RSA -232 - Page 2

jasonp · 2020-02-20, 14:31

Maybe Ilya Popvyan (poily on the forum here) was involved. He contributed Lanczos optimizations several years ago and I know he was running jobs on this machine.

R.D. Silverman · 2020-02-20, 14:32

Quote:

Originally Posted by axn

But there is no mention of speed (i.e. elapsed time),

Huh? What have you been smoking? Speed is not the same as elapsed time.
I assume that you studied the use of units?

And 30 seconds with Google reveals that these particular CPUs run at 3GHz.

Speed is irrelevant to the discussion.

R.D. Silverman · 2020-02-20, 14:35

Quote:

Originally Posted by axn

As for the singular "processor", they also wrote "polinomial".

A clear typo or spelling error. Speculating that they meant processors (plural) is
the wrong thing to do.

Uncwilly · 2020-02-20, 14:54

Quote:

Originally Posted by R.D. Silverman

What was written was "processor". Singular.

Quote:

50 core-years on Intel Xeon Gold 6136 processor.

As opposed to say 100,000 core years on an 8086 processor. They are referring to the amount of work on a processor class, not a single actual unit. This is a press release, not a journal article. "Details will be published in forthcoming scientific paper." Wait for them.

ldesnogu · 2020-02-20, 15:33

-- deleted --

R.D. Silverman · 2020-02-20, 18:35

Quote:

Originally Posted by Uncwilly

As opposed to say 100,000 core years on an 8086 processor. They are referring to the amount of work on a processor class, not a single actual unit. This is a press release, not a journal article. "Details will be published in forthcoming scientific paper." Wait for them.

You are putting words in their mouths. You are speculating about what was meant.
If they meant to say "processor class", they did not do so. I go by what is written and
do not try to guess what is meant.

I have been advised to do so by professional editors. I will take their advice over yours.

ldesnogu · 2020-02-20, 19:25

Quote:

Originally Posted by R.D. Silverman

To quote: " Computational cost of linear system solution step is 50 core-years on Intel Xeon Gold 6136 processor."

This processor has 12 cores. So it took over 4 years to do the LA.......

Since Xeon Gold 6136 was released mid-2017, we have to assume there's some error in the reporting no matter what the editors told you.

Ref: https://ark.intel.com/content/www/us...-3-00-ghz.html

xilman · 2020-02-20, 19:27

Quote:

Originally Posted by R.D. Silverman

Read what was written. They said "processor", (singular) not processors. I assume
that a technical paper/announcement is accurate.

Running on multiple cores on a single chip still qualifies as parallel.

I did exactly that. I also recognised that English is not their native language.

Never ascribe to malice that which is adequately explained by incompetence.

R.D. Silverman · 2020-02-20, 19:42

Quote:

Originally Posted by xilman

I did exactly that. I also recognised that English is not their native language.

Which is why one thing that editors do, before handing off a paper to a referee, is to
recommend to an author that the paper be reviewed by a native speaker before it is
submitted if there is any possibility of language difficulties.

Quote:

Never ascribe to malice that which is adequately explained by incompetence.

Huh? There is no suggestion of either in this context. It is certainly
possible, absent any other information, that they started this computation more than 4
years ago.

Indeed. One might question why they would undertake such a computation at all
given that RSA-240 and RSA-768 were larger. [the latter only very slightly]. One
answer might be [here, I am speculating] that they started the work quite some time
ago.

Uncwilly · 2020-02-20, 20:16

Quote:

Originally Posted by R.D. Silverman

Which is why one thing that editors do, before handing off a paper to a referee, is to recommend to an author that the paper be reviewed by a native speaker before it is submitted if there is any possibility of language difficulties.

This is a press release, not a paper. Learn the difference.

Want to wager how many cores were actually used? The loser would pay to charity of the winners choice and need to post proof to the tread.

R.D. Silverman · 2020-02-20, 22:45

Quote:

Originally Posted by Uncwilly

This is a press release, not a paper. Learn the difference.

Want to wager how many cores were actually used? The loser would pay to charity of the winners choice and need to post proof to the tread.

It was a research announcement of the Russian Academy of Sciences.
It had their letterhead. It was not a letterhead from the news media.

The same standards of accurate scholarship apply to it as they do a paper.

As for the actual number of cores, why does it matter?

2020-02-20, 15:33	#16
ldesnogu Jan 2008 France 2·271 Posts	-- deleted -- Last fiddled with by ldesnogu on 2020-02-20 at 15:35

2020-02-20, 14:31	#12
jasonp Tribal Bullet Oct 2004 2·3·19·31 Posts	Maybe Ilya Popvyan (poily on the forum here) was involved. He contributed Lanczos optimizations several years ago and I know he was running jobs on this machine.