Bruce Sutherland

33 posts

Bruce Sutherland

Bruce Sutherland

@bonus_sco

Senior Software Engineer - Epic Games. Ex-Activision, EA, 4J Studios.

Katılım Aralık 2023
115 Takip Edilen6 Takipçiler
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@Jonathan_Blow Nvidia have had ARM CPUs for a long time. Denver was over a decade ago. I'll be amazed if it's anything other than ARM.
English
1
0
0
306
Jonathan Blow
Jonathan Blow@Jonathan_Blow·
I doubt that's what it is, because the modern computer industry does not seem to be able to do anything that is not incremental and accretive. But one can hope.
English
14
4
194
12.2K
Rilwan
Rilwan@Real1_balogun·
My most legendary response from a manager to a journalist was from David Moyes. It was in 2022. West Ham lost 1-0 to Everton in that game and this journalist called Moyes “Moyesy”. He laughed and said: “Moyesy?! F*******g hell, I don’t think we’re that close”😂😂😂😂😂 Lives rent free in my head. Let me see if I can find the video.
English
14
28
1.2K
477.9K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@Jonathan_Blow @rfleury One of the most interesting bugs I encountered was on the PS3 port of Oblivion. QA reported that standing still for 24 hours then turning 360 degrees crashed the game. It turned out to be memory fragmentation caused by dynamic animals spawning in the world.
English
1
0
6
117
Ryan Fleury
Ryan Fleury@rfleury·
"The stack" is a per-thread address space range, dynamically reserved by a kernel when a thread is created. The reason why "stack" is often presented as preferable to "heap" is that, when using a thread's stack, the expensive part of allocation - address space reservation, and preparation of physical pages for backing the address space - has already been performed when the thread was created. But kernels also provide mechanisms for doing your own address space reservation (mmap, VirtualAlloc), and there is nothing stopping you from using these to do bulk allocations up-front to create your own stacks. This can make common case allocations as cheap as "the stack", but the advantage is that you now control the semantics and lifetime of the stack you've created. Thus, it does not need to be coupled to - for example - the lifetime of a scope or function, as the thread stack is. The "stack versus heap" dichotomy is an unfortunate mythology because it seems to, in practice, communicate the idea that when a thread stack is insufficient for some purpose (allocations must exceed scope boundaries, allocations may need to exceed thread stack limits, allocations require more fine-tuned reserve/commit behavior, and so on), then the only alternative is the heap, particularly for very granular allocations. This is, again, a mythology, and it has confused the C++ world in particular for decades.
Boost C++ | Open Source Libraries@Boost_Libraries

std::vector always heap allocates. std::array can't change size. For decades, there's been no standard container that gives you a dynamically sized array with a compile-time capacity limit and zero heap allocation C++26 finally adds std::inplace_vector. Guess where they got the idea 🧵👇

English
33
60
1.1K
72.7K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@Jonathan_Blow @rfleury Visceral tech had different heaps of fixed size for different game systems, think online heap, audio heap, etc. Call of Duty had two bump allocators. One growing up, one growing down with the left over space in the middle used for streaming.
English
0
0
2
128
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@GergelyOrosz Anthropic: no, you can't use Claude. xAI: ok, run it from my data centre then. I won't peek at it, pinky promise. Or maybe they just agreed a deal to have usage of the inference running in that data centre and Anthropic gets more compute. It's probably that.
English
0
0
1
3.1K
Gergely Orosz
Gergely Orosz@GergelyOrosz·
So let me get this right: 1. Anthropic bans xAI from using Claude (to stop them from perhaps distilling Claude for their own model) (...) 2. xAI gives up ~a quarter of its DC capacity for Anthropic to rent and run Claude A win for Anthropic no doubt. What's in it for xAI tho?
English
386
18
1.6K
320K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@SebAaltonen @MegaPunkGames I'd love some insight into your context management strategies because these experiences align with mine in long context sessions. Rarely see things like this in low context sessions.
English
0
0
0
14
Sebastian Aaltonen
Sebastian Aaltonen@SebAaltonen·
@MegaPunkGames It's still 2x-3x faster throughput than me. I don't have a rendering team currently. AI is my team. It does a pretty good job generally, but sometimes makes these crazy mistakes.
English
1
0
11
424
Sebastian Aaltonen
Sebastian Aaltonen@SebAaltonen·
Me: Why did you remove water features such as waves? What was the reasoning? Codex: There was no good rendering reason. It was a bad port.
English
8
0
68
12.8K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@ssamat On device Gemma in the OS exposed through an API so apps can use it without each one needing to download their own? Go on, you know you want to.
English
0
0
1
2K
Sameer Samat
Sameer Samat@ssamat·
Biggest. Android. Updates. Ever. The Android Show: I/O Edition returns May 12 @ 10am PT. 🎬You won't want to miss it: android.com/io-2026
English
213
359
4.5K
1.2M
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@cmuratori @tomwarren Can Windows actually allow something to show on screen faster than the run dialog here? That would be an "interesting" (and likely soul destroying) investigation. The run dialog could latency here could be as good as it gets in modern Windows 😬
English
0
0
2
221
Casey Muratori
Casey Muratori@cmuratori·
@tomwarren And as you can clearly see from my post, I'm not suggesting that the new one is slower than the old one. I'm saying that if you supposedly rewrote something to prioritize performance, you should be getting a lot better than a 1/10th of a second response time.
English
2
0
82
2.7K
Tom Warren
Tom Warren@tomwarren·
if we're misleadingly conflating 2 different perf metrics (time to first frame latency and fps smoothness) then just factor in OS boot time so it's less than 1fps 🙃The reality is the new Run loads faster (94ms) than the existing one (103ms) that nobody has ever moaned about
Casey Muratori@cmuratori

