Peter, it appears to be the same as this bug:
https://github.com/xianyi/OpenBLAS/issues/2168
I added my info to the discussion.
Thanks for the reminder.
And thank you again to Ivan for the help.
Sarah
On Thu, Jul 18, 2019 at 3:17 PM Peter Langfelder
wrote:
>
> Sarah, if you haven't done so
Sarah, if you haven't done so already, please do us (OpenBLAS users) a
big favor and report the bug, either to Fedora or directly to OpenBLAS
maintainers.
Peter
On Thu, Jul 18, 2019 at 11:46 AM Sarah Goslee wrote:
>
> Wow. You are entirely correct. I would not have been able to pinpoint
> the pr
Wow. You are entirely correct. I would not have been able to pinpoint
the problem, or how to test it. Thank you.
I am unhappy you are right, since these are the fast workstations I
use for all of my heavy-duty analysis, and it's not even *possible* to
rerun everything.
Oh dear.
Here, with the op
On Thu, 18 Jul 2019 13:30:09 -0400
Sarah Goslee wrote:
> I'm not even remotely a hardware expert: if the difference is due to
> changes in the instruction set, I assume that has potential
> consequences for other things, and I just happened to spot it in this
> particular case because it's visual
I'll address all of your questions below, but starting with this:
> Could any of the computers
exhibiting the bizarre behaviour be equipped with an AVX-512-capable
CPU?
Yes. Both computers that are giving the bizarre results have Intel
i9-7900X CPUs; the X-series apparently is AVX-512-capable.
T
On Thu, 18 Jul 2019 11:50:17 -0400
Sarah Goslee wrote:
> The problem is in the conversion from RGB to Lab.
Hmm. Assuming defaults and skipping all checks, convertColor(red.rgb,
from = "sRGB", to = "Lab") amounts to the following:
red.rgb <- t(col2rgb(rep('red',8), alpha = 0)/255)
# let's hope t
I didn't include enough detail, despite the length of my original
email. The problem is in the conversion from RGB to Lab. I converted
it back to RGB because most of us are more familiar with that. Below
are the intermediate steps.
And to make it even more bizarre, it fails with eight or more colo
This is, indeed, bizzare!
On Thu, 18 Jul 2019 10:35:57 -0400
Sarah Goslee wrote:
> I've gotten as far as locating the problem in this line from
> grDevices::convertColor()
>
> xyz <- from$toXYZ(color, from.ref.white)
So if you manually feed the arguments:
Lab <- structure(c(53.48418378524
I am getting unexpected results when converting from RGB to Lab. This
is clearly some kind of configuration problem, but I cannot find it. I
have three linux workstations, both fully updated (except all are
running R 3.6.0 instead of 3.6.1, because that's the latest Fedora 30
binary). I'm running R
9 matches
Mail list logo
