Kernel Newbies archive on
 help / color / Atom feed
From: "Valdis Klētnieks" <>
To: "Irfan Ullah" <>
Subject: Re: Software Prefetching using Machine learning
Date: Wed, 09 Oct 2019 15:08:32 -0400
Message-ID: <170918.1570648112@turing-police> (raw)
In-Reply-To: <>

[-- Warning: decoded text below may be mangled --]
[-- Attachment #1.1: Type: text/plain; charset=utf-8, Size: 2404 bytes --]

On Wed, 09 Oct 2019 12:37:51 +0900, Irfan Ullah said:

> Thanks in advance. I am a PhD candidate, and currently I have started
> working on kernel development. My professor told me to implement this paper
> <>. In this paper authors have used machine
> learning to predict the next missed addresses.

From the abstract:

On a suite of challenging benchmark datasets, we find that neural networks
consistently demonstrate superior performance in terms of precision and recall

Delving further into the paper, we discover that the researchers have learned that
if you run Spec CPU2006 enough times, a neural network can learn what memory access
patterns Spec CPU2006 exhibits.

But they don't demonstrate that the patterns learned transfer to any other programs.
And nobody sane runs the exact same program with the exact same inputs repeatedly
unless they're doing benchmarking.

Ah, academia - where novelty of an idea is sufficient to get published, and considerations
of whether it's a *useful* idea are totally disregarded.

> 1) How can I directly store the missed addresses, and instruction addresses
> from kernel handle_mm_fault() to a file?

Don't do that.  Pass the data to userspace via netlink or debugfs or shared
memory or other means, and have userspace handle it.

> 2) How can I use machine learning classifier in the kernel for predicting addresses?    

Well... in general, you won't be able to do much actually *useful*, because of
time scales.  If your predictor says "Oh, program XYZ will need page 1AB83D 20
milliseconds from now", but it takes 10 milliseconds to bring a page in, your
predictor has only 10 milliseconds to make the prediction in order to be

And in fact, you probably have even less, because your predictor has to be fast
enough and use little enough memory that it doesn't significantly affect CPU,
cache, or RAM usage.

> 3) Is there any way to do the machine learning in the user space in python, and
> then transfer  the classifier in bytes forms to the kernel space for address
> predictions ?

Sure, there's plenty of ways, from using shared memory to creating an ioctl().

But all of them are going to have the same "you need to do it in less time than
it takes for the program you're predicting to reach the point for the prediction".

Good luck, you will need it.

[-- Attachment #1.2: Type: application/pgp-signature, Size: 832 bytes --]

[-- Attachment #2: Type: text/plain, Size: 170 bytes --]

Kernelnewbies mailing list

  parent reply index

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-09  3:37 Irfan Ullah (울라 이르판)
2019-10-09  7:44 ` Greg KH
2019-10-09 19:08 ` Valdis Klētnieks [this message]
2019-10-10  0:21   ` Ruben Safir
2019-10-10  0:24   ` Ruben Safir
2019-10-10  2:10     ` Irfan Ullah (울라 이르판)
2019-10-10  2:48       ` Irfan Ullah (울라 이르판)
2019-10-10  5:43         ` Valdis Klētnieks
2019-10-10  5:51         ` Kernel development tools (was Re: Software Prefetching using Machine learning) Valdis Klētnieks
2019-10-10  6:21           ` Irfan Ullah (울라 이르판)
2019-10-10  6:40             ` Greg KH
2019-10-10 13:22       ` Software Prefetching using Machine learning Rik van Riel

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=170918.1570648112@turing-police \ \ \ \

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Kernel Newbies archive on

Archives are clonable:
	git clone --mirror kernelnewbies/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 kernelnewbies kernelnewbies/ \
	public-inbox-index kernelnewbies

Example config snippet for mirrors

Newsgroup available over NNTP:

AGPL code for this site: git clone public-inbox