nteon

10 ms·

61.
by nteon 12y ago
the glibc allocator is based off of Doug Lea's malloc that was developed before multithreading was common. Optimizations for multiple threads + arenas were retrofitted on top of the existing code, whereas jemalloc was designed more re
62.
by nteon 12y ago
I would have loved to see some data that showed their bump pointer regions and thread local stuff performing significantly better than tcmalloc or jemalloc, both of which do thread-local caching that avoids locks for the vast majority of al
63.
by nteon 12y ago
I do see it now... wasn't testing right before. Goes away as you say when disabling transparent THP. KSM has no effect, contrary to what I thought earlier.
64.
by nteon 12y ago
Netflix is a Java-heavy shop (the client talked up in the blog post is the Java client). I was pleasantly surprised to see that dynomite is implemented in C with an eye towards both BSD + Linux. I'm curious if this was because a GC&#
65.
by nteon 12y ago
as a followup, I don't see large latency with huge pages, but I have kernel samepage merging disabled. I would be interested if this also decreases latency for you: echo 2 > /sys/kernel/mm/ksm/run Also check
66.
by nteon 12y ago
It has already been fixed with thegetrandom(2) syscall: https://lwn.net/Articles/605828/ EDIT: this is in 3.17, so it will be a while before it is commonly available.
67.
by nteon 12y ago
I just tested - while running the redis-benchmark command you listed after a BGSAVE smaps lists all anonymous memory as AnonHugePages. So it ends up as THP, although I suppose its possible that it is broken into 4k pages and then adjacent
68.
by nteon 12y ago
After the latency spike does the smaps output still list the same amount of AnonHugePages?
69.
by nteon 12y ago
Are you working on this code publicly? I would love to take a look, as I've become quite interested in the gc compiler lately.
70.
by nteon 12y ago
The _DEFAULT_SOURCE change is annoying. Well maintained C codebases that use -Werror combined with _BSD_SOURCE are now broken until they add the additional _DEFAULT_SOURCE define to their build, like Go: https://code.google.com&
71.
by nteon 12y ago
I'm not sure I see the benefit of plugin discovery or separate namespaces. Loading a plugin gives you a Plugin struct, which is a namespace to lookup symbols (functions + globals) from that plugin. Most plugins would I think by necess
72.
by nteon 12y ago
On linux, system calls don't result in a TLB flush - kernel data structures and code are in a different portion of the virtual address space (starting from the top of VM memory, if I remember right) that is tagged as not being availabl
73.
by nteon 12y ago
Its actually pretty impressive that 88% of tor operators have deployed OpenSSL updates. Curious how that compares to the web.
74.
by nteon 13y ago
I do this too, and agree. The benefits and straightforward version control diffs are fantastic when working on a distributed team with multiple devs touching the UI.
75.
by nteon 13y ago
haha, yup. I just straight up read the code wrong :)
76.
by nteon 13y ago
I don't think so. I don't see any direct calls to recvfrom_into in Django, and it seems little used. It is more likely to be called with a UDP listener - with TCP sockets there isn't usually a need to use recvfrom (which re
77.
by nteon 13y ago
nicely, it seems that the python3 version of socketmodule doesn't use these on-stack Py_buffer objects, so wasn't affected.
78.
by nteon 13y ago
I would be really surprised if the rest of the function was still there - this seems like the perfect example for a dead code elimination compiler pass.
79.
by nteon 13y ago
>> I'm no kernel hacker, but doesn't that make the GP argument better? Nope. Dynamic linking is done each time the program is loaded - the kernel calls out to the dynamic linker to open shared libraries, resolve symbols, cr
80.
by nteon 13y ago
I use https://github.com/bpowers/psm every day at this point. Simple, fast and filterable memory reporting. disclaimer: I also wrote it.
81.
by nteon 13y ago
they're not talking about the plan 9 OS. Go ships with forked copy of several plan9 libs (lib9, libbio, libmach) & the plan9 c compiler. The plan is to obsolete the use of all of that code.
82.
by nteon 13y ago
most likely on case-insensitive file systems, like the default HFS+ on the Mac and NTFS on Windows. EDIT: to expand, reference lookups are filysystem lookups: 01:42:24 (gh-pages) [bpowers@fina myproject]$ strace git show HEAD 2>&am
83.
by nteon 13y ago
The original behavior, where a substring pins the full original string, is also how Go's substring slicing works. I much prefer this behavior, as it lets you profile and add a copy yourself if you find yourself pinning large strings.
84.
by nteon 13y ago
The strangeness I see is that the /delegate call isn't specific. I can't say that I want Joe to be able to decrypt LaunchCode3, so I could end up inadvertently allowing Mary to decrypt SecretLocation without really wanting t
85.
by nteon 13y ago
I don't think there is something inheritantly wrong with a lib allocating on your behalf. Most non-trivial problems require memory allocation, and if you have a struct that has dynamically allocated fields (like char*'s of data d
86.
by nteon 13y ago
What he describes is how the Linux kernel implements linked lists: http://kernelnewbies.org/FAQ/LinkedLists
87.
by nteon 13y ago
I can't stand macro comments - it makes it impossible to figure out whats going on without being sure to open the file in a specific editor with a specific project loaded. Say you have a separate lib project used by an application - if you
88.
by nteon 13y ago
I've written a small production web project and several medium-sized personal projects in Go, and have not yet found a need for a web framework outside of net/http. I've used several libraries that work as nested/stacked handlers under net
89.
by nteon 13y ago
You're forgetting the largest smartphone OS provider. And as Android is embedded into devices like televisions and cars, having a great libre codec available could be a significant advantage.
90.
by nteon 14y ago
excellent suggestion, I will definitely try something like that soon. There are in fact only 2 variable sized lines - the first VMA info line and the last VmFlags line. In the fast path the middle hunk of map info can be accessed as a sing

More ›