aboutsummaryrefslogtreecommitdiffhomepage
diff options
context:
space:
mode:
authorAndrzej Janik <[email protected]>2024-02-16 00:00:22 +0100
committerAndrzej Janik <[email protected]>2024-02-16 00:00:22 +0100
commitad8632d579c2112eea5f7dfba6815efcae8694f0 (patch)
treec3d71909a09c369e98ea2ccdf16321c3ca0f7481
parent13bf965784bcee152e9591e1b35bc73c60eda723 (diff)
downloadZLUDA-ad8632d579c2112eea5f7dfba6815efcae8694f0.tar.gz
ZLUDA-ad8632d579c2112eea5f7dfba6815efcae8694f0.zip
Add llama.cpp to READMEllama2
-rw-r--r--README.md10
1 files changed, 10 insertions, 0 deletions
diff --git a/README.md b/README.md
index 70b376a..0365540 100644
--- a/README.md
+++ b/README.md
@@ -127,6 +127,16 @@ If an application fails to start under ZLUDA or crashes please check [Known Issu
### Applications
+#### llama.cpp
+
+If you are building llama.cpp with cmake and don't want it to crash on ZLUDA then you should use `CUDA_DOCKER_ARCH=compute_61` like this:
+```
+make CUDA_DOCKER_ARCH=compute_61
+```
+Alternatively, building with cmake should work with no changes.
+
+Performance is currently much lower than the native HIP backend, see the discussion in #102.
+
#### Arnold
* ZLUDA implements minimum of OptiX framework to support Arnold. ZLUDA's OptiX is buggy, unoptimized and incomplete. It's been tested with Arnold 7.1.4.1 command line rendering on Linux.