summaryrefslogtreecommitdiffstats
path: root/doc/tuning.md
diff options
context:
space:
mode:
authorfireice-uk <fireice-uk@users.noreply.github.com>2017-10-22 14:14:57 +0100
committerfireice-uk <fireice-uk@users.noreply.github.com>2017-10-22 14:18:32 +0100
commit203fad6181ee224762d5e37828849605c10f5adf (patch)
treefa45d1816538ee964ee13f1f82f7fff550e3dcbb /doc/tuning.md
parent610f4f0fa72c60daa14dc0661f58ba23563afb61 (diff)
downloadxmr-stak-203fad6181ee224762d5e37828849605c10f5adf.zip
xmr-stak-203fad6181ee224762d5e37828849605c10f5adf.tar.gz
Proper removal of gpu_thread_num
Diffstat (limited to 'doc/tuning.md')
-rw-r--r--doc/tuning.md9
1 files changed, 5 insertions, 4 deletions
diff --git a/doc/tuning.md b/doc/tuning.md
index 806f059..6459235 100644
--- a/doc/tuning.md
+++ b/doc/tuning.md
@@ -53,15 +53,16 @@ In the most cases a `worksize` of `16` or `8` is optimal.
### Add More GPUs
-To add a new GPU you need to add a new config set to `gpu_threads_conf` and increase `gpu_thread_num"` to the number of gpus (entries in `gpu_threads_conf`).
-`index` is the number of the gpu.
+To add a new GPU you need to add a new config set to `gpu_threads_conf`. `index` is the OpenCL index of the gpu.
+`platform_index`is the index of the OpenCL platform (Intel / AMD / Nvidia).
+If you are unsure of either GPU or platform index value, you can use `clinfo` tool that comes with AMD APP SDK to dump the values.
```
-"gpu_thread_num" : 2,
-
"gpu_threads_conf" :
[
{ "index" : 0, "intensity" : 1000, "worksize" : 8, "affine_to_cpu" : false },
{ "index" : 1, "intensity" : 1000, "worksize" : 8, "affine_to_cpu" : false },
],
+
+"platform_index" : 0,
```
OpenPOWER on IntegriCloud