vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

CANN: Fix SOC_TYPE compile bug (#10519)

Browse source 
* CANN: Fix the bug build fail on Ascend310P under two cases:
1) Manual specify SOC_TYPE
2) Under some unusual compile environment

* Update the cann backend News content: Support F16 and F32 data type model for Ascend 310P NPU.

* fix CANN  compile fail bug: the assert in ascend kernel function doesn't supportted on some CANN version
This commit is contained in:
leo-pony 2024-11-28 15:25:24 +08:00 • committed by GitHub
parent b7420131bf
commit 605fa66c50
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
8 changed files with 57 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

3
docs/backend/CANN.md
View file

@ -23,6 +23,8 @@ The llama.cpp CANN backend is designed to support Ascend NPU. It utilize the abi
## News
- 2024.11
- Support F16 and F32 data type model for Ascend 310P NPU.
- 2024.8
- Support `Q4_0` and `Q8_0` data type for Ascend NPU.
- 2024.7
@ -43,6 +45,7 @@ The llama.cpp CANN backend is designed to support Ascend NPU. It utilize the abi
| Ascend NPU | Status |
|:-----------------------------:|:-------:|
| Atlas 300T A2 | Support |
| Atlas 300I Duo | Support |
*Notes:*