Mark istrans as inactive

[wsmoses]

Purely a performance optimization

[coveralls]

Pull Request Test Coverage Report for Build 10827927966

Details

• 0 of 1 (0.0%) changed or added relevant line in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage decreased (-0.02%) to 77.486%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
ext/DynamicPPLEnzymeCoreExt.jl	0	1	0.0%

Totals
Change from base Build 10701838580:	-0.02%
Covered Lines:	2712
Relevant Lines:	3500

---

💛 - Coveralls

[wsmoses]

One may equally mark many others, but from the linked issue I saw this was differentiated and we might as well not (and other tools mark as non differentiable already)

…

On Wed, Sep 11, 2024 at 12:50 PM Markus Hauru ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In ext/DynamicPPLEnzymeCoreExt.jl <#658 (comment)> : > @@ -10,4 +10,6 @@ end @inline EnzymeCore.EnzymeRules.inactive_type(::Type{<:DynamicPPL.SamplingContext}) = true ***@***.*** EnzymeCore.EnzymeRules.inactive_noinl(::typeof(DynamicPPL.istrans), args...) = nothing Cool, thanks. Is there a particular reason why marking this one manually has having 0 grad is significant, or is it the case that we might equally mark many other functions, and this one just came up in EnzymeAD/Enzyme.jl#1807 <EnzymeAD/Enzyme.jl#1807>? — Reply to this email directly, view it on GitHub <#658 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJTUXFS2UQECWMT7DNICE3ZWBYDXAVCNFSM6AAAAABOBJI23CVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDEOJXHEZTAOJUHE> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[mhauru]

Thanks @wsmoses!

[wsmoses]

Can this get a patch release?

…

On Thu, Sep 12, 2024, 6:02 AM Markus Hauru ***@***.***> wrote: Merged #658 <#658> into master. — Reply to this email directly, view it on GitHub <#658 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJTUXD6FFLBKAWILEXGGI3ZWFRC3AVCNFSM6AAAAABOBJI23CVHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJUGIZTAMJQGQZTOMY> . You are receiving this because you modified the open/close state.Message ID: ***@***.***>

[mhauru]

Looking into it. A major PR (#575) has been merged into master since the last release, will probably rather make sure that everything is in order with that, and make a breaking release that would include this too.