Xavier Hallade xavierh
  • Joined on 2022-02-21
Xavier Hallade created pull request blender/blender#123562 2024-06-21 17:36:09 +02:00
Cycles: Update ocloc to 101.5518 on Windows
Xavier Hallade created branch windows_ocloc_update in xavierh/blender 2024-06-21 17:34:55 +02:00
Xavier Hallade pushed to windows_ocloc_update at xavierh/blender 2024-06-21 17:34:55 +02:00
203c547939 Cycles: Update ocloc to 101.5518 on Windows
2950098012 Fix: Issue in openimagedenoise cmake config harvest
3b4d03968b Build: Fix various issue for 4.2 macOS library update
e72730d763 Fix: assert when switching to sculpt mode
23e497d1af Fix: Memory leak in volume with packed file
Compare 10 commits »
Xavier Hallade commented on issue blender/blender#123522 2024-06-21 10:27:02 +02:00
Cycles: Compress kernel binaries

PyPI limit is about the size of the files already compressed, right? You may want to treat disk space and distribution sizes issues separately.

oneAPI kernel DLL is large (mostly because it…

Xavier Hallade pushed to xavier/PackedState at xavierh/blender 2024-06-04 21:54:32 +02:00
fcf5f4d421 Cycles: extend packed GPU integrator state to CUDA and HIP devices
Xavier Hallade pushed to main at blender/blender 2024-06-04 20:01:55 +02:00
a7171e0391 Cycles: oneAPI: Cleanup: Change packed_float3 to float3
Xavier Hallade commented on pull request blender/blender#122242 2024-06-04 18:34:43 +02:00
Build: Upgrade DPC++ (5.2) and Embree (4.3.2)

gitea didn't link all the commits during rebase+merge, here is the list:

  1. d690b08c1f - DPC++ upgrade 2.…
Xavier Hallade deleted branch sycl_update from xavierh/blender 2024-06-04 18:26:29 +02:00
Xavier Hallade merged pull request blender/blender#122242 2024-06-04 18:26:26 +02:00
Build: Upgrade DPC++ (5.2) and Embree (4.3.2)
Xavier Hallade pushed to main at blender/blender 2024-06-04 18:26:25 +02:00
8fa578dcc2 Build: Backport sse2neon MSVC compiler fixes to Embree
d690b08c1f Build: Upgrade Embree to 4.3.2-blender release
d8b3f852b9 Build: Upgrade DPC++ compiler to 5.2 release
Compare 3 commits »
Xavier Hallade commented on pull request blender/blender#122670 2024-06-04 17:16:45 +02:00
WIP: extend "Struct-of-array-of-packed-structs" to all devices

the tricks with offsets is needed only for Apple as other platforms do have un-padded float3, maybe that simplifies things? I've rebased the changes on top of master, not started to implement a…

Xavier Hallade pushed to xavier/PackedState at xavierh/blender 2024-06-04 17:10:51 +02:00
bf931d3d1c Cycles: extend struct-of-array-of-packed-structs to oneAPI device
5be30b7d2b Cycles: "Struct-of-array-of-packed-structs" for parts of the integrator state
c7807a425a EEVEE: Alias/remove legacy RNA material attributes
8db709e149 deps: oiio, openexr, tiff updates
6e0a49265a Fix #122622: Cycles GPU performance regression in certain scenes
Compare 638 commits »
Xavier Hallade pushed to sycl_update at xavierh/blender 2024-06-04 10:47:04 +02:00
0016d836e4 Build: Backport sse2neon MSVC compiler fixes to Embree
Xavier Hallade commented on pull request blender/blender#122242 2024-06-04 09:32:49 +02:00
Build: Upgrade DPC++ (5.2) and Embree (4.3.2)

@Anthony-Roberts having Embree upgrade sse2neon is doable but certainly not on a quick enough timeline for bcon3. It's safer to use your new diff that backports the fix. I have checked the…

Xavier Hallade created pull request blender/blender#122670 2024-06-03 17:51:52 +02:00
WIP: extend "Struct-of-array-of-packed-structs" to oneAPI device
Xavier Hallade pushed to xavier/PackedState at xavierh/blender 2024-06-03 17:48:54 +02:00
89e261739a Cycles: extend struct-of-array-of-packed-structs to oneAPI device
91e28e69bf Address PR feedback
c548500d76 make format
87919f1ac8 Autogenerate packed structs (for structs declared using KERNEL_STRUCT_BEGIN_PACKED)
9fe74d1d21 Cycles: Struct-of-array-of-packed-structs for some bits integrator state
Compare 5 commits »
Xavier Hallade created branch xavier/PackedState in xavierh/blender 2024-06-03 17:48:53 +02:00
Xavier Hallade commented on pull request blender/blender#122015 2024-06-03 14:00:12 +02:00
Cycles: "Struct-of-array-of-packed-structs" for parts of the integrator state

How much does it matter to have a different layout for ray and packed_ray ? I suspect we actually want to apply this for the Intel GPUs too. I've started testing with latest compiler and I see a…