mgaunard

16 ms·

271.
by mgaunard 10mo ago
I never understood why people claim the Macbook is so good. Bad keyboard, bad aluminium body, soldered ram... Is it just the Apple Silicon that somehow makes it worth it? It's ARM, most software is still written and optimized for x86.
272.
by mgaunard 10mo ago
That's Milan to Amsterdam though. Aren't there some mechanisms to have a handoff as close to Milan as possible? I was pretty sure there was a way to have something that's just a fixed cost, maybe with a partner third party se
273.
by mgaunard 10mo ago
Just waiting for a MTU-sized frame to come in through the network at 10Gbps is 1.2us. Reacting to incomplete frames in software is possible, but realistically at this point just use FPGAs already.
274.
by mgaunard 10mo ago
Some comments from skimming through the code: - spin loop engine, could properly reset work available before calling the work function, and avoid yielding if new work was added in-between. I don't see how you avoid reentrancy issues as
275.
by mgaunard 10mo ago
Those numbers seem to be TSC sampled in software from the moment it receives a full frame to the moment it starts sending a packet. The traditional way to measure performance in HFT is hardware timestamps on the wire, start of frame in to s
276.
by mgaunard 10mo ago
There are times where a big company needs to build something new (albeit within a constrained ecosystem and a very narrow swimming lane). To do so, one good way is to hire the experts of that domain that have built it before. That can mean
277.
by mgaunard 10mo ago
Big Tech are large organizations with different needs than that of small organizations. They care about setting standards over widely different teams and managing large-scale upgrades etc. It's not optimized for velocity or cost effici
278.
by mgaunard 10mo ago
It is quite common for CTOs of smaller companies to be hired as team leaders into bigger companies; nothing wrong with that.
279.
by mgaunard 10mo ago
Personally, I write my own build systems. Any readily available build system is more of a meta-language onto which you code your own logic, but with limited control and capabilities. Might as well take control of the whole stack in a real
280.
by mgaunard 10mo ago
To my knowledge direct connect is a fixed cost, as is running your own dedicated infrastructure, so you don't pay for bandwidth.
281.
by mgaunard 10mo ago
The X1 Carbon is fine, not sure what is wrong with the 2k IPS display. Do you need 4k for a 13 inch laptop?
282.
by mgaunard 10mo ago
Well of course, if you were CTO of a company of 10, you can't expect to be hired as CTO of Google.
283.
by mgaunard 10mo ago
13 interviews suggests he was interviewing for multiple roles within the same company; in which case it's not that shocking. In many places every team runs their own interviews.
284.
by mgaunard 10mo ago
It might be nicer to go work for startups, acquire experience there as you build everything from scratch across the whole stack, then get hired at a high responsibility position. Though most people into entrepreneurship never go back to big
285.
by mgaunard 10mo ago
So it's like C++ consteval?
286.
by mgaunard 10mo ago
My understanding is that S3 egress is only a problem if you need to take data out of AWS, which you can simply avoid by having some kind of dedicated AWS direct connect or some such to route the traffic yourself?
287.
by mgaunard 10mo ago
The compiler is only able to perform certain optimizations that have no observable behaviour. For example it can only parallelize code which is inherently parallelizable to begin with, and unless you design your algorithm with that in mind,
288.
by mgaunard 10mo ago
My experience is that anything involving Bazel is slow, bloated, and complicated, hammers your disk, copies your files ten times over, and balloons your disk usage without ever collecting the garbage. A lot of essential features are missing
289.
by mgaunard 10mo ago
My question is in the context of doing those optimizations yourself, understanding what can be done to make the code more efficient and how to code it up, not the compiler engineering to make that happen.
290.
by mgaunard 10mo ago
I guess what's surprising here is that compilers are able to perform those optimizations systematically on arbitrary code, not the optimizations themselves, which should be obvious to a human.
291.
by mgaunard 10mo ago
My interview aims to assess whether the candidate understands that the dependency of each iteration on the previous one prevents effective utilization of a superscalar processor, knows the ways to overcome that, and whether the compiler is
292.
by mgaunard 10mo ago
I spoke in the context of coding those optimizations yourself.
293.
by mgaunard 10mo ago
My point is that it's not a personal amount, but an absolute constant. If you need to mix other things with it, then the coffee isn't good.
294.
by mgaunard 10mo ago
The comparison to coffee is bad, since it's obvious the best coffee is with no sugar and no milk.
295.
by mgaunard 10mo ago
RTP is specifically designed for real-time video.
296.
by mgaunard 10mo ago
Actually, with newer Linux kernels and io_uring, it appears it is now possible to do multiple writes asynchronously concurrently, by annotating each one with a sequencing constraint. Whether that's really useful or not depends on wheth
297.
by mgaunard 10mo ago
TCP is not message-oriented. Retransmitted bytes can be at arbitrary offsets and do not need to align with the way the original transmission was fragmented or even an earlier retransmission.
298.
by mgaunard 10mo ago
There are a number of alternatives to VDHL and Verilog, many of which lower to Verilog, for example MyHDL.
299.
by mgaunard 10mo ago
It's just a standard TLS layer, works with any TCP protocol, nothing WebSocket-specific in it. You should ideally design your messages to fit within a single Ethernet packet, so 2 bytes is more than enough for the size. Though I have s
300.
by mgaunard 10mo ago
Doing more system calls isn't really a good idea for performance. Also if you're doing asynchronous writes you typically can only have one write in-flight at any time, you should aggregate all other buffers while that happens. Tho

More ›