dotstdy

dotstdy, 2 hours ago to random

I can't believe the whole Linux CVE mess. I had Linux CVEs hooked up to my keyboard heater, but now they're generating so many useless CVEs my keyboard is far too hot for comfort! I'm angrily waiting for the kernel team to take this problem seriously.

dotstdy, 7 hours ago to random

It's kind of funny when people complain about things which are the direct result of something being useful for a long time. Like yeah it's complex and has a bunch of weird extensions, it's been solving (hopefully real) problems for 20 years. And guess what, if you throw all that away to build something "clean" you're still going to end up with a complete clusterfuck in N years (where N can be as low as 0, depending on how good you are at designing complex systems)

dotstdy, 1 day ago to random

subgroup ops are a mindfuck on top of the existing mindfuck of work groups, local work groups, dispatches and invocations.

dotstdy, 1 day ago to random

There's a bit of stuff in this article which phrases it in terms of changes over time, e.g. compute capability has grown and we no longer need big data. But it seems closer to reality that it was never required, and continues to not be required. (looking forward to the same style of post happening in a few years vis-a-vis microservices)

https://mastodon.social/@ltratt/112518285832831004

dotstdy, 2 days ago to random

i keep getting results for the church of the latter day saints when i look for LDS smh

dotstdy, 2 days ago to random

Anyway I'm curious whether I can move the coarse bitmap to LDS, merging the two culling passes. Basically dispatch a "coarse" tile, which loops over all the incoming primitives, writing the bitmap to LDS (16KiB per tile, too much?), then switch to one invocation per "fine" tile, looping over the data in LDS and writing out the final culled tiles.

dotstdy, 2 days ago to random

I wonder why the performance counters don't work in RGP captures on my APU. I guess something isn't plumbed up yet for RDNA3 :'(

dotstdy, 3 days ago to random

Has anyone ever made a good system for gameplay queries? (of the "give me all the apples nearby" or "trigger when x happens" style) I'm somewhat of the opinion that any generalized approach here is going to end up in tears, but I'm also of the opinion that gameplay scripting in particular wants to have a pretty generalized approach to help with auditing content (e.g. you want to tightly control what data scripts can access, and when they can access it).

dotstdy, 4 days ago to random

I find it funny the way people give away their own lack of understanding sometimes. Reading this comments on this (why i do this) https://bikepacking.com/plog/man-or-bear-debate/ and there's the classic "[RE: meeting women on the trail] Never in my silly mind crossed the thought of me making them uncomfortable."

If it has never crossed your mind that you might be making somebody uncomfortable - that's a you problem - and specifically one of the core issues the whole damn piece is talking about!

dotstdy, 4 days ago to random

Anyway here's the two level binning code. https://github.com/jsimmons/narcissus/commit/a2f73579cd48f1fec26503a170f02e50719ee7d2 And as usual it has been written following the "copy pasta approaches from themaister" methodology. https://themaister.net/blog/2019/10/12/emulating-a-fake-retro-gpu-in-vulkan-compute/

dotstdy, 5 days ago to random

To be honest I'm surprised the "one small rock a day" diet hasn't caught on outside of Sweden. Arguably it's one of the biggest reasons swedes are all so healthy and attractive. The rest of the world could learn a lot from Europe.

dotstdy, 6 days ago to random

It's always fun when you remove a parallel foreach and do some basic re-organization and end up with a result that's 25% faster on a single core than the old version was spread across 10 cores.

dotstdy, 8 days ago to random

Err is the GPU supposed to fault if you try to bind a descriptor set to a bind point before you bind a pipeline? I couldn't see any mention of there being an ordering requirement on the spec, but maybe I missed it in the sea of VUIDs

vkCmdBindPipeline()  
vkCmdBindDescriptorSets()  
vkCmdDispatch() // totally fine  

vkCmdBindDescriptorSets()  
vkCmdBindPipeline()  
vkCmdDispatch() // totally explosion  

Well, perhaps I should also mention that there's a graphics happening before this too.

dotstdy, 9 days ago to random

CPU optimisation guide: You should try vectorizing
GPU optimisation guide: You should try scalarizing

CHOOSE A LANE

dotstdy, 9 days ago to random

I feel like the most difficult part of subgroups and GPU programming in general, is getting all the terminology straight in your head. Sometimes it seems like it would be easier just writing rdna asm directly. :')

dotstdy, 10 days ago to random

I bet a lot of people think of themselves as important game development influencers, but the simple truth is that in the last decade or so the single most impactful game development thought leader is the "can you pet the dog" twitter account

dotstdy, 10 days ago to random

Masahiro Sakurai's youtube channel continues to be very wholesome https://www.youtube.com/

dotstdy, 11 days ago to random

One thing that I think really needs to happen is that datacenters are treated more like "Industry" and are forced to actually report their resource consumption publicly. The bullshit games of obfuscation from google, microsoft, facebook, et al, need to end.

dotstdy, 11 days ago to random

Unironically wrote linked list cycle detection today with the tortoise and hare trick. Made it out of the interview brain teaser and into real life.

dotstdy, 12 days ago to random

The Firefox mobile app is really not good :(

dotstdy, 12 days ago to random

Heck yeah, now I can use a bpf program to disable preemption in my spinlock bodies! Checkmate spinlock haters. (It doesn't actually allow this) https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=55d30cc90fd4

dotstdy, 12 days ago to random

So now we have replaced the fast fixed-function UI text rendering with an extremely slow (and probably slightly incorrect) compute version. Success? I guess so.

https://github.com/jsimmons/narcissus/commit/3851fb6d2cf8c5d56ecb029a12476dd1d220c10b#diff-84a451052161b35dc8c74ea47c846ed82d81a4951010407196e1b36997840f52

dotstdy, 13 days ago to random

the phrase "blazingly fast" really upsets me

dotstdy, 13 days ago to random

Bug of the week (nay, year) it turns out was caused by a mix of bad data setup in a particular area, combined with hashmap[key] inserting an item if no such item existed. I think this particular behavior is a pretty common source of quite painful bugs, and shoutout the colleague who finally debugged it.

dotstdy, 13 days ago to random

"the good AI can do for t̶h̶e̶ ̶w̶o̶r̶l̶d̶ my bank account will outweigh its environmental impact" - microsoft