bought the Dell laptop some five years ago. Granted, the Dell was thicker and noisier. It’s battery never lasted longer than a few hours. But it was about as fast as the new Macbook.

Computer games have evolved a lot during the last years. Demanding faster and faster video cards, today’s games are relatively lax on CPU requirements. Manufacturers followed the trend, continuing the performance race. GPUs have picked up where CPUs have left.

NVIDIA has recently released a reference design for GTX 1080 boards based on the new Pascal architecture. Elcomsoft Distributed Password Recovery 3.20 adds support for the new architecture. What does it mean for us?

GPU Acceleration: The Present and Future of Computer Forensics

Today’s desktop video cards pack significantly more grunt compared to contemporary desktop CPU’s. The powerful GPU units can deliver unmatched performance in massively parallel computations, offering 100 to 200 times greater performance compared to today’s CPUs. All this performance is still relatively useless when it comes to regular computing. The several hundred individual GPU cores are built specifically for “one code, different data” scenarios, while general-use CPUs can run different code on each kernel. Since breaking passwords involves executing the same code repeatedly, just with different data (encryption keys or passwords), a large array of GPU units makes lots of sense.

How does it scale to real-world applications? A low-end NVIDIA or AMD board will deliver 20 to 40 times the performance of the most powerful Intel CPU. A high-end accelerator such as the NVIDIA GTX 1080 can crack passwords up to 250 times faster compared to a CPU alone.

Just how important is GPU acceleration, exactly? As an example, a common 6-character password (lower-case letters with numbers) has just about 2 billion combinations (http://projects.lambry.com/elpassword/). If that password protects a Microsoft Office 2013 document, you’ll spend 2.2 years trying all possible combinations. Using the same computer, add a single NVIDIA GTX 1080 card, and the same password will be cracked in under 83 hours. That’s 3.5 days vs. 2.2 years!

bench_office_2013

NVIDIA Pascal Architecture

NVIDIA’s latest GPU architecture code-named Pascal gives a significant performance boost compared to NVIDIA’s past flagship. With 21 half-precision teraflops, GTX 1080 boards are 1.5 to 2 times faster breaking passwords compared to GTX 980 units.

pascal10x

According to ElcomSoft’s internal benchmarks, Elcomsoft Distributed Password Recovery can try 7100 passwords per second for Office 2013 documents using a single NVIDIA GTX 1080 board compared to 3800 passwords per second on an NVIDIA GTX 980. When recovering RAR 5 passwords, using a single NVIDIA GeForce GTX 1080 results in 25000 passwords per second compared to 13000 passwords per second on a GTX 980.

The New GTX 1080 Board

Those new video cards are still hard to come by. We were able to source a GTX 1080 board, and used it to build support for the new architecture directly into our products. Distributed Password Recovery 3.20 became the first ElcomSoft tool to receive support for NVIDIA Pascal architecture. So how does the new gaming monster fare when it comes to crunching innocent victims’ passwords?

In fact, it does pretty well. Just look at these benchmarks and decide for yourself whether you just want the new card or need it right away!

bench_office

bench_rar5

Cannot see the numbers for CPU-based benchmarks without a magnifying glass? In case you wonder, we were only able to try 30 (yes, THIRTY) MS Office 2013 passwords per second on an Intel Xeon E5 2603 without GPU acceleration. Compare that to 7100 passwords per second using a single NVIDIA GTX 1080 board!

Conclusion

NVIDIA Pascal is a major break-through in GPU computations. If you need a reliable powerhouse to break passwords faster, consider adding a GTX 1080 board to your workstation.

What if your computer already has a GTX 980 installed? If you have a free PCIe slot, if there is sufficient cooling, and if your computer’s power supply can deliver enough juice for an extra GTX 1080 board, then you can just add the new board without removing the old one. Elcomsoft Distributed Password Recovery will use both video cards together for even faster attacks. Does it make sense keeping a GTX 980 along with the new GTX 1080? By keeping the old card together with the new GTX 1080, you’ll get an additional performance boost of about 20 to 30 per cent. Whether this extra performance is worth the increased power consumption and excess heat is debatable, but if your power supply and cooling can reliably manage both cards working at their maximum performance, by all means go for it!

