Bill,
Following your great recommendation I created the verb `right` below.
right=: 3 : 0
wh=. gl_qwh''
vd=. (gl_Rotate y, 0 0 1) mp~ 0,.~_3]\_1 fc vertexData
p=. symdat <'vertexData'
d=. 1&fc ,vd
d memw p,0,(#d),2
gl_sel HD
gl_paint''
)
After running the script I can see that `vertexData` has b
Ah, yes... I was taking the length of the representation from 3!:1 as
the size of the array.
You're right that that would leave out some memory management values.
That said, while I have seen m {~^:n or ^&m^:n get used (and that one
tends to be used with a list value for n), I have not seen }.^:n
It's hard to know what typical is. Selection is pretty common, as are
reshape, take, and drop, and they can operate on data without moving
it. Perhaps the most common operations are arithmetic, and they can be
performed in place much of the time. Many operations, like u;.1, u/,
u"n, can opera
Hmmm...
We're working with rank 1 lists here, so 40 bytes overhead for each array.
Also, I think we should assume a 64 bit J implementation, so another 8
bytes for each value in these arrays.
So... 'a' would be 800040 bytes and 'b' would be 80 bytes.
Anyways... the expression b { 999 ,~ ] a
It's hard to know what the interpreter might get away with. How many
bytes will be used by the last sentence in this sequence?
a =. 1e5 ?@$ 1e5
]b =. a i. 4 5 6 7 8
10 4592 7680 5095 18231
b { 999 ,~ ] a
999 5 6 7 8
Henry Rich
On 3/4/2020 1:19 PM, Raul Miller wrote:
Unless you're working with memory mapped files -- which are a
reflection of an external (os) mechanism -- named nouns in J are copy
on write.
For example:
a=:p:i.4
b=:a
a=:a 1}~ 9
a
2 9 5 7
b
2 3 5 7
Thanks,
--
Raul
On Wed, Mar 4, 2020 at 1:14 PM Hauke Rehr wrote:
>
> talking a
J uses reference counts on blocks, and makes a copy whenever a block is
modified and there is need to keep the unmodified version.
Henry Rich
On 3/4/2020 1:14 PM, Hauke Rehr wrote:
talking about copying: I don’t remember having read about
whether J is copy-on-write or when copying will take pl
talking about copying: I don’t remember having read about
whether J is copy-on-write or when copying will take place
is there a defined set of verbs that does copying or does
that take place at a different level?
Am 04.03.20 um 15:29 schrieb Raul Miller:
Yes.
The general case, implementated in
Yes.
The general case, implementated in J, would be more like
>(, <@u)^:n
But, also, if side effects are involved, then computational efficiency of
the ^: implementation will often not be the key priority.
And... copying is one of the faster operations that computers support. Not
the fastest, b
(,u)^:n is a clever form. It does do a lot of copying, which is what
the JE design is trying to avoid; and if u returns a list it will
produce a list rather than a table.
I think it's harsh to say that the code with restarts is not 'working
right'. It just has the side effect. I agree that
And if you do need to keep the results, something like (,u)^:n would do it...
That said, I hope the other forms are fixable and get fixed to work
right. (Working correctly should take precedence over working fast,
and in the extreme cases where that's a significant problem, that just
means you nee
11 matches
Mail list logo
