About two years ago, I switched teams at Google to focus on fully homomorphic encryption (abbreviated FHE, or sometimes HE). Since then I’ve got to work on a lot of interesting projects, learning along the way about post-quantum cryptography, compiler design, and the ins and outs of fully homomorphi...

It’s April Cools! Last year I wrote about friendship bracelets and the year before about cocktails. This year it’s parenting. Parenting articles are a dime a dozen and always bury the lede behind a long story. I’ll skip that. How to think about your child and your role as a parent These are framing ...

Table of Contents In this article we’ll implement a global optimization pass, and show how to use the dataflow analysis framework to verify the results of our optimization. The code for this article is in this pull request, and as usual the commits are organized to be read in order. The noisy arithm...

Table of Contents In the last article we lowered our custom poly dialect to standard MLIR dialects. In this article we’ll continue lowering it to LLVM IR, exporting it out of MLIR to LLVM, and then compiling to x86 machine code. The code for this article is in this pull request, and as usual the com...

Table of Contents In previous articles we defined a dialect, and wrote various passes to optimize and canonicalize a program using that dialect. However, one of the main tenets of MLIR is “incremental lowering,” the idea that there are lots of levels of IR granularity, and you incrementally lower di...

Can you find a set of cards among these six, such that the socks on the chosen cards can be grouped into matching pairs? (Duplicate pairs of the same sock are OK) Spoilers: If the cards are indexed as 1 2 3 4 5 6 Then the following three subsets work: $\{ 1, 2, 4, 5, 6 \}$, $\{ 2, 3, 6 \}$, and $\{ ...

Table of Contents In a previous article we defined folding functions, and used them to enable some canonicalization and the sccp constant propagation pass for the poly dialect. This time we’ll see how to add more general canonicalization patterns. The code for this article is in this pull request, a...

In cryptography, we need a distinction between a cleartext and a plaintext. A cleartext is a message in its natural form. A plaintext is a cleartext that is represented in a specific way to prepare it for encryption in a specific scheme. The process of taking a cleartext and turning it into a plaint...

Table of Contents Last time we defined folders and used them to enable some canonicalization and the sccp constant propagation pass for the poly dialect. This time we’ll add some additional safety checks to the dialect in the form of verifiers. The code for this article is in this pull request, and ...

Table of Contents Last time we saw how to use pre-defined MLIR traits to enable upstream MLIR passes like loop-invariant-code-motion to apply to poly programs. We left out -sccp (sparse conditional constant propagation), and so this time we’ll add what is needed to make that pass work. It requires t...

Table of Contents Last time we defined a new dialect poly for polynomial arithmetic. This time we’ll spruce up the dialect by adding some pre-defined MLIR traits, and see how the application of traits enables some general purpose passes to optimize poly programs. The code for this article is in this...

Problem: Compute 16% of 25 in your head. Solution: 16% of 25 is equivalent to 25% of 16, which is clearly 4. This is true for all numbers: $x\%$ of $y$ is always equal to $y\%$ of $x$. The first one is $\frac{x}{100} y$ and the second is $\frac{y}{100}x$, and because multiplication is commutative an...

Table of Contents In the last article in the series, we migrated the passes we had written to use the tablegen code generation framework. That was a preface to using tablegen to define dialects. In this article we’ll define a dialect that represents arithmetic on single-variable polynomials, with co...

Table of Contents In the last article in this series, we defined some custom lowering passes that modified an MLIR program. Notably, we accomplished that by implementing the required interfaces of the MLIR API directly. This is not the way that most MLIR developers work. Instead, they use a code gen...

Table of Contents This series is an introduction to MLIR and an onboarding tutorial for the HEIR project. Last time we saw how to run and test a basic lowering. This time we will write some simple passes to illustrate the various parts of the MLIR API and the pass infrastructure. As mentioned previo...

Table of Contents Last time, we covered a Bazel build system setup for an MLIR project. This time we’ll give an overview of a simple lowering and show how end-to-end tests work in MLIR. All of the code for this article is contained in this pull request on GitHub, and the commits are nicely organized...

Table of Contents As we announced recently, my team at Google has started a new effort to build production-worthy engineering tools for Fully Homomorphic Encryption (FHE). One focal point of this, and one which I’ll be focusing on as long as Google is willing to pay me to do so, is building out a co...

Today my team at Google published an article on Google’s Developers Blog with some updates on what we’ve been doing with fully homomorphic encryption (FHE). There’s fun stuff in there, including work on video processing FHE, compiling ML models to FHE, an FHE implementation for TPUs, and improvement...

Before I discovered math, I was a first year undergrad computer science student taking Electrical Engineering 101. The first topic I learned was what bits and boolean gates are, and the second was the two’s complement representation of a negative n-bit integer. At the time two’s complement seemed to...

It’s April Cools again. For a few summers in high school and undergrad, I was a day camp counselor. I’ve written before about how it helped me develop storytelling skills, but recently I thought of it again because, while I was cleaning out a closet full of old junk, I happened upon a bag of embroid...