gnufx

33 ms·

331.
by gnufx 5y ago
> FORTRAN has a REAL 16 type Fortran (note the post-'77 spelling) doesn't. It has KINDs. REAL x is a non-standard extension, and there's definitely no requirement to implement 128-bit floating point. I don't know w
332.
by gnufx 5y ago
What are those applications? They've passed me by unless that's meant to cover large-scale physics measurements of some sort. I don't understand how it would even make sense in mainstream supercomputers. It should be good
333.
by gnufx 5y ago
This bit of [1], referred to in [2], seems relevant: "their site owners, moderators, administrators, and users are required to comply with the Fedora Code of Conduct. Community sites and accounts which are unable to meet this standard
334.
by gnufx 5y ago
Beware that the Polly matmul example apparently pattern-matches and expands the loop structure to the Goto-like form rather than deriving it a priori. (So I was told when raising those results -- I haven't checked the code.) GCC also
335.
Central Bank Accounts are Dangerous and Unnecessary: A critique of two papers(taler.net)
3 points by gnufx 5y ago|1 comments
336.
by gnufx 5y ago
Presumably you use it if you want its advertised feature of providing a highly secure operating system. Does it matter who else uses it, particularly if it provides an advantage over people who don't?
337.
by gnufx 5y ago
There's a difference between commercial and proprietary, and it's a pity the Genode site confuses them, but if you want to make a proprietary version of it Genode Labs advertises the option to pay them, as a commercial software bu
338.
by gnufx 5y ago
Indeed, the BLIS test suite with generic C kernels passes, not just level 3, though only the -funsafe-math-optimizations of -ffast-math is used by default. (I don't know the specifics for OpenBLAS, but it must be similar.) At least so
339.
by gnufx 5y ago
I'm unconvinced by the need for intrinsics in many cases (other than possibly for prefetch). I took out those out of the final version of [1] and got code that actually ran faster on a contemporary processor, targeting the same old SI
340.
by gnufx 5y ago
Yes, GCC fully vectorizes (and unrolls) a micro-kernel loop nest.
341.
by gnufx 5y ago
-ftree-vectorize, but you'll typically want something like -Ofast to allow lax numerics; and -mcpu=native, of course. You also want to ensure you use C "restrict" appropriately, though GCC will multi-version for possible ali
342.
by gnufx 5y ago
https://libre-soc.org/ is going for free hardware with free tools, but the value of "soon" might be quite high. There was a previous item on 180nm silicon: https://news.ycombinator.com/item?id=277
343.
Compilers and 3 orders of magnitude in 60 minutes (2019) [pdf](venge.net)
1 points by gnufx 5y ago|0 comments
344.
by gnufx 5y ago
Submitted before but with little interest.
345.
by gnufx 5y ago
> Legit point. IBM POWER has 128-bit IEEE floats and decimal hardware coprocessors, but I believe the rest of the CPU is still 64-bits wide. Yes (POWER9+, I think), but also 128-bit "IBM" format. Transitioning to an IEEE ABI f
346.
Methodology of Window Management(chilton-computing.org.uk)
30 points by gnufx 5y ago|5 comments
347.
by gnufx 5y ago
An interesting piece of history, which features Gosling on SunDew, which became NeWS.
348.
by gnufx 5y ago
I was a fan of HyperNeWS, but although PostScript could be fun, I found it hard work, and used literate programming for the experimental data analysis system I worked on. It was notable that you could easily try different designs for GUI e
349.
by gnufx 5y ago
I do research computing support, and bioinformatics has long been thought in my circles the nightmare area to support (perhaps supplanted by machine learning these days). To pick up some of that, compiler maintainers obviously disagree abou
350.
by gnufx 5y ago
I thought there'd be mention of CHERI as an up-to-the-minute architecture (as Arm Morello). I don't remember whether it requires modifications to standard C, but there's a C programming guide [1]. People who say everything i
351.
by gnufx 5y ago
Perhaps best experienced with Rhodri Davies calling "The Seamier Side of Playford" at the UK Inter-Varsity Folk Dance Festival (all ages of student mentality welcome), but this year it's the Techno version in Sheffield soon:
352.
by gnufx 5y ago
Argument passing is, or was, defined in the "argument association" section of the standard, which obviously doesn't preclude implementing by reference. I don't know whether the same language was in the '66 standard
353.
by gnufx 5y ago
Someone who wasn't on JANET in the '90s, for instance?
354.
by gnufx 5y ago
Hodgson said — I don't remember where — something like "We'll make it work or die trying".
355.
by gnufx 5y ago
The language doesn't specify optimization, though it has features for optimizers to take advantage of, and the classic loop optimizations were introduced in Fortran H if I remember correctly. The storage association rules aren't
356.
by gnufx 5y ago
I don't remember if it exists for LAPACK, but there's a generic BLAS interface with ifort/MKL. It's just a pity it was never (quasi-)standardized. (The trivial example is Fortran intrinsics.) I was happily using CLOS-l
357.
by gnufx 5y ago
Well, Fortran relies on you obeying the storage association rules that people regularly used to assure me didn't exist.
358.
by gnufx 5y ago
With GCC (now the GNU Compiler Collection) for instance, it's the same compiler for C and Fortran, just different front ends. The statement isn't useful, anyway, as it doesn't say compiled Fortran will be faster. Given the
359.
by gnufx 5y ago
Fortran (and previously FORTRAN) passes by value-return, not by reference, even if Unix f77 and VAX Fortran implemented it that way. If you use f77(1) in the form of f2c, which I don't have to hand, I changing the value of 1 won't
360.
by gnufx 5y ago
I don't remember if it was a rejoinder to Dijkstra, but Knuth wrote "Structured Programming with go to Statements".

More ›