I connected with supdoggie and he indicated the real problem here is gas metering but generally agrees lowering the gas limit is the best immediate solution. Also Assaf indicated that lowering the gas limit or increasing performance were our only options, outside of improving the metering again which is non trivial and has been worked on extensively in the past by cashmaney, assaf, etc.
As for the issue of pushing out an emergency fix. Shade is already refusing to push out updated contracts that are more efficient until they go through testing and audits. In an emergency 24 hours would not be enough time for shade to test and do proper audits on new contracts. Is he suggesting that they would skip security audits in an emergency?
Furthermore, the gas limit, even when lowered would not be too low for them to push out a temporary contract that simply paused functionality, which would be the goal in such an emergency, right?
So the reason austin listed isn’t valid on several levels.
Leaving as is, is not an acceptable solution for anyone who cares about deposits / withdraws working during periods of time where several people use the inefficient contracts simultaneously. Since deprioritizing this would be harmful to the price, the path forward seems obvious at this juncture.
Thats the current status from discussions on this topic.
I think we will need to have some additional discussion about what are the effective ways to improve performance. Lowering gas limit may be helpful, but from my data, the heaviest transactions are around 3,8M gas, so lowering gas limit to 5M will not effectively block those transactions.
Separately, Labs is working with Shade to test the optimized Swap contracts, and we will share some results on that during the week.
could we move everything to the nvidia tee? maybe this direction will get enough funding to solve the tech debt because it represents new exciting territory.
Nvidia tee isn’t for the same work loads as sgx (or tdx). It would also require every single validator have an extremely expensive Nvidia GPU. We’re talking about tens of thousands of dollars just for the graphics card. Nvidia tee makes more sense as a worker, that isn’t required for each node on the network.
Moving from sgx to tdx may be a better long term solution though. That would simply require 4th gen xeon or newer, not as expensive as requiring nvidia for all nodes.
There is an idea Guy floated about having sidecars though. Which would be tee’s dapps could use dedicated to their dapp only, so all nodes didnt have to do the same computations. This is another longer term idea, but it’s extremely interesting.
could you expand on what the chain capabilities would look like if this WAS a requirement? I would love for you to be rich enough that this wasn’t an unreasonable investment.
It’s not about what i personally can afford (a little confused on that comment), unless you want like 3 validators on the chain. I could certainly get my hands on one if i wanted to buy one today😅
But the abilities of the chain, do not increase compared to having Nvidia required vs as workers. You can have a small handful of nvidia tee workers and the chain gets the benefits of the new tee, without enforcing everyone has them. Also keep in mind…. Nvidia workers ARE coming already. They are currently testing them.
