Andrew Fitzgerald retweeted
I'll use this as an excuse to solicit more consumer/app dev involvement in the SIMD process. Would love to see more banter from informed devs around protocol changes!
This was discussed in depth as a tradeoff.
That said, we have the power to change things. I've heard a lot of complaints about the shorter expiry time, so let's fix it.
Andrew Fitzgerald retweeted
Durable Nonces will be deprecated next year!
It will move out of the protocol and into onchain programs.
Now is the time to try out the new solutions and give feedback to make the tooling around it great for the migration period.
Everything you need to know in the deprecation guide:
solana.com/upgrades/durable-…
Andrew Fitzgerald retweeted
The Alpenglow feature gate activates on testnet tomorrow.
Feature gate flips at epoch 1042, ~17:30 UTC Thursday Sept 24, starting a 5,000 slot countdown. ~17 minutes later, Alpenglow takes over.
We'll post when countdown starts and when Alpenglow is officially running. ⏱️
Andrew Fitzgerald retweeted
Alpenglow is coming to testnet this week. 🧵
Testnet exercises the exact migration devnet and mainnet-beta will run.
The Alpenglow community cluster rehearsed the migration numerous times and has been running Alpenglow for over 4 months.
Now Alpenglow moves to the formal activation path.
Andrew Fitzgerald retweeted
250ms slots live on mainnet-beta. Last stop: 200ms.
250ms slots activation now pending on mainnet-beta.
400 → 350 → 300 → 250ms
Reminder, SIMD-0525 features take effect one epoch after they activate thus will go live at the epoch 1037 boundary ~05:01 UTC, Friday Sept 18.
- epoch 1035: pending activation
- epoch 1036: feature active
- epoch 1037: 250ms slots live
1 more step to 200ms.
Andrew Fitzgerald retweeted
Transaction V1 (SIMD-0385) is live on mainnet-beta.
Max transaction size grows from 1,232 to 4,096 bytes. ZK proofs, large multisigs, BLS signatures, and confidential transfers that required multiple transactions now fit in one atomic operation.
V1 also moves resource requests (compute unit limit, priority fee, loaded accounts data size, heap size) into the transaction header instead of compute budget instructions.
Andrew Fitzgerald retweeted
Omg omgomg its coming!!!! If your transaction parsing fails in 6 hours here is the guide on how to fix it!
solana.com/upgrades/larger-t…
I still wish they didn’t call them accounts and instead called them “pieces of state”.
More accurate and far more fun - we all need a bit more pirate humor in our lives
Andrew Fitzgerald retweeted
Transaction V1 activation on mainnet-beta is moving to the start of epoch 1035, expected Sept 15 at ~1:20 AM UTC.
Teams across the ecosystem told us they needed more time to test and integrate V1 support. We listened.
How to get ready:
- V1 is live on devnet, test your integration now.
- Compute budgets in V1 are set using a new transaction config. Compute budget instructions still work on legacy and V0 but no-op in V1.
- App developers: even if you don't plan to send V1 transactions, some changes may impact your app. Review the guide below.
Details: solana.com/upgrades/larger-t…
somehow OpenAI models solved navier stokes, but also forgot to include salt in a beef stew recipe for me
Andrew Fitzgerald retweeted
if you are running an application on solana, please please please run this page through your LLM
Transaction V1 is about to go live soon. And the whole network is getting ready in case things break!
Please update your dependencies. And run your apps on Devnet if you can
@solana @solana_devs
solana.com/upgrades/larger-t…
Andrew Fitzgerald retweeted
wincode has surpassed 5M downloads 📦📈
Our Rust crate started as a fix for serialization performance in the Agave validator client.
RustSec now lists it as a recommended bincode replacement. Built to make Solana faster, now adopted across the Rust community.
Credit to Anza IBRL engineering culture who keep pushing the envelope on speed.
How it works by wincode creator @zbr0wn: anza.xyz/blog/wincode-bincod…