Tags: , , , , ,

Sign up for free ElcomSoft Password Recovery Software newsletter

Leave a Reply

15 Comments on "NVIDIA Pascal: a Great Password Cracking Tool"

Notify of
avatar
Darel King
Guest

Greetings, I am wondering if all GTX-1080s use the Pascal architecture? While looking at the different options there is no mention of Pascal in the specs.

RedLeader
Guest
@Darel King – yes, every GTX 1080 is Pascal. Pascal is the name for the new line of cards on Nvidia’s 16nm production line. It includes (so far) the 1060, 1070, 1080, and just now the newly announced Titan X (there are two of these, so make sure you look for the one announced in July 2016). The Titan X is a further ~35% performance improvement over the 1080, so if you thought these numbers were impressive, the Titan X will go even faster! @Oleg: could you provide some benchmarks with AMD cards? I know they don’t have a current… Read more »
Vladimir Katalov
Admin
@ReadLeader, thanks for the answer on Pascal architecture! AMD benchmarks are on the way (though we do not have all the latest boards, in contrary with NVIDIA). I should say, however, that we concentrate on NVIDIA for several reasons. The main one is: AMD drivers are really buggy, and quite often the new release causes our software to crash, so we have to implement some kind of workaround. Even more, problems with AMD drivers sometimes cause *system* crash (on high loads). And last but not least, AMD cards have problems with cooling, they often overheat, and definitely not intended for… Read more »
RedLeader
Guest

Thanks for the insight. The last bit there is interesting, because I know the last 2-3 generations of AMD GPUs were prized for cryptocurrency mining operations 24/7 which caused the secondary market to get all kinds of crazy. I know they’ve been largely phased out for ASICs, but I assumed that the crypto mining was a similar operation to these password cracks. Can you speak to their similarities or differences?

Anonymous
Guest

Limit the amount of tries some one can get each day to type in wrong password and it becomes irrelevant, because you no longer can brute force it.

RedLeader
Guest
That’s not how this works. What these GPUs do is work off of downloaded password sets that are encrypted. That’s why you often hear news about “160,000 passwords stolen” or “20,000 customer credit cards stolen”. The thieves are downloading the hash-encrypted sets and then using the GPUs to crack the passwords in massive volumes at a time. Since most people use the same password or variations of the same password, the thieves can then use cracked passwords and usernames from Website #1 on Website #2, Website #3, etc. There is no need to repeatedly hit the same login prompt over… Read more »
JuanPC
Guest

Can’t wait for the AMD tests,..

JuanPC
Guest

In some tests I’ve seen: Luxmark 2.0 Fury X does not improve, and GTX980 kills AMD,
But in Luxmark 3.1 things change in favor of AMD, both using OpenCL. Luxmark 3. also has a C++ mode that kills OpenCL. Very disappointing.

Vladimir Katalov
Admin

JuanPC,

Unfortunately, we do not have all modern AMD cards for testing (in contrary with NVIDIA; now we run banchmarks on new Titan X). GTX 1080 is an absolute winner so far, and in general, we prefer NVIDIA over AMD for many reasons — simpler development (with CUDA), much better and stable drivers, less problems with cooling, and top suppoprt from NVIDIA.

JuanPC
Guest

I thought Render tests could translate 1:1 to brute force password hacking, until I saw that.
Worse case gtx1080 has the same pwrds/sec as the HD7970.
Also HardDrive vs. RamDrive, I’ve seen hard drive performs faster with FlacCL.
Crazy, unbelievable!

JuanPC
Guest

Also there is FPGA PCIe cards that claims to perform 15x faster than HD7970 in render tests.

JuanPC
Guest

For some reason I don’t understand, Nvidia kills AMD in Folding@Home,
But in Boinc things change.

Guest
Guest

I wonder how fast DGX-1 would benchmark. Any speculations?

Vladimir Katalov
Admin

Well, that system works on Linux, while our software is currently for Windows only. But you can estimate the performance itself (DGX-1 uses 8x Tesla P100 GPUs).

Brian K Wallace
Guest
Brian K Wallace

Vlad – are NVIDIA K80 GPUs supported? Reason is that I could spin up some AWS P2 instances with up to 16x GPUS (2496 cores ea) and 64 vCPUs each and really tackle some internal pwd testing for some of our medium/large corporate clients. I did tried one of the elcomsoft tools on a G2 instance with 1536 tesla cores but it didnt run anywhere near as fast as I was expecting…

wpDiscuz