Just want to make sure I'm reading this right: Microsoft rewrote the run dialog with performance "top-of-mind", and the best they could manage to do when putting up a single text box was 10fps?

English
31
15
422
83.2K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@HarshithLucky3 @xm_build I'd love them to embed the model in the OS and expose an API. Multiple apps all downloading their own models will not be a good place.
English
0
0
0
23
Harshith
Harshith@HarshithLucky3·
Google is testing an app called COSMO on play store its a huge 1.13gb download because it includes local Gemini Nano asks for big system permissions like screen access and voice match lets you choose between local offline processing or hybrid cloud has a browser agent called mariner that automates web tasks for you can do deep research write documents and auto schedule calendar events, etc.. has a recall feature and can pull up photos mid conversation probably an accidental release to test features for Google I/O 2026
Harshith tweet mediaHarshith tweet mediaHarshith tweet mediaHarshith tweet media
English
10
24
477
37.4K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@mattpocockuk @codevibesmatter Scripting is cheap, architecture is hard. We'll move to a software world where scripts created at runtime by LLMs is commonplace. UI scripts generated on the fly by an LLM. The engine will still be the hard part.
English
0
0
0
54
Matt Pocock
Matt Pocock@mattpocockuk·
@codevibesmatter Yeah what it sounds like you're describing is what I advocate for - thinking about the design of the system every day. One issue here is that designing the system for longevity is something that AI running AFK still struggles with.
English
5
2
18
1.1K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@staylorish This single policy has put me off voting for them. Won't make the slightest difference in Dundee but this is just bonkers.
English
0
0
0
167
The Herald
The Herald@heraldscotland·
Does Scotland need more Malcolm Offords? 👇
The Herald tweet media
English
549
67
821
118.2K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@criplefish @SebAaltonen You can still pull them off pretty easily. Just that tiny bit of friction makes an improvement. Complete non-issue. If these were the norm before no one would be running campaigns to remove them 🤷‍♂️
English
1
0
2
32
Sebastian Aaltonen
Sebastian Aaltonen@SebAaltonen·
Most Europeans actually like the new bottle caps. It's convenient that you don't need to hold the cap in your other hand. And you can't lose it. There were some first-gen bottles with badly designed caps, but most manufacturers have fixed their initial design flaws now.
Rodney@_ChanFace

Why were Europeans so whiney about the attached bottle caps?? Regardless of environmental reasons, it does not interfere with drinking at all and it’s actually super convenient to not need to hold on the cap! 10/10 Canada should adopt!

English
137
12
494
46.6K
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@OlafSNP @JohnSwinney Utterly bizarre incentivisation through legislation here. Why is the current market competition not enough to ensure prices are as low as they can be?
English
0
0
0
2
Bruce Sutherland
Bruce Sutherland@bonus_sco·
@OlafSNP @JohnSwinney Why wouldn't they compete on price with each other just on those lines? What incentive would they have to not sell every other line above the minimum? What happens to every other business that can't compete in that space? You entrench the big supermarkets even further.
English
1
0
0
4
Olaf Stando
Olaf Stando@OlafSNP·
Why on earth is Anas Sarwar so desperate to stop @JohnSwinney from cutting food prices? Weak, weak, weak.
English
107
165
404
12K