Comments (3)
KornevNikita
commented on June 12, 2024
@sarnex could you take a look?
from llvm.
sarnex
commented on June 12, 2024
Yes sure
from llvm.
sarnex
commented on June 12, 2024
Unfortunately this is a known limitation with the IGC vector compiler (used by ESIMD) and AOT mode. Per-kernel specification is not currently supported. There is already an open internal ticket for IGC tracking this limitation. If you have to use AOT mode, per-kernel GRF specification won't work, so I recommend using the IGC option to set auto GRF mode for all kernels, which should pick large GRF for kernels where it will have a benefit and leave others with small GRF mode. The IGC option is -ze-intel-enable-auto-large-GRF-mode. You can also force large mode for all kernels with -ze-opt-large-register-file. These will go inside your -Xsycl-backend string.
from llvm.
Related Issues (20)
- Turn on LLVM_USE_SPLIT_DWARF for Windows Debug build when split-dwarf option is supported
- Enhance `nd_item` code coverage in CI validation.
- [CI][CUDA] No GPU device has been found on the system with NVIDIA GPU in pre-commit. HOT 1
- Add environment variable to redirect SYCL_PI_TRACE to file
- Add WarpReduce to GENX
- [CUDA][HIP] user_defined_reductions HOT 1
- Align terminology used by ONEAPI_DEVICE_FILTER and SYCL_DEVICE_ALLOWLIST HOT 4
- [SYCL] Non-templated aligned_alloc_xxx functions return non-null when the alignment argument is not a power of 2 on GPU
- [SYCL] throw-exception-for-out-of-registers-on-kernel-launch.cpp failing in WW06'24 pulldown HOT 3
- [SYCL] Several tests failed with OpenCL FPGA in precommit HOT 2
- [SYCL-FUSION] [DebugInfo] Enable support of non-instrinsic debug-info HOT 2
- `sycl::kernel_handler::get_specialization_constant` calls `__sycl_getComposite2020SpecConstantValue` on enumerated `value_type` HOT 5
- [SYCL][CUDA][BUILD] cuPTI error while building the compiler with CUDA backend HOT 3
- SYCL stream print nothing when kernel is in endless loop HOT 3
- Device code not split in modules when using aspect::atomic64
- SYCL :: Graph/RecordReplay/host_task_multiple_roots.cpp HOT 2
- Error in configure.py on windows 11 HOT 1
- [SYCL][Matrix] Replace size_t with uint32_t to specify joint matrix rows and columns HOT 4
- Test "ESIMD/unified_memory_api/scatter_lacc_dg2_pvc.cpp" failing on unrelated changes HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from llvm.