Hey this is my work, although the exploits etc aren't my ideas by any means.
This is just a way to wrap the whole thing up in a much less annoying package than having to patch the nvidia driver. I originally got inspired by the CMP90/50 windows unlocks that run as standalone UEFI modules that chainloads into the OS.<p>Happy to answer questions!
Thanks so much. 27b is a favourite model and last night my dev ran some benchmarks to achieve 133tks about 3.2x which changes the economics.
Super well done, I wished all mods took your approach. Patched drivers are a nuisance, especially if you have more than one kind of card. Thank you.<p>edit: now do the 4090 ;)
Yeah I hate the patched drivers too, I have a CMP170 as well I want to see if I can find a way to get fully working the same way without or minimal driver patching.<p>4000s is WIP, forks has already unlocked it but I'm doing some refactoring to better support different families and after that I'll port it over :-)
Can't wait, I have two 48G modded ada 4090's and 4 stock Blackwell 4000's (though I doubt those will be as easy). The 4090's alone would be an amazing thing to have boosted, they're my main workhorses and effectively any improvement there will immediately make my days more productive. If I can help out with testing, my email is in my profile.
This one is real, but there are some caveats:<p>- you will need to run your card hotter than you might be comfortable with<p>- the real pay-off (on some cards, apparently, see downthread) requires increasing the maximum power to 450W<p>- This is well above what the stock cooler was designed for so a bit more cooling would help<p>What's nice is that you can stack this mod on top of (or rather, below of, this one loads first) the p2p mod that has been out for a while, they don't seem to interfere with each other.
It's quite nuanced. In a nutshell the unlock just flips the 2 bits that halves how often the core can issue the specific instructions Nvidia locks to full rate on the professional and datacenter cards.<p>Obviously doing more compute requires more power but what it does is shift the card down on the frequency curve to keep the wattage in check.
So you're right that at stock power limits it hits a cap where the full issuing rate isn't reached but it depends heavily on the exact workload how much juice is left.<p>The frequency curve also takes heat into account so in general you end up with a nice perf boost for a slightly hotter card.<p>I've been running this on my 2 3090s for a month now with no issues but they are also water cooled :-)
> - you will need to run your card hotter than you might be comfortable with<p>Can you explain this one? On my GPU I didn't have to change the power limit. Obviously the card might spend more time at max power usage, but it's still within the limit.
If you increase the power limit on the card you will be able to get much more mileage out of this modification. At 450W the difference is much, much larger but that is not something I'm comfortable with doing for longer runs. Note that the stock settings are already pretty extreme for that cooler, and that ambient temp and free airflow are going to be impacting the resulting GPU temp in a very direct way. I have my 3090's in an external box (with its own separate PCIe backplane) so they can talk p2p and have 'fresh air' rather than in the enclosure.
Is your external box something you put together or is it a product you purchased?
Yes, but you can get up to a 40-50% increase in some workloads even with the same power limit, and some cards have it locked anyway. I just felt like it was a litle misleading for you to mention the power in your original post, since the unlock works regardless of that.
Interesting, I have 7 3090's here of different plumage and they all require a power increase to get > 25% gain with this modification, would you mind sharing the type of card you have and how it is cooled?<p>Obviously, if there are 3090's out there that respond differently then I will be happy to retract that.
These number claim 240W locked power draw on 4070:<p><a href="https://github.com/JustAResearcher/gpu-unlock-ada/releases/tag/v1.1.1" rel="nofollow">https://github.com/JustAResearcher/gpu-unlock-ada/releases/t...</a>
I've tested myself and the results are legit, this is incredible, even if a little risky (it relies on exploiting firmware): <a href="https://github.com/Zagitta/gpu-unlock/blob/main/docs/How%20Ampere%20was%20unlocked.md" rel="nofollow">https://github.com/Zagitta/gpu-unlock/blob/main/docs/How%20A...</a><p>> Unfortunately for Nvidia the booter is vulnerable to a classic buffer overflow that writes straight into the stack which gives us the ability to execute code via Return Oriented Programming (ROP).<p>Repo for RTX 4090 and 4070 Ti Super: <a href="https://github.com/JustAResearcher/gpu-unlock-ada" rel="nofollow">https://github.com/JustAResearcher/gpu-unlock-ada</a>
Really interesting.<p>On the other hand, this only applies to a limited set of workloads (training is often memory bandwidth bound for example).<p>I also wonder how this affects power consumption.
Training is compute bound.<p>If its memory bandwidth bound you need to raise your batch sizes to leverage more parallelism.
It's also useful for prefill in inference, VLLM sees up to 30% boost depending on the exact kernels and prefill size
Uh, no. This headline is wrong. It's FP32 unlocked from being halved, not 16 bit ops. And to be clear, over the last week I've looked many places and I've not seen <i>any</i> reports of this unlock being applied to any 3xxx card except the 3090/3090ti series. Even if it might apply more generally in <i>theory</i>.
It's the FP/BF16 FMA to FP32 instructions.<p>And also it does work on most 3000 series cards as they all run the same GA102 booter that's vulnerable :-)
There are some reports of cards with weird vbios layouts having issues but i'll try to get that fixed
The repo has fp16 numbers.<p>Also this one for 4070s shows that its for fp8 as well:<p><a href="https://github.com/JustAResearcher/gpu-unlock-ada/releases/tag/v1.1.1" rel="nofollow">https://github.com/JustAResearcher/gpu-unlock-ada/releases/t...</a>
The repo has a lot of reports of people applying it to other 3xxx card, see <a href="https://github.com/Zagitta/gpu-unlock/issues" rel="nofollow">https://github.com/Zagitta/gpu-unlock/issues</a><p>There's at least 3080 Ti, 3070 Ti, 3070, 3060