<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On 7 November 2017 at 09:28, Samuel Pitoiset <span dir="ltr"><<a href="mailto:samuel.pitoiset@gmail.com" target="_blank">samuel.pitoiset@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5"><br>
<br>
On 11/07/2017 10:18 AM, Michel Dänzer wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On 07/11/17 10:08 AM, Samuel Pitoiset wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
It seems safe and it improves performance by +4% (73->76).<br>
<br>
Signed-off-by: Samuel Pitoiset <<a href="mailto:samuel.pitoiset@gmail.com" target="_blank">samuel.pitoiset@gmail.com</a>><br>
---<br>
src/amd/vulkan/radv_device.c | 20 ++++++++++++++++++++<br>
1 file changed, 20 insertions(+)<br>
<br>
diff --git a/src/amd/vulkan/radv_device.c b/src/amd/vulkan/radv_device.c<br>
index 1ecf70d4a9..29bbcc5a43 100644<br>
--- a/src/amd/vulkan/radv_device.c<br>
+++ b/src/amd/vulkan/radv_device.c<br>
@@ -341,6 +341,24 @@ radv_get_perftest_option_name(<wbr>int id)<br>
return radv_perftest_options[id].stri<wbr>ng;<br>
}<br>
+static void<br>
+radv_handle_per_app_options(s<wbr>truct radv_instance *instance,<br>
+ const VkApplicationInfo *info)<br>
+{<br>
+ const char *name = info ? info->pApplicationName : NULL;<br>
+<br>
+ if (!name)<br>
+ return;<br>
+<br>
+ if (!strcmp(name, "Talos - Linux - 32bit") ||<br>
+ !strcmp(name, "Talos - Linux - 64bit")) {<br>
+ /* Force enable LLVM sisched for Talos because it looks safe<br>
+ * and it gives few more FPS.<br>
+ */<br>
+ instance->perftest_flags |= RADV_PERFTEST_SISCHED;<br>
+ }<br>
+}<br>
+<br>
VkResult radv_CreateInstance(<br>
const VkInstanceCreateInfo* pCreateInfo,<br>
const VkAllocationCallbacks* pAllocator,<br>
@@ -400,6 +418,8 @@ VkResult radv_CreateInstance(<br>
instance->perftest_flags = parse_debug_string(getenv("RAD<wbr>V_PERFTEST"),<br>
radv_perftest_options);<br>
+ radv_handle_per_app_options(i<wbr>nstance, pCreateInfo->pApplicationInfo)<wbr>;<br>
+<br>
*pInstance = radv_instance_to_handle(instan<wbr>ce);<br>
return VK_SUCCESS;<br>
<br>
</blockquote>
<br>
There should probably be a way to explicitly disable sisched.<br>
</blockquote>
<br></div></div>
mmh, yeah probably. RADV_DEBUG="nosisched" could be a thing then.</blockquote><div><br></div><div>Can't this sort of app-specific stuff go to a drirc-style config file?</div><div><br></div><div>Alex</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5"><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
<br>
</blockquote>
______________________________<wbr>_________________<br>
mesa-dev mailing list<br>
<a href="mailto:mesa-dev@lists.freedesktop.org" target="_blank">mesa-dev@lists.freedesktop.org</a><br>
<a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/mesa-dev</a><br>
</div></div></blockquote></div><br></div></div>