Add sub-RFC for increased availability of NUMA API #1545

aleksei-fedotov · 2024-11-04T14:56:16Z

Description

Add sub-RFC to #1535 for increased availability of NUMA API.

Type of change

Choose one or multiple, leave empty if none of the other choices apply

Add a respective label(s) to PR if you have permissions

bug fix - change that fixes an issue
new feature - change that adds functionality
tests - change in tests
infrastructure - change in infrastructure and CI
documentation - documentation update

Tests

added - required for new features and some bug fixes
not needed

Documentation

updated in # - add PR number
needs to be updated
not needed

Breaks backward compatibility

Yes
No
Unknown

Notify the following users

List users with @ to send notifications

Other information

vossmjp

Some small wording issues.

rfcs/proposed/simplified_numa_support/increased_availability/README.org

Co-authored-by: Mike Voss <[email protected]>

rfcs/proposed/numa_support/increase-numa-support-availability.org

akukanov

I support this proposal.

ilya-lavrenov · 2024-12-09T18:34:13Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+
+Having a dependency on a shared HWLOC library has advantages:
+1. Code reuse with all of the positive consequences out of this, including
+   relying on the same code that has been tested and debugged, allowing the OS


but in fact most of Linux OSes has obsolete hwloc versions and relying on it does not provide benefits.
IMO, having most up-to-date static HWLOC together with recent versions of oneTBB has benefits and fixes / new features are available to oneTBB immediately

Thanks, I did not know that. I will consider this in the future changes.

I'd rewrite it to smth like this:
1. Reliability. Using a tested and debugged shared library, oneTBB benefits from established, reliable functionality.
2. Code Reuse. Reuse the same code across different processes, improving cache locality and reducing memory footprint, which is the primary purpose of shared libraries.
3. Drop-In Replacement. Use your version of HWLOC without recompiling oneTBB. It can be useful in the following cases:

You need to apply a hotfix to support your hardware that has not yet been integrated into the HWLOC project.

You use a HWLOC version that may never be upstreamed. For example, if hardware unavailable to the broader market.

You want to test a development version of HWLOC on your system.

ilya-lavrenov · 2024-12-09T18:39:26Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+   so that the necessary variant of ~tbbbind~ library can be found and loaded.
+2. The drop of support for HWLOC 1.x allows to not introducing additional
+   ~tbbbind~ variant of the library, yet maintaining support for popular
+   versions of HWLOC.


actually, you still introduce new variant of tbbbind, as ALL distributions already ship only libtbbbind_2_5.so

It means, all TBB distribution across all package managers need to be updated

Or they may continue supporting the current variant with system-specific version of HWLOC as it is still a working approach.

ilya-lavrenov · 2024-12-09T18:40:06Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+
+** Disadvantages
+By default still no diagnostics if users failed to setup environment with their
+own version of HWLOC library correctly. Although, specifying ~TBB_VERSION=1~


why user would be its own version of hwloc?

We introduced that approach in the past and it may be backward incompatible if we change it at once. Perhaps, we decide to prioritize own version of HWLOC someday, but AFAIK still need to leave the possibility to specify a user one.

aepanchi · 2024-12-18T10:10:38Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

@@ -0,0 +1,125 @@
+# -*- fill-column: 80; -*-
+
+#+title: Link ~tbbbind~ with static HWLOC to improve predictability of NUMA support API


the title is too long. its okay though since it is not a part of the main doc set, but i suggest rephrasing it to something like that:
Link tbbbind with Static HWLOC for NUMA API Predictability

aepanchi · 2024-12-18T10:12:30Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+
+#+title: Link ~tbbbind~ with static HWLOC to improve predictability of NUMA support API
+
+*Note:* This is a sub-RFC of the https://github.com/oneapi-src/oneTBB/pull/1535.


Suggested change

*Note:* This is a sub-RFC of the https://github.com/oneapi-src/oneTBB/pull/1535.

*Note:* This document is a sub-RFC of the https://github.com/oneapi-src/oneTBB/pull/1535.

aepanchi · 2024-12-18T10:12:49Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+#+title: Link ~tbbbind~ with static HWLOC to improve predictability of NUMA support API
+
+*Note:* This is a sub-RFC of the https://github.com/oneapi-src/oneTBB/pull/1535.
+Specifically, its section about "Increased availability of NUMA support".


Suggested change

Specifically, its section about "Increased availability of NUMA support".

Specifically, the "Increased availability of NUMA support" section.

aepanchi · 2024-12-18T10:22:33Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+Specifically, its section about "Increased availability of NUMA support".
+
+* Introduction
+oneTBB has a soft dependency on several variants of ~tbbbind~, which are loaded


Suggested change

oneTBB has a soft dependency on several variants of ~tbbbind~, which are loaded

oneTBB has a soft dependency on several variants of ~tbbbind~, which

aepanchi · 2024-12-18T10:22:50Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+
+* Introduction
+oneTBB has a soft dependency on several variants of ~tbbbind~, which are loaded
+by the library as part of its initialization stage. In turn, each ~tbbbind~ has


Suggested change

by the library as part of its initialization stage. In turn, each ~tbbbind~ has

the library loads during the initialization stage. Each ~tbbbind~, in turn, has

aepanchi · 2024-12-20T11:16:55Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+** Common Advantages
+- Explicitly tells that the functionality being used is not going to work
+  instead of just being silent.
+- Does not require additional variant of ~tbbbind~ library to be distributed


Suggested change

- Does not require additional variant of ~tbbbind~ library to be distributed

- Avoids the need to distribute an additional variant of ~tbbbind~ library.

aepanchi · 2024-12-20T11:17:00Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+- Explicitly tells that the functionality being used is not going to work
+  instead of just being silent.
+- Does not require additional variant of ~tbbbind~ library to be distributed
+  along with the others.


Suggested change

along with the others.

aepanchi · 2024-12-20T11:17:47Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+- Requires additional step from the user side to resolve the problem. In other
+  words, it does not provide complete solution to the problem.
+
+** Disadvantages of Issuing a Warning


should this section and the next one be subparts of the common disadvantages section?

aepanchi · 2024-12-20T11:18:29Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+  words, it does not provide complete solution to the problem.
+
+** Disadvantages of Issuing a Warning
+- The warning may still not be visible, especially if standard streams are


Suggested change

- The warning may still not be visible, especially if standard streams are

- The warning may be unnoticed, especially if standard streams are

aepanchi · 2024-12-20T11:18:53Z

rfcs/proposed/numa_support/tbbbind-link-static-hwloc.org

+  closed.
+
+** Disadvantages of Throwing an Exception
+- May break existing code as it does not expect an exception to be thrown.


Suggested change

- May break existing code as it does not expect an exception to be thrown.

- May break existing code that does not expect an exception to be thrown.

aleksei-fedotov requested review from vossmjp, akukanov and pavelkumbrasev November 4, 2024 14:56

github-actions bot added the documentation label Nov 4, 2024

vossmjp reviewed Nov 6, 2024

View reviewed changes