Graham Markall

Professional interests: Python, CUDA, @numba_jit.
Personal interests: RISC-V, PSXDev, OSHW, 日本語.
Fun: family, cycling, running.
Also @[email protected]

[Graham Markall @gmarkall]

Apparently it's #MyTwitterAnniversary, so I'll use it as an opportunity to link to the Mastodon account I now use: https://t.co/8Qcrn8yLDs

[Graham Markall @gmarkall]

Quick reminder that I'm mainly t[oo|wee]ting on Mastodon these days: https://t.co/8Qcrn8yLDs

[Graham Markall @gmarkall]

Me on Mastodon: https://t.co/8Qcrn8yLDs

I don't plan to disable this account, but I'm generally posting there these days.

[Graham Markall @gmarkall]

Do you know about automatic differentiation and / or Enzyme? Could you help sketch out how AD support for @numba_jit could be implemented? Issue / thread: https://t.co/jISDyxtEfW

[Graham Markall @gmarkall]

I've been on Mastodon a while but I got out of the habit of using it - this is my account that I'm starting to get active with again: https://t.co/8Qcrn8yLDs (is this usually written as @[email protected]?). It has an old profile pic, still need to upload my current one.

[Graham Markall @gmarkall]

@__mharrison__ Faster Python with @numba_jit

[Graham Markall @gmarkall]

@9600 Uploading to pachube from my na-no-day

[Graham Markall @gmarkall]

@jonathandotcel @anthonypjshaw Yeah, the CPython Internals book is great for this, and I think there's a few blog posts on https://t.co/WJ7BjkyMra that cover some of the same / similar material too

[Graham Markall @gmarkall]

I've been experimenting with using CUDA Unified Memory for the Python heap, towards a general goal of experimenting with a more unified CPU / GPU execution model within CPython, implemented as PyMemAllocatorEx instances. I have a few thoughts so far...

[Graham Markall @gmarkall]

@apmasell If you optimise for size (-Os) I would imagine you have the best chance of that happening. I'm sure I've been aware of examples of that sort of thing but need to do a bit of digging.

[Graham Markall @gmarkall]

@apmasell Maybe answering my own question, but it seems that constants in a constant pool have their own label in the asm source, so perhaps not part of the function. e.g. for 0x5f3759df in https://t.co/TtUGvTIPiI

[Graham Markall @gmarkall]

@apmasell Years ago I recall trying to infer the size of functions, and I was working with elf files, working it out by looking for the following symbol. I maybe was an idiot and unaware of the size entry! How does it consider constant pools located near functions? Are they part of it?

[Graham Markall @gmarkall]

@apmasell Does ELF have a way to do that if you haven't used something like -ffunction-sections?

[Graham Markall @gmarkall]

@jonathandotcel @anthonypjshaw tutorial in the official docs suggests (https://t.co/D7lbOmDuJb)... Though things might have been different if I'd been working in a different domain - I've not really used Python to implement a server app or anything like that, mainly compiler toolchain support / compiler work

[Graham Markall @gmarkall]

@jonathandotcel @anthonypjshaw From a language user standpoint, I don't feel it's been that important to understand the internals of CPython - before I started working on Numba I think I was pretty happy just knowing the basics of the GIL (basically only one thread executing Python code) and what the...

[Graham Markall @gmarkall]

@vellamike @anthonypjshaw Whoops, totally forgot to link to https://t.co/rg9Vsykb9U earlier - basically it's memory that can be accessed from CPU and GPU and migrated transparently between devices and host as required

[Graham Markall @gmarkall]

@anthonypjshaw 6. Finally, maybe this should have been a blog post with some code! Would that be interesting? What else should be answered / elaborated on if I write this up?

[Graham Markall @gmarkall]

5. "CPython Internals" by @anthonypjshaw is a great intro / reference to CPython, saved me a lot of time, and got me up to speed quickly for this endeavour. Highly recommended if you want to poke about with this sort of stuff! (https://t.co/VmTiicnfYY)