gnufx

63 ms·

781.
by gnufx 6y ago
Previously on the Taler privacy-friendly online payment system: https://news.ycombinator.com/item?id=15274110
782.
GNU Taler Operational at Bern University of Applied Sciences(taler.net)
4 points by gnufx 6y ago|1 comments
783.
by gnufx 6y ago
I don't know why you'd want to write lambda like that. Long ago, I made a prototype mixfix syntax for Scheme intended for a shell-like REPL with ISWIM overtones. The paren-less function call syntax (terminated by newline or semic
784.
by gnufx 6y ago
How does it compare with JACAL, which is the old Scheme one I know? https://people.csail.mit.edu/jaffer/JACAL
785.
by gnufx 6y ago
Probably the most penetrating review I've ever seen, but only seen after we got somewhat burned by Mathematica early on.
786.
by gnufx 6y ago
This was being touted for the Fedora build system recently, and Jeff Law made typically sensible remarks to temper the enthusiasm. I don't have a link to the devel archive to hand.
787.
by gnufx 6y ago
Unix in the 1970s was completely irrelevant to the context of the article. I don't know to what extent we paid separately for the OS, but we had source for at least MVT/MVS and OS4000, of ones I used, which we didn't later f
788.
by gnufx 6y ago
At least in the previous decade, and even at that time in other parts of UK science, I don't think there was a "software priesthood". It probably invaded astronomy and HEP earlier than elsewhere, and to some extent now resid
789.
by gnufx 6y ago
Some context might be useful. Starlink was one of two somewhat similar subject-specific networks set up in a period of enlightenment by the UK Science Research Council (as I think it was then) operating in the 1980s, particularly for largel
790.
by gnufx 6y ago
Not something I've used, and it's difficult to countenance anything called that, but what Globus (globus.org) has become seems to be quite popular for transferring large research datasets between academic sites which subscribe. Fo
791.
by gnufx 6y ago
I think I got more appreciation of abstraction from physics and programming than what mathematical education I had; I'm weak at maths -- experimental physics doesn't need much :-/ -- but I do understand that weakness, and am
792.
by gnufx 6y ago
It's a pity it's "open source" and not free software including Emacs, in particular, which has been rather successful itself and as a model long-term. I wonder how some of these architectures have really helped with main
793.
by gnufx 6y ago
So true. Also with respect to documentation, which you have to put in a blog rather than contributing it to the project where it can be maintained. It's bizarre to be literally the last port of call for work you've actually done,
794.
by gnufx 6y ago
Yes, teaching that sort of thing is sorely needed in my area of research computing support specifically, and something I've wanted to do. Unfortunately the collaboration I'm involved with only seems concerned with regurgitating S
795.
by gnufx 6y ago
So what are the results with libxsmm and current AMD BLAS, as that must be for small dimensions? The reason it's serial BLAS that mainly matters is that HPC codes are usually parallelized above the BLAS; why do you want the nesting? S
796.
by gnufx 6y ago
It remains to be seen how bad AMD support is on Frontier and El Capitain compared with Aurora (eventually). The new UK Tier 1 system is all Rome CPU.
797.
by gnufx 6y ago
Why this obsession with MKL on AMD hardware? I've long made measurements of OpenBLAS on "large" dimension serial DGEMM, at least. Even on Intel hardware between Westmere and SKX, with the exception of KNL, it was always at
798.
by gnufx 6y ago
I don't remember, because the measurements on the £1M purchase was maybe five years ago, but they taught a useful lesson. I didn't see figures in what I was responding to. If I'd had more influence on the purchase, as oppos
799.
by gnufx 6y ago
Amusingly, OpenBLAS significantly beat the bought-in ACML, on DGEMM, over the six(?) generations of Opteron I had available. AMD learnt.
800.
by gnufx 6y ago
> This is an overstatement. To be generous... Where do you typically see lack of inlining and vectorization with GCC? I'm curious because most times people have said GCC wouldn't vectorize code that I've been able to try,
801.
by gnufx 6y ago
So which are the optimizations the Intel compiler does which GCC can't is asked? I could guess at the reason for a factor of two, but what does the detailed profiling say with equivalent compiler flags? I can also say that GCC is a f
802.
by gnufx 6y ago
You can see BLIS on Intel's home turf at https://github.com/flame/blis/blob/master/docs/Performance.m... (52-core SKX) and compare with OpenBLAS on 32-core Zen1. (Multithreaded BLAS isn't
803.
by gnufx 6y ago
If OpenBLAS' CPU detection fails, you can force it with an environment variable, but why omit AMD's implementation?
804.
by gnufx 6y ago
I don't know about VASP because it's proprietary, but I have compared other DFT code on 64-core Bulldozer with 12-core Sandybridge nodes. I don't remember the numbers, but the high core count was rather effective in reducing
805.
by gnufx 6y ago
The mythology surrounding the Intel tools and libraries really ought to die. It's bizarre seeing people deciding they must use MKL rather than the linear algebra libraries on which AMD has been working hard to optimize for their hardw
806.
by gnufx 6y ago
There's support in hardware tokens to consider too.
807.
by gnufx 6y ago
What makes Julia code shorter and more elegant generally? I strongly disagree with static typing not being useful in scientific computing. Most of that I see isn't dealing with data frames, though I've seen confusion with types
808.
by gnufx 6y ago
An example of OCaml use in scientific computing is to derive the framework for the FFTW library, which is rather important. (No, the library isn't written in OCaml, as I've seen claimed.)
809.
by gnufx 6y ago
> Why do you think that? Observation in research support, I'd guess. It typically no longer seems to be the case that you do whatever you need to for your data.
810.
by gnufx 6y ago
It is a continuing source of wonder how many scientists seem unable to apply scientific method/research to computing. However, yes, things like debugging, measurement, and hardware features should be taught and rarely are apart from c

More ›