shahmishal ( Mishal Shah )

Okay I’ll try that. But other frameworks (say, PyTorch) consume constant memory during training regardless of the size of dataset. I don’t understand why memory keeps increasing continuously over time until program crashes here in Turicreate. 🤔 This looks like a bug.

rahulbhalley

comment created time in an hour

pull request commentapple/swift-package-manager

[Collections] API changes to support signed collections

@swift-ci please smoke test

yim-lee

comment created time in an hour

pull request commentapple/swift

[CodeCompletion][Sema] Don't filter out any viable solutions when solving for code completion

Build failed Swift Test Linux Platform Git Sha - 7d8b4858ea8ecb5ee0f7a41caa540655acab5210

nathawes

comment created time in an hour

push eventapple/swift-package-manager

Yim Lee

commit sha d08a0198d2124b8d8967a43df7714d501e37f6b8

Move package collection JSON models to a separate module (#3199) Motivation: Some publishers may opt to generate package collections on their own, which will require the JSON models. It might not be desirable to pull in all the dependencies that come with PackageCollections module. Modifications: Move PackageCollections/JSONModel into its own PackageCollectionsModel module and make it a product. Result: Package collection JSON models can be used with minimal dependencies.

view details

push time in an hour

PR merged apple/swift-package-manager

Reviewers

Move package collection JSON models to a separate module next

Motivation: Some publishers may opt to generate package collections on their own, which will require the JSON models. It might not be desirable to pull in all the dependencies that come with PackageCollections module.

Modifications: Move PackageCollections/JSONModel into its own PackageCollectionJSONModel module and make it a product.

Result: Package collection JSON models can be used with minimal dependencies.

+599 -417

6 commit in yim-lee/swift-package-manager/tree/json-model-module apple/swift-package-manager/tree/main

7 comments

11 changed files

yim-lee

pr closed time in an hour

push eventapple/swift

Michael Gottesman

commit sha 47f76f175cf6dd8f81711146e6845006c1912c2f

[sil] Add a variant of SmallSetVector that uses SmallPtrSet internally instead of SmallDenseSet.

view details

Michael Gottesman

commit sha f39df1d73b169d728a3cccd8c1ebe9ee9b0a9734

[sil-combine] Address comments from post-commit review of cd7c9e97e46a0141e37f10fbfb74c094e0576ff3.

view details

Michael Gottesman

commit sha 90b23a6990f17d975640166eb97aef7e6675d13d

[sil] Replace a thinko in getNextInstruction(). We were using std::next on the pointer, not the iterator meaning we were accessing bad memory.

view details

Michael Gottesman

commit sha 1e039cce38201adcdd909089def60bd06aff8043

Merge pull request #35563 from gottesmm/pr-cd7c9e97e46a0141e37f10fbfb74c094e0576ff3-followups [ownership] cd7c9e97e46a0141e37f10fbfb74c094e0576ff3 followups

view details

swift-ci

commit sha d303fac16e663ea1a99acbc01b4b9108c1662e22

Merge remote-tracking branch 'origin/main' into rebranch

view details

push time in an hour

push eventapple/swift

Michael Gottesman

commit sha 47f76f175cf6dd8f81711146e6845006c1912c2f

[sil] Add a variant of SmallSetVector that uses SmallPtrSet internally instead of SmallDenseSet.

view details

Michael Gottesman

commit sha f39df1d73b169d728a3cccd8c1ebe9ee9b0a9734

[sil-combine] Address comments from post-commit review of cd7c9e97e46a0141e37f10fbfb74c094e0576ff3.

view details

Michael Gottesman

commit sha 90b23a6990f17d975640166eb97aef7e6675d13d

[sil] Replace a thinko in getNextInstruction(). We were using std::next on the pointer, not the iterator meaning we were accessing bad memory.

view details

Michael Gottesman

commit sha 1e039cce38201adcdd909089def60bd06aff8043

Merge pull request #35563 from gottesmm/pr-cd7c9e97e46a0141e37f10fbfb74c094e0576ff3-followups [ownership] cd7c9e97e46a0141e37f10fbfb74c094e0576ff3 followups

view details

push time in an hour

PR merged apple/swift

Reviewers

[ownership] cd7c9e97e46a0141e37f10fbfb74c094e0576ff3 followups

Fixes stuff from #35522.

+114 -29

3 commit in gottesmm/swift/tree/pr-cd7c9e97e46a0141e37f10fbfb74c094e0576ff3-followups apple/swift/tree/main

3 comments

4 changed files

gottesmm

pr closed time in an hour

pull request commentapple/swift

[Test] Ran some execution tests on more platforms.

Build failed Swift Test Linux Platform Git Sha - d94ae1d44589a8fb0ea9f8a400a4bb17f869d746

nate-chandler

comment created time in an hour

PR opened apple/foundationdb

Encapsulate PacketWriter methods

Encapsulates serializeBytesAcrossBoundary and nextBuffer methods.

Style

[x] All variable and function names make sense.
[x] The code is properly formatted (consider running git clang-format).

Performance

[x] ~All CPU-hot paths are well optimized.~
[x] ~The proper containers are used (for example std::vector vs VectorRef).~
[x] ~There are no new known SlowTask traces.~

Testing

[x] ~The code was sufficiently tested in simulation.~
[x] ~If there are new parameters or knobs, different values are tested in simulation.~
[x] ~ASSERT, ASSERT_WE_THINK, and TEST macros are added in appropriate places.~
[x] ~Unit tests were added for new algorithms and data structure that make sense to unit-test~
[x] ~If this is a bugfix: there is a test that can easily reproduce the bug.~

+3 -3

1 commit in sfc-gh-tclinkenbeard/foundationdb/tree/encapsulate-packetwriter apple/foundationdb/tree/master

0 comment

1 changed file

pr created time in 2 hours

push eventapple/swift

Mike Ash

commit sha 216e555ad6e60ad651e53ea29462e87f5359301e

[Runtime] Mark swift_asprintf with __attribute__((__format__)) This gives us build-time warnings about format string mistakes, like we would get if we called the built-in asprintf directly. Make TypeLookupError's format string constructor a macro instead so that its callers can get these build-time warnings. This reveals various mistakes in format strings and arguments in the runtime, which are now fixed. rdar://73417805

view details

tbkka

commit sha 35dad1161a7c825dea48d9eb93a5d0a654354671

Fix reflection_Optional_Any to support non-Apple platforms (#35546) * Fix reflection_Optional_Any to support non-Apple platforms Apple and non-Apple platforms happen to use different number of extra inhabitants in 64-bit pointers. This is already supported by the test framework, we just need to fill it in properly for this one test. * Remove OS= restriction

view details

Mike Ash

commit sha 4988127e997b20af58aa6dadf762ad00296a3a05

Merge pull request #35525 from mikeash/swift_asprintf-static-checking [Runtime] Mark swift_asprintf with __attribute__((__format__))

view details

swift_jenkins

commit sha 81b22f3330ac4f13b7c0767121663e1a79ebab87

Merge remote-tracking branch 'origin/main' into next

view details

push time in 2 hours

pull request commentapple/swift

Support SILBoxWithLayout in TypeDecoder

@swift-ci test

adrian-prantl

comment created time in 2 hours

pull request commentapple/swift

Add debug info support for boxed arguments in async functions.

@swift-ci test

adrian-prantl

comment created time in 2 hours

pull request commentapple/swift

Fix two issues with the SwiftGlibc module map

@drodriguez, no problem, I figured you were busy. If we get this modulemap change in first, we may not need to modify that C++ test in the NDK pull, so I will try to get this working first.

martinboehme

comment created time in 2 hours

Pull request review commentapple/swift

[Runtime] #ifdef __clang__ on printf __attribute__ in Portability.h

 size_t _swift_strlcpy(char *dst, const char *src, size_t maxlen); #ifdef SWIFT_RUNTIME_ATTRIBUTE_ALWAYS_INLINE SWIFT_RUNTIME_ATTRIBUTE_ALWAYS_INLINE #endif-__attribute((__format__(__printf__, 2, 3))) inline static int+#ifdef __clang__

Yeah, we could, but at that point, its probably better to introduce a bigger wrapper like the SWIFT_RUNTIME_ATTRIBUTE_ALWAYS_INLINE.

mikeash

comment created time in 2 hours

pull request commentapple/swift

Fix reflection_Optional_Any test to work on non-Apple platforms

Ooops. I already submitted and merged this.

tbkka

comment created time in 2 hours

PR closed apple/swift

Reviewers

Fix reflection_Optional_Any test to work on non-Apple platforms

This fixes up the new reflection_Optional_Any test to correctly work on Linux and other non-Apple platforms.

Background: Apple reserves 4G at the beginning of the address space for 64-bit processes to catch 32-bit errors in software. As a result, any 64-bit pointer with value < 4G is invalid, which allowed Swift to allocate 2G invalid pointer values as "extra inhabitants" for enum coding. But many other platforms reserve only a 4K initial page, so Swift for non-Apple platforms has standardized on 4K "extra inhabitants" for such.

We put in some infrastructure to account for this difference, this PR just adapts this new test to use that infrastructure.

+19 -20

2 commit in tbkka/swift/tree/tbkka/LinuxReflectionUpdate apple/swift/tree/main

0 comment

1 changed file

tbkka

pr closed time in 2 hours

Pull request review commentapple/swift

[Runtime] #ifdef __clang__ on printf __attribute__ in Portability.h

 size_t _swift_strlcpy(char *dst, const char *src, size_t maxlen); #ifdef SWIFT_RUNTIME_ATTRIBUTE_ALWAYS_INLINE SWIFT_RUNTIME_ATTRIBUTE_ALWAYS_INLINE #endif-__attribute((__format__(__printf__, 2, 3))) inline static int+#ifdef __clang__

Per Config.h, couldn't you just check __has_attribute(format)?

mikeash

comment created time in 2 hours