hmpc

7 ms·

1.
by hmpc 5mo ago
Similarly, if not performance-focused, I can wholeheartedly recommend Building Git[0], which walks you through building your own git clone in Ruby (although the language is immaterial). [0]: https://shop.jcoglan.com/building
2.
by hmpc 5mo ago
It's definitely possible to do correctly, but looking through the code for both crates it doesn't look like they take the necessary precautions (issuing a fence or using RDTSCP). Which is a little weird because at least quanta exp
3.
by hmpc 5mo ago
Thanks for sharing this! I hadn't seen it before, I'll definitely add a mention in the post. This approach is a smart way to get the same precision as the kernel without tying it to the vDSO implementation as in the post, so presu
4.
by hmpc 5mo ago
This is explicitly called out in the post as well as the Intel instruction manual. Every codebase I've ever seen that reads the TSC either issues an LFENCE or uses RDTSCP. In my benchmarks RDTSCP has a slight advantage, despite the slo
5.
by hmpc 5mo ago
You might've misunderstood the requirements. The time scale was 1-10 micros per component; 100 ns was the overhead per span we were aiming for. In this case distributed tracing absolutely was the right choice. These were not simple com
6.
by hmpc 5mo ago
Yep, I should probably have mentioned this option for completeness. In practice, however, it only saves you a few cycles/nanos and adds more complexity and failure points downstream.
7.
The fastest Linux timestamps(hmpcabral.com)
71 points by hmpc 5mo ago|18 comments
8.
by hmpc 6mo ago
Check the specification at the top. The range for x is [-1, 1]. For the range you provided the accuracy of the 0.5x alternative is reported as only 33%: https://herbie.uwplse.org/demo/570b973df0f1f4a78fe791858038a...
9.
The machines are fine. I'm worried about us(ergosphere.blog)
21 points by hmpc 6mo ago|3 comments
10.
Vtables Aren't Slow (Usually)(louis.co.nz)
1 points by hmpc 6mo ago|0 comments
11.
by hmpc 6mo ago
It depends on your use case, as always. Correctness is not always black and white (hence my favourite compilation flag, -funsafe-math-optimizations) and time complexity can be misleading (O(log N) with a large base is O(1) in practice), but
12.
by hmpc 6mo ago
There's an interesting comment on this over on Lobsters: https://lobste.rs/s/e4y5ps/two_studies_compiler_optimisation...
13.
by hmpc 6mo ago
(Author here) >It all seems very brittle, though. And that something has gone very wrong with our ecosystem of tools, languages, and processes when it becomes advisable to massage source until specific passes in a specific version LLVM d
14.
Two studies in compiler optimisations(hmpcabral.com)
128 points by hmpc 6mo ago|21 comments
15.
The C++ input iterator pitfall(hmpcabral.com)
3 points by hmpc 7mo ago|0 comments
16.
Making Waffles with Group Theory(hmpcabral.com)
1 points by hmpc 7mo ago|0 comments
17.
by hmpc 6y ago
Don't you need your read to be line-oriented to avoid breaking up lines?
18.
by hmpc 8y ago
Full disclosure: I work for Veniam on connected vehicle technology, including LTE, Wi-Fi, and DSRC stacks. Unfortunately there's also still no agreement on what 5G will actually be, what technologies it will entail, or how it will be d
19.
by hmpc 10y ago
Backticks are discouraged for any new scripts, since they complicate nesting, quoting, and escaping, besides being graphically less explicit. See also http://mywiki.wooledge.org/BashFAQ/082 . The Open Group also discour
20.
by hmpc 11y ago
Actually no. From the manpage: Use "-C -" to tell curl to automatically find out where/how to resume the transfer. It then uses the given output/input files to figure that out.
21.
by hmpc 11y ago
Veniam | Mountain View, CA/Porto, Portugal/Singapore | Multiple openings | ONSITE We are looking to fill several openings across Engineering, Sales, and Management: https://veniam.com/jobs/ . We have opportuni
22.
by hmpc 11y ago
You can also remove all the "STCP | PortoDigital" networks in Porto, Portugal, which are WiFi hotspots in the public buses. Interesting map! (Disclaimer: I work for the company who implemented that network.)
23.
by hmpc 11y ago
This is the service supervision approach, which is indeed the best, avoiding race conditions and nasty polling. If you really must monitor processes which you cannot parent, at least under Linux you can do this using ptrace to sort-of-repar
24.
by hmpc 12y ago
Veniam, Portugal - Software Engineer - http://www.veniam.com/ Veniam just raised a $4.9M series A to build the networking fabric for the Internet of Moving Things ( http://techcrunch.com/2014/12/02&
25.
by hmpc 12y ago
Unless I misunderstood, that's what the TIMER_ABSTIME option is for. This is actually what I use for soft real-time loops, when I'm concerned about both drift and jitter.
26.
by hmpc 12y ago
You can already do this in POSIX.1-2001-conforming systems like Linux using `clock_nanosleep` ( http://man7.org/linux/man-pages/man2/clock_nanosleep.2.html ).
27.
by hmpc 12y ago
Veniam, Portugal - Software Engineer - http://www.veniam.com/ Veniam just raised a $4.9M series A to build the networking fabric for the Internet of Moving Things ( http://techcrunch.com/2014/12/02&
28.
by hmpc 12y ago
𝐏𝐨𝐫𝐭𝐨, 𝐏𝐨𝐫𝐭𝐮𝐠𝐚𝐥 - 𝐒𝐨𝐟𝐭𝐰𝐚𝐫𝐞 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 - 𝐕𝐞𝐧𝐢𝐚𝐦 http://www.veniam.com/ We are looking for a Full-Stack Software Engineer to help us build the networking fabric for the Internet of Moving Things. Someone with entrepreneurial sp
29.
by hmpc 12y ago
Are you in the US? Because in Europe all electrical products are certified under strict EU directives, and (at least in my country) there are routine inspections to ensure that all products for sale comply. You can walk into just about any
30.
by hmpc 12y ago
Message body just as in <subject> use imperative, present tense: “change” not “changed” nor “changes” Is it me or is this directly contradicted by the examples at the end? Which, by the way, seem much more natural than they

More ›