The Mill CPU Architecture - the Belt (2 of 13)
Channel:Mill Computing, Inc. · Length:1:26:49 · Published:2013-07-29 · Views:37729
Summary
The second talk in the Mill CPU series by Mill Computing, covering the Belt, the Mill's operand-passing model that replaces general registers. It explains the cost of register renaming in conventional wide CPUs and how the Belt reaches high instruction-level parallelism.
Why HN recommends it
HN points to the Mill talks as an excellent, deep look at an unconventional CPU design, especially in threads on registers and instruction sets.
the Mill was described in an excellent series of talks — HN, 2023
Subtitles (excerpt)
Good evening. And thank you very much Google for providing the site and the invitation. This is one of a series of talks about the Mill general purpose CPU architecture. Um this one is about the internal data structures and how data is communicated within the machine from operation to operation and the programmer's model that works on that the programmer sees. We'll touch a little bit about how it's implemented on the inside as well. To do this, I need to touch lightly on the available choices for architectures which are out there. These are a couple of top-of-the-line or at least they were top-of-the-line when I went to go get the numbers. Top-of-the-line CPUs from major manufacturers. This one has got four cores that are on the chip, each putting out four operations per cycle peak, 3.3 GHz, burning 130 W, giving you 52 giga operations per second and costing $885 quantity of 1,000. Architecturally, this is a out-of-order superscalar architecture. You can probably guess which one, but what constitutes top-of-the-line changes week by week, so look a couple of months back. A different one has got a single core which can issue eight operations peak per cycle, half a GHz, only 1.1 W, and only 3.6 giga instructions per
On Hacker News
This is reminiscent of the Mill CPU which had temporal addressing and called the history a “belt”. the Mill was described in an excellent series of talks: [VIDEO] — wood_spirit, 2023-12
No ordinary registers. Instead you have a FIFO queue( belt ), where every element( register ) is accessible and has a relative address( offset ). [VIDEO] — w0000t, 2015-08
False. Mill does not do register renaming. ( Or use "lies". I failed to comprehend you second sentence. ) Here is a talk from one of the designers explaining how it is done: [VIDEO] — w0000t, 2015-08
How far has development come with the Mill CPU? This video[1] mentions that "nothing with pins on the bottom" has been made yet. [1] [VIDEO] — runeks, 2015-02
The studies would vary over time because CPU design and bottle necks have changed. Early designs were of course limited by transistor count, now we have OoOe and physical registers are limited by muxers and latency (see the presentations by the mill CPU guy [1] Saving registers in functions is mostly irrelevant - you only save what you'd use, so saving more means fewer spills within the function.… — brigade, 2013-10
Related videos
- 8-bit computer update · Ben Eater · 6:53
- Light Years Ahead | The 1969 Apollo Guidance Computer · TNMoC · 1:21:22
- “Hello, world” from scratch on a 6502 — Part 1 · Ben Eater · 27:24
- Finding the Missing Memristor - R. Stanley Williams · The Qualcomm Institute · 47:52
- The Ultimate Game Boy Talk (33c3) · media.ccc.de · 1:01:06
- The Thirty Million Line Problem · Molly Rocket · 1:48:55
- USENIX ATC '21/OSDI '21 Joint Keynote Address-It's Time for Operating Systems to Rediscover Hardware · USENIX · 1:06:19
- Jim Keller: Moore's Law, Microprocessors, and First Principles | Lex Fridman Podcast #70 · Lex Fridman · 1:34:43
- 34C3 - The Ultimate Apollo Guidance Computer Talk · media.ccc.de · 1:01:41
- Jim Keller: Moore’s Law is Not Dead · UC Berkeley EECS · 1:04:53
- "Programming a 144-computer chip to minimize power" - Chuck Moore (2013) · Strange Loop Conference · 40:12
- NaN Gates and Flip FLOPS · suckerpinch · 19:07
- Richard Feynman Computer Science Lecture - Hardware, Software and Heuristics · Muon Ray · 1:15:47
- Breaking the x86 Instruction Set · Black Hat · 44:29
- Guarding Against Physical Attacks: The Xbox One Story — Tony Chen, Microsoft · Platform Security Summit · 58:05
- Simple Code, High Performance · Molly Rocket · 2:50:14