Re: matmul trouble
- From: nospam@xxxxxxxxxxxxx (Richard Maine)
- Date: Sun, 8 Jul 2007 21:59:10 -0700
Richard Maine <nospam@xxxxxxxxxxxxx> wrote:
Wade Ward <invalid@xxxxxxxxxxxx> wrote:
I thought matmul was the culprit, but it's the write statement.
Well. That's where you see the error. That doesn't necessarily mean that
is actually where the problem is....
Afraid I don't have any good ideas, anyway not ones I could plausibly
work remotely. If I were sitting in front of the machine with code that
demonstrated the problem, I'd probably fiddle around with lots of things
for clues. But I can't do that remotely. Not even worth trying at that
level. I'm talking about the kind of fiddling where I'd probably go
through several dozen fruitless things before getting much of a hint.
Even if you have the patience for that kind of thing remotely, I don't.
That being said, and on a few more minutes of reflection.... :-)
Probably the first pointless fiddle I'd do would be to take the matmul
out of the picture by doing something trivial like z = x instead of the
And another might be to take the advance='no' off of the prior write.
(No, I don't have any good reason for that. It is probably completely
pointless. But well... maybe if the advance='no' doesn't leave the I/O
system in a good state for some reason...)
Both along the lines of simpifying as much as possible.
Richard Maine | Good judgement comes from experience;
email: last name at domain . net | experience comes from bad judgement.
domain: summertriangle | -- Mark Twain
- Prev by Date: Re: matmul trouble
- Next by Date: Re: matmul trouble
- Previous by thread: Re: matmul trouble
- Next by thread: Re: matmul trouble