valarauca1

13 ms·

61.
by valarauca1 8y ago
you have a lot of people in the channel Which is common you have a loud background Which is also extremely common. Fans, open windows, keyboard, neighbors yelling, etc. you don't want people hearing stuff not
62.
by valarauca1 8y ago
AVX-512 does not do to micro-op splitting (on intel server class CPU's, it does on some intel consumer cpu's). Amusing Zen is emulating AVX512 and AVX2 via micro-op splitting and it performs better under some workloads. The real i
63.
by valarauca1 8y ago
This post cites sources that actively contradict the points it attempts to make. It provides a wealth of useful aggregation of information against the points it makes. Let's us begin However, there are also deterministic frequenc
64.
by valarauca1 8y ago
A malloc implementation is free to divvy up a page in order to service several allocations But like the hardware won't enforce that. So yeah you can malloc 50bytes, but the hardware is just going to assume there is a 4096
65.
by valarauca1 8y ago
`malloc` is part of the standard library. the underlying implementation is environmental specific, or implementation specific depending on the host application. Generally speaking `mmap` is the system call that * Allocates memory * Moves al
66.
by valarauca1 8y ago
my irritation is C's primitive type system. And programmers willingness to conflate their standard library with system level operations.
67.
by valarauca1 8y ago
These numbers are based on values from Google's production stack. Insofar as these numbers are based on Jeff Dean's post: http://highscalability.com/blog/2011/1/26/google-pro-tip-use... That be
68.
by valarauca1 8y ago
Gradient Decent aims to please the average consumer, which often is not what is in-demand in the market place. Had Henry Ford used ML he would've invented a faster horse.
69.
by valarauca1 8y ago
you mean `mmap`? Also yes I'm radical enough to suggest the 50+ year old C standard library API is bad. Which isn't really a radical position >.>
70.
by valarauca1 8y ago
The first provided link is extremely useful as it lets you jump to declarations, files, and line numbers in the current Linux Kernel source. For digging up kernel code snippets it it is extremely useful. Also like the kernel is fairly sanel
71.
by valarauca1 8y ago
code sample + name (and already dealing with SIMD alignment issues) told me what the problem is. Pointer casts are ALWAYS bad in C. Try to avoid them. Reply and down vote you old hat C programmers who will defend this. Your old practices ar
72.
by valarauca1 8y ago
> I guess I wish their original programming didn't skew so heavily that direction. no it really did. "Shipping" is a classic trope where 2 characters have a will they won't they relationship for several seasons. It is
73.
by valarauca1 8y ago
Generics are just syntax sugar around `interface{}` [1] well hopefully the runtime JIT optimizes things for me. I'm rather impressed that when `go` actually implements something I wanted they still manage to implement it in a way that
74.
by valarauca1 8y ago
blatantly p hacking
75.
by valarauca1 8y ago
Great article. If you could over the 8008 and 4004 that'd be awesome :D
76.
by valarauca1 8y ago
In short they don’t, neither does the compiler. I wrote a rather large macro to auto-generate a Java-Ops parser. The I spent the next 4 days re-writing it so I didn’t have to give the Rust compiler 4GiB of stack just to compile. Deeply recu
77.
by valarauca1 8y ago
For TR kind of. In the newer models only half the cores have direct access to a DDR channel. This shows in memory bounded benchmarks, but Epyc (server line) elevates this problem by have 8 (!!) memory channels, instead of the consumer TR’s
78.
by valarauca1 8y ago
I think a lot of the HN community seems themselves too much like grand inventors. Prometheus bringing fire down from the mountain ignorant of the future of war, gun powder, and bombs. My career in technology has taught me its a lot more obv
79.
by valarauca1 8y ago
Geekbenchmark runs it tests in bursts with short delays between them to allow for the SoC to dissipate heat. They implemented that so laptops and cellphones wouldn't overheat, or thermally throttle while running the benchmark. https:&
80.
by valarauca1 8y ago
"Are optimizing compilers a good idea?" Look if you replace 1 representation with another that is _correct_ (to within the human author's reason), but faster. You have the identical discussion, and an answer (yes).
81.
by valarauca1 8y ago
>Geometry processing that's more complex than moving vertices Have you heard about our lord and savior abstract vector spaces? everything is just moving vertices around in linear transformations.
82.
by valarauca1 8y ago
wouldn’t shader libraries be better?
83.
by valarauca1 8y ago
So you use tapes? Seriously tapes are way more dependable than disks. Less moving parts and less chance internal eletral mechanical break down. And check all the other boxes you mention.
84.
by valarauca1 8y ago
Firefox process per tab is behind a feature flag as it’s in testing still
85.
by valarauca1 8y ago
Solutionism SBT has an extremely non trivial start up, and spin up time. Even using a client JVM (more aggressive JITing). You need to know the hot paths, before you can JIT them. And as a CLI tool, by the time you start optimizing your hot
86.
by valarauca1 8y ago
My dude should quit his architect positions. Consistent hashing of messages requires that every application which handles the message will lay the fields out in the same manner, this is a recursive problem for every layer that will put its
87.
by valarauca1 8y ago
how do you deal with power surges during start up and shut down. I imagine the inductive loads/peak inrush would be massive?
88.
by valarauca1 8y ago
Your Intel based CPU’s Management Engine does so it can run that NSA spyware the Internet tells me [1]. Also embedded java, (kind of the JVM without floats, or GC) is common on base band processors for cellphones and satellites. The full mo
89.
by valarauca1 8y ago
>This depends heavily on your workflow. If you find TR memory bound, then you want a server part. TR has 4 memory controllers that will each do DDR4-3200 in TR2 per TFA. That should be able to push ~95 GiB/s. Wrong. You neither read
90.
by valarauca1 8y ago
2/4 CCX’s lack direct memory access. This is kind of a deal breaker. Threadripper is already memory bound, so the 2 new packages will just be waiting on RAM most the time. Cite: http://en.community.dell.com/cfs-file&#x2

More ›