Midterms 2026See who we think should earn your vote, based on our standardsThe guide →
WRITTEN IN PLAIN AMERICAN ENGLISH.
CLAY TRIBUNE.
Advertisement

AMD’s random number generator produces no zeros, and one user says it’s not random

A hobbyist's 9-day test finds AMD's random number generator never produces a 0, across two processors. The company has escalated the case.

By mitch·4 min read
A close-up of a processor with binary patterns, representing a malfunctioning random number generator.

A hobbyist has reported a strange problem with AMD’s random number generator: his testing software has not produced a single zero in months, across two processors. The result is so improbable it reads like a joke, but the tester insists the results are real.

The tester, who describes himself as both an assembly language programmer and an electronics engineer, says he has hundreds of ways to build a high-quality random number generator using white noise from semiconductors. He does not believe the task is beyond the engineering teams at Intel and AMD.

The Tester’s Case

The problem centers on a simple question: can AMD’s random number generator produce a zero? The tester’s software, which he has updated to benchmark the generators, has run for 9 days straight without a single zero appearing.

Advertisement

The tester has not received a response from AMD yet, though he says the company has escalated his case internally. He is still waiting for a precise update, which he plans to share publicly.

Benchmarking the Generators

The tester’s new testing update also allows benchmarking the random number generators on processors that do not have the ‘rdseed’ instruction. That change has produced some surprising results.

The tester reports that different processors perform very differently. A Core i5 from 2012 generates over 12 million numbers per second, while a Core i7 7700 from 2017 manages fewer numbers per second. His AMD Ryzen 7 sits in the middle, at about 2.6 million numbers per second.

The patterns in clock behavior offer a clue. The Core i5 from 2012 appears 100% dependent on the clock, the tester says, while the AMD generator seems completely inert to clock variation. Those observations come from the tests he did.

The Numbers Problem

The absence of zeros is the headline, and it is the strangest part of the report. The tester has not found any information on how AMD’s RNG unit works, which leaves the question of why the hardware would behave this way.

There is no official statement from AMD yet. The tester has not received a response, and he continues to wait for a precise update.

What the Benchmark Says

The benchmark results show large differences between processors, even within the same manufacturer. The tester says this leads him to realize that the random number generator circuit must have large differences and probably completely different versions or revisions, even for models from the same manufacturer.

The tester’s observations on clock dependence point in a specific direction. If AMD’s generator is completely inert to clock variation, that would set it apart from the Core i5, which relies entirely on the clock.

A Waiting Game

The tester is still holding the two AMD processors. He has not received a response from AMD, but he says the company has escalated his case internally.

He will share the update from AMD here when it arrives, and he says he is still waiting for that precise update.

“Since 9 days before posting here, until the present moment, not a single 0 came out of the 2 AMDs that I had access to!”

The story is a strange one, and it has drawn attention online. The tester’s report has reached 177 points on Hacker News, a sign of the curiosity his numbers have sparked.

The absence of zeros is remarkable, but it is not necessarily a bug. The tester has run the test for 9 days, and the result is so improbable it reads as a joke.

But the tester is not laughing. He is waiting for a response from AMD, and he will share whatever the company tells him.

The story is a reminder that hardware is not magic. It is a circuit, and it can fail in strange ways. The fact that the tester has not found any information on how AMD’s RNG unit works adds to the mystery.

For now, the two AMD processors sit in his testing rig, churning out numbers that refuse to be zero.

  • The tester has not received a response from AMD
  • The company has escalated his case internally
  • The test has run for 9 days without a single zero
  • The Core i5 from 2012 generates over 12 million numbers per second
  • The Core i7 7700 from 2017 manages fewer numbers per second
  • The Ryzen 7 generates about 2.6 million numbers per second

The story is still developing. The tester will share whatever the company tells him, and the waiting continues.

Source material: “AMD's random number generator can't generate a 0?,” flatassembler.net.

The Notebook

Get the Notebook.

The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.

We send one note to confirm. Every issue has a one-click way out.

Advertisement

Leave a Reply

Your email address will not be published. Required fields are marked *

As an Amazon Associate, Clay Tribune earns from qualifying purchases.