Compiling Backend Code to a Colder Chip
April 21, 2022

Backend infrastructure can move from instruction-heavy servers toward a colder, massively parallel chip without abandoning the business logic that gives the system value. Brian Greenforest’s vision compiles languages, databases, schemas, procedures, and configuration into reconfigurable silicon.

The result promises immediate queries, elastic capacity, redundant execution, stronger object-capability isolation, and a dramatically shorter path from an engineer’s change to deployed computation.

Move the Whole Backend, Not One Kernel

The architecture absorbs PHP, C#, Python, Perl, Node, COBOL, C, C++, web-server rules, DNS, email configuration, SQL schemas, stored procedures, and proprietary logic as parts of one system.

A compiler must preserve names, state, data shape, ordering, observability, and access rights while reorganizing the work into parallel hardware. That continuity makes migration commercially meaningful.

Turn Reconfiguration Into Elastic Service

FPGAs can reshape resources around momentary load. A multi-tenant object-capability runtime can allocate that fabric while keeping each customer’s data and business logic under explicit ownership.

Cloud providers, backend teams, and semiconductor partners can collaborate on this machine. Bring a demanding backend path and help turn colder execution into a production platform.

Build the Fabrication Path Behind Compiling Backend Code to a Colder Chip

The Missing Maker Fab carries this work into the tools and processes required to fabricate fast active devices locally.

The Missing Maker Fab · https://youtube.com/watch?v=VbxcAJiBkps&feature=share

Originally posted on LinkedIn

Brian Greenforest · (2022-04-21 17:32:51 UTC)

Open the original LinkedIn post · LinkedIn activity 6922960356663332864

LinkedIn status when archived: Visible to anyone on or off LinkedIn.

Soon you'll be able to parallelize and get a supercompute concurrency speedup, putting ALL your custom backend code, no matter PHP, C#, Python, Perl, Node, COBOL, C, C++, Nginx configs, DNS, Email configs with any SQL server, with ALL your tables and schemas, with all your stored procedures and very sensitive business logic. Transparently compile your entire infrastructure, run it in on a COLDER CHIP (a very expensive chip though; but it'll manage multiple REDUNDANT copies of your server infrastructure) that will enable unimaginable capabilities! Your customers will never wait complex queries to complete. Your engineers will love you because they will deliver 10x quicker. Your hiring budget will be optimized to favor #AI, a trade off of Machine Learning and Advanced In-Memory Superscalable Silicon Compute. It'll be VERY expensive to buy a chip, but savings are insane! You'll run your backend code in the cloud, sharing your own super-sensitive resources in a PERFECTLY SECURE Object Capabilities and Object Oriented Reactive Programming auto-parallelizing environment, absolutely elastic to satisfy your changing needs, and EVEN YOUR MOMENTARY LOADS (that's what FPGAs are good at: they can be reconfigured in the "field" based on current statistical analysis of data). Choose to WAIT when we finish building THIS, or simply start collaborating with us directly NOW. Just PM me, I don't byte.

Original LinkedIn media

Cloud Compiling
Link preview thumbnail shown with the original LinkedIn post. Open the saved full-resolution image.

Comments added by Brian Greenforest on LinkedIn

These 2 comments were also preserved verbatim from Brian Greenforest’s LinkedIn data export or the public post page.

Comment 1 · (2022-04-21 17:36:37 UTC)

Pardon me? COBOL an d PL/1? 😎 😅

View the LinkedIn post

Comment 2 · (2022-04-21 17:39:54 UTC)

Our advanced transpiler will be able to preserve your high-level class structures, variable names, table names. We'll encourage you to stop writing in the old language though that new reactive programming model does not have call stack, no concept of "current line being executed", unless you're running a special heavy-weight numbers crunching DSP algorithm. Objects and classes will work just like in normal OOP. Again, the answer to "what's the catch?" is the ENTRY COST of running your code on such a platform. But you'll make more money and scale faster in the long run! Your TEAM WILL LOVE YOU!

View the LinkedIn